The Successor to AC'97: Intel High Definition Audio

An anonymous reader writes "A few days back Intel announced the name to its previously dubbed 'Azalia' next-generation audio specification due out by midyear, under royalty-free license terms. The Intel High Definition Audio solution will have increased bandwidth that allows for 192 kHz, 32-bit, multi-channel audio and uses Dolby Pro Logic IIx technology 'which delivers the most natural, seamless and immersing 7.1 surround listening experience from any native 2-channel source'. The architecture is designed on the same cost-sensitive principles as AC'97 and will allow for improved audio usage and stability."

Comments

It is still onboard sound

[Anonymous Coward]

Will it still also suffer from the same effects of background noise from the rest of the voltage going through the motherboard, or have they found a way to block that out also? 32/192 is fine as a standard... but it is still onboard sound. It needs some seperation from the motherboard to maintain a high S/N ratio

Re:It is still onboard sound

[UrGeek]

Mmmm, what would really be nice if the DAC's were not on the sound chip but in a sheilded housing if it's own and then some nice connectors. And the sound chip would have that digital audio interface - i forgot what it is called - if it even supports something as insane as 32-bits/192kbps

Re:It is still onboard sound

[treat]

If only there were some way to have a digital output from the computer, and do the D/A conversion in a dedicated box.

Re:It is still onboard sound

[xlyz]

use a DAC out of your case

just use digital out to a good A/V receiver

Re:It is still onboard sound

[xlyz]

If only there were some way to have a digital output from the computer, and do the D/A conversion in a dedicated box.
br> there is

digital out is common on today pc (either optical or coax) and any good A/V receiver with integrated decoder is able to convert the signal from digital to analog

Re:It is still onboard sound

[EventHorizon]

Thankfully many onboard audio systems do have an SPDIF optical/coaxial out that you can connect to a dedicated DAC.

Or you can just use USB; the EMagic 2|6 and Edirol UA-5 work well under ALSA and are prosumer studio quality for $200-$300. If you're into softsynth the USB stuff also tends to have low latency, and still works nicely on a laptop.

Re:It is still onboard sound

[Anonymous Coward]

If only there were some way to have a digital output from the computer

Uh... the computer I'm typing this on, with an Asus P4P800 motherboard has a built-in digital coaxial S/PDIF audio output (yes, the motherboard's built-in sound chip - and it does also have analog outputs). I can plug it into my Sony receiver which finds and decodes the digital signal just fine. I've not tried to get surround sound going because I don't have any surround sources, unfortunately.

Of course, that doesn't mean that the shovelware manufacturers (Dell, HP) are going to have digital outputs any time soon - but the moral is "build it yourself" to get the good stuff.

Re:It is still onboard sound

[ethanms]

You're right... but keep in mind that most of the motherboards out there that give out lousy sound from onboard are due to poor layout from the manufacturers... who giving poor layouts because want to save money and physical space on the motherboard, at the expense of analog components like sound...

more bits and more kHz are useless for onboard until you clean up the analog paths to the jack, and properly isolate the codecs on the motherboards using ground moats. Nothing worse then a company that routes a processor +12V feed trace right under the analog side of the codec... or worse, a noisy signal like PS/2 or NIC.

Dear Boss: please don't fire me for this post

Re:It is still onboard sound

[alienw]

There is no physical possibility of having *good* onboard audio. Even with all the above construction techniques, it's damn near impossible to completely isolate the prodigious amounts of digital noise that a typical computer produces.

A much better idea is to run a digital link to an outboard DAC that has its own power supply and is outside the computer. That would actually give you extremely high quality audio, assuming the DAC box is properly designed.

Re:It is still onboard sound

[steve_l]

yes, having the DAC on the same mainboard as the rest of the system, inside a case with a 3+GHz CPU is doomed to create noise.

What I'd like is a pair of wireless headphones that use bluetooth, so my laptop or pc would just stream the digital data out to the BT port; the headphones would pick it up and convert it there and then. That'd be great at work, and great travelling (esp if they were noise cancelling 'phones).

Re:It is still onboard sound

[Detritus]

If they were really serious about noise, they could use RF construction techniques and put the analog components in a shielded can on the motherboard, with bypass capacitors on the power/ground connections. You can shield anything if you are willing to spend some money.

Re:It is still onboard sound

[j3110]

If they are going through that much work, I wouldn't be suprised if there wasn't a seperate card with the DAC that you put in a slot and run cables to. It's been done before, just not for this purpose.

That said, I actually think 32bit audio may be at least 8 bits overkill. I'm all for 192Khz, because we can actually hear a difference in the resolution of the wave. 16bit audio allowed for 64K levels that were smoothed between. Most audio is pretty smooth sounding, and I doubt you can hear any difference between 16 and 32 bit unless you crank the volumn up to a level that could damage your hearing.

Also, 32bit DACs are practically impossible to buy last time I checked. A full 16bit DAC is pretty expensive relatively and it's exponentially more complicated with each bit to build a proper DAC. I'm expecting a lot of shortcuts. A 32bit ADC for recording is prohibitively expensive, so I gaurantee you won't be doing any 32 bit recording any time soon on a PC.

Basically, the 32bit idea is dead in the water. The machine will be long gone before any audio is distributed that takes advantage of it. You probably can't use it for mixing because you probably won't be able to record at 32bit. It's also going to be more expensive in components. Speakers aren't going to be accurate enough to 32bits of resolution. They may shoot for 24 bit, because you can get an OK DAC and ADC for working with 24 bits, but it'll still cost.

The 192Khz thing is awesome. Right now, you can get 48Khz out of some consumer cards, but 192 would be excellent. Maybe we'll get digital audio up to proffesional quality some day. Right now if you go get a recording from a studio, you get tape (unless you can't afford it). All professional audio equipment is not only analog end-to-end, it's also usually tube based. The average transistor is pure sewage, and even MOSFETs are lacking. There's gotta be a lot more R&D into just transistors before we have professional grade audio going anywhere near digital. This is still going to be helpful to the end user that likes music, but we are still a long way off from having no audible differences. Amazingly enough, I think speaker technology has advanced more over the last decade than digital audio.

Re:It is still onboard sound

[foonf]

Like AC97, this is a generic interface for a codec. It doesn't only apply to onboard sound. In fact most consumer sound cards use AC97 codecs, and their sample rate restrictions (requiring everything to be resampled to 16bit/48khz) are a big limitation of most of them. So this will improve the quality of almost anything outside of professional cards and external USB devices.

As for the issue of internal noise, it is really orthogonal to the codec interface specification. You can find cards now, including onboard sound, that are very quiet, using AC97 codecs. It depends on the quality of the codec as well as the motherboard manufacturer (Crystal codecs are very nice, Realtek/Avance Logic codecs are not, they are both AC97 though). This new standard will be no different, it will still depend on the manufacturer. If you really are concerned about noise, you should look into a USB device, although as I'm finding out Linux support for them isn't that great right now.

Re:It is still onboard sound

[Hoser McMoose]

Yeah, it's terrible how nVidia MADE 3Dfx screw up their entire distribution channels, how they made them buy out STB and try to become a card manufacturer. Absolute horrible how they made 3Dfx deliver all their products a year late (or more) and missing much-needed features. And it was especially bad how nVidia made 3Dfx release crappy drivers (or no drivers) for so long.

Face it, 3Dfx killed themselves, nVidia just moved in to pick up the slack. Even if the whole lawsuit between the companies had gone in 3Dfx's favor it's unlikely that they would have managed to survive long enough to see the results of it. A combination of bad decisions and products that were a day late and a dollar short (but still expensive) killed them, not nVidia.

Re:It is still onboard sound

[gidds]

Am I missing the point, or is analogue sound output a red herring here? ISTM that computers should have a digital output (optical or coax or whatever), which would obviate these sorts of problems. You could then use your own separate DAC if you needed an analogue signal, or connect directly to an amp or other bit of hi-fi gear.

(You could have an analogue output as well, of course, but if you were using that in preference to the digital one, then you wouldn't be worrying about sound quality anyway.)

Re:It is still onboard sound

[sfe_software]

It is still onboard sound

Not necessarily. The specification can be used for PCI cards as well, and in fact AC97 is used on some lower-end audio cards. It's more of a specification for minimum supported features and other specs.

The fact that it is on-board in itself doesn't mean it is bad. It's all in the implementation. With proper design techniques (ground-loop isolation, etc) you can get quite a good S/N ratio. It doesn't need "separation from the motherboard", rather, it needs a buffered power bus, separate audio and digital grounds, etc.

The bottom line is, you get what you pay for. If you spend $100 for a motherboard with onboard sound, video, nic, modem, etc... you're likely to get cheap versions of each. If you spend $130 on a PCI sound card, you'll probably get really good specs, whether it is based on AC-97, this new spec, or its own details.

Re:It is still onboard sound

[Anonymous Coward]

32bit DACs are impossible. True 24bit DACs don't even exist.
Also, the highest sample rate available is 192khz not 192bps.

Re:It is still onboard sound

[Valar]

Impossible? Impossible why? I don't see why this would be the case. In fact, I imagine that with minimum wiring you could run two 16bit DACs in parallel, one handling the top 16 bits at twice the voltage, the other handling the low 16.

Re:It is still onboard sound

[Anonymous Coward]

Rubbish. The pci Lynx cards get a real (tested, not just specs) 115db A for the ADC and 113db A for the DAC. That's on an unshielded PCI card. It's all about good circuit design. There is no reason (apart from cost) the specs could not be achived on an on board soundcard.

Re:It is still onboard sound

[ethanms]

I'm guessing you don't work in the industry... it's not only possible, but it's been done on many designs...

Codec construction is important, for example two major suppliers from Taiwan: C-Media and Realtek, are both pretty much crap even on their high end parts... they've traded features and low BOM cost for audio quality...

Other codec supplies, like Analog Devices & Sigmatel (or even Wolfson, Phillips, etc) have put audio quality as a priority to feature sets.

Unfortunately if Realtek rolls out some new feature then the others need to follow or be left behind.

Using ground layers properly, moats and keeping traces near the edge of the board... or even better, making sure you keep the codec as physically close to the jacks as possible, will yield very good results easily rivaling your average sound card.

Let's also keep in mind that an AC'97 or "HD-Audio/Azalia" codec goes for between $0.50 and $1.25...

Where-as a typical SoundBlaster will go from $50-200... they're able to use a lot higher grade support components, and since they are on a PCI card they're able better isolate from the rest of the motherboard (which speaks to your point...)

As for digital out...

Many motherboard manufacturers are finding that the masses are demanding SPDIF (digital) output from onboard sound, it's been available for the past several years from AC'97 vendors, even on most of the low end codecs, but adding the TOS (or even RCA) jacks cost too much in BOM and board real estate (surprise, surprise)...

I think the next big requirement from users will be that SPDIF provide an AC3/DTS signal for all 4/6/8 channel audio. I'm surprised that this wasn't a requirement for Azalia, but we'll see what happens in the near future... After all, AC'97 is currently at version 2.3, there's room for change...

Currently nearly all (even the $200 SB Audigy2) provide only PCM (2-ch) when playing non-DVD audio (when playing DVD they will all pass the AC3/DTS signal out, but they do not generate their own based on a multi-channel game or sound file).

This is mainly due to the licensing fees from Dolby to encode AC3/DTS signals, and partly due to the processing overhead that would be required for implementation in soft-audio.

The exception to this are boards equipped with the nVidia nForce2 audio, they build a DSP into the southbridge(ICH) that encodes AC3 out of any 4/6-ch source being played.

Re:It is still onboard sound

[mlyle]

Impossible? Impossible why? I don't see why this would be the case. In fact, I imagine that with minimum wiring you could run two 16bit DACs in parallel, one handling the top 16 bits at twice the voltage, the other handling the low 16.


You mean the top 16 bits at 65536x the voltage, and the other handling the low 16. Else you've just produced a 17 bit DAC.

Re:It is still onboard sound

[Anonymous Coward]

Well at 32 bit the resolution is smaller than the voltage created by thermal noise, so yes, 32 bits is overkill and useless. To properly sample a signal without aliasing, you need to sample at least twice the highest frequency. Since we can hear up to about 20kHz, you just need to sample at around 40kHz. Higher rates give more accurate representaion, but you get diminishing returns. I would say no more than 4 times the highest frequency is really needed. 192kHz also seems overkill.

As for the tubes comment... I am rolling my eyes. The transistor is well understood. Mosefets operate like the way tubes *should* have worked. Tubes are flawed, get over it. You may like the sound, but think of it as a sort of postprocessing, it is not the true signal.

Re:It is still onboard sound

[Valar]

Err... good point. I was getting my early morning /. fix, so it is no wonder the connections failed :/ I should point out, by the way, that running it at 65536 the voltage would be impractical. You would have to run the bottom one towards the bottom of its speced voltage and the more significant towards the high end. Obviously, the whole thing would need to be pretty sensitive to small changes in voltage...

Re:It is still onboard sound

[crazy_monkey]

There is no physical possibility of having *good* onboard audio. Even with all the above construction techniques, it's damn near impossible to completely isolate the prodigious amounts of digital noise that a typical computer produces.

A much better idea is to run a digital link to an outboard DAC that has its own power supply and is outside the computer. That would actually give you extremely high quality audio, assuming the DAC box is properly designed.


As far as the "onboard" audio, I don't know. But Lynx TWO is a multi-channel professional sound card that fits in a PCI slot (and I mean real professional, not fake-spec professional) Check out the freaking specs!

Re:It is still onboard sound

[Tough Love]

If only there were some way to have a digital output from the computer, and do the D/A conversion in a dedicated box.

Here's one

This box allows you to use spdif with your existing analog stereo.

Specs here

Re:It is still onboard sound

[j3110]

Tubes generally have a flatter curve when comparing amps out to signal voltage, but MOSFETs have a flatter RMS Watts out compared to RMS signal level. Basically, MOSFETs screw up the wave form more than tubes, but manage to preserve the loudness at various frequencies better.

I would rather take the one that can be fixed with an equalizer. :)

Where transistors really rule though is low power usage. A class A tube amp will keep you warm at night without even actually making noise. We need better transistors, but I'm not saying we need tubes everywhere.

What you are saying about needing twice the sampling rate is complete BS. Between that remark and the tube vs MOSFET remark, I can tell you care very little about the wave form.

If I tried to make a 20khz sound with 40khz sampling, I wouldn't hear a thing. Take two samples of one wave equidistant, and you get the middle and the beginning of the wave, which are always the same value. At 4X the sampling, you get a saw wave (or worse, a muffled trapazoid if you phase shift 45 deg). So, if you do 20Khz at 80Khz, you're still screwed. How many points do you need on a wave to make it smooth? I would say at least 8 at high frequencies (and that has a chance of only getting you about 66% of the power). That's about 160Khz for 20Khz sound.

I hate that someone actually thought the whole 2X the audible range was good enough to begin with. You may not hear a difference, but I can. If you don't, check the frequency response of your speakers.

If you want to calculate how many samples it takes to get >90% power, you should calculate the distance on the wave that is 90% power, then divide the length of the full wave by that distance, then multiply by 20Khz. So, here's a trusty sine table for you already measured in percentage of wave: .15 : .81 .18 : .90 .20 : .95 .25 : 1 .30 : .95 .32 : .90 .35 : .81 .32-.18=.14
1/.14 = 7.14
20KHz*7.14=143KHz

This isn't RMS calibrated, but so what.

192KHz: /20KHz=9.6
2pi/9.6=.65 radians
sin(pi/2-pi/9.6)= 94.69% power output.
Basically, it's the sine of half the distance away from the peak to the furthest point out. Why is is it not the average? If you take the average, then you are forgetting that you will miss the right side of the wave completely. The only case better is sometimes when you hit the exact peek.

Also, you have to consider that this is going to create distortion too. Consider that the resonance of 20KHz with the actual output level, since it varies around the 94% mark.

Re:It is still onboard sound

[slaker]

Man, I can't STAND nvidia graphics cards. I have a bunch of overheating testicle-shitting rectal-wart Geforce2s and 3s nailed to the wall at my workbench. I rant about how shitty they are and I'm flamed regularly by people who apparently either like their AGP cards to come with leaf blowers, or who hate ATI even more.

But nforce soundstorm is EASILY the best matrixed stereo-to-surround technology available to consumers right now. I connected the SPDIF out on my HTPC's Asus A7N8X to my Integra 8.2 receiver, and, seriously, it does an amazing job with crummy stereo content on my PC like MP3s. Sometimes better than "real" DD5.1 tracks on DVDs.

Really, honestly, it's as good as you can do, and I don't have any complaints about nforce2 as a motherboard chipset, either.

So here's the thing: nvidia doesn't make chipsets for Intel. DTS Neo:6 is also better than DPL2 IMO. Why the fuck didn't Intel license either of those superior technologies?

Re:It is still onboard sound

[slaker]

Just as a small correction, the nvidia Soundstorm chip matrixes any source (even mono if you're a wild man) to 6 channels. Mostly, it works very well, well enough that I'd probably pass the output from my CD jukeboxes through it if my A7N8X had a digital input.

Re:It is still onboard sound

[oneishy]

At one point in time I would have agreed. I do a lot with live sound [worship teams, bands, spoken messages, etc], and to that end, recording audio as well. I was really frustrated that the audio cards in PC's (even some of the after market ones) are really crap when it comes to recording. Just run some spectrometer software and you might be amazed. Even though the onboard cards do ok for reproducing sound, they do horrible for lower tones when recording.

This is a big reason why the last two computers I have purchased are Macs. The onboard audio on my powerbooks is crystal clear when recording (and I do a lot of it). The same goes for the other G4 I purchased. Granted a high end audio card in a pc might do the same, but this isn't rocket science; why is it not perfect out of the box? Audio has been around for quite a while.

You could make the argument that the extra money i spent on my Macs would have been better served in a high end audio card, but really: quality is worth the extra $$

Re:It is still onboard sound

[SnowZero]

I think the problem is people mix up information coding theory with human perceptual theory. A 40KHz signal *can* encode a 20KHz wave. It will have some pretty wicked aliasing however (as the parent points out), since the volume will be random depending on where you start your sampling (you'll get something randing from 0 to a full amplitude saw wave). The thing about information theory is that it is talking about BITS, i.e. you can construct a sender and receiver that detect the binary presence of a 20KHz signal. The volume will not be detectable with any kind of reasonable accuracy however; you get *only* 1 bit of resolution.

Human perception however, is not even based on evenly spaced intermittent sampling. Instead we have a coiled tube in our ear where different frequencies resonate at different locations (warning, massive oversimplification). Because it is so different, a fixed bandwidth range from 20Hz-20KHz is just a coarse approximation. The best way to test things like that is to do double-blind comparison testing with listeners to see if they can tell the difference between quality levels (blind in that you don't tell the subjects and experimenters which signal is which during testing). Coding theory can guide how we generate the signal, but you still have to test it to see if people can notice the difference.

At least this doesn't bug me as much as people claiming that humans have a 10fps-30fps vision system...

Re:It is still onboard sound

[SnowZero]

My guess is that the point of the "insane" specs is so they don't have to make a new version of the spec any time soon. This should cover anything we need for the forseeable future, even though implementations will not be up the full resolution of the spec for a long time (if ever). For a comparison, we don't need 64bits worth of addressable memory now, as 40bits would be fine for several years. If you're going to change however, might as well make a big change (from 32 to 64) so we won't need to change again any time soon.

Re:It is still onboard sound

[MikeHunt69]

Couple of points:

(1) Name a musical instrument that plays frequencies at 20KHz. The highest note on a violin is 3.5KHz. The top of the piano scale is 4.1KHz. In an orchestra, the only instrument faster than 5KHz is the pipe organ.

(2) Even if you manage to name one, one of the most frequent misconceptions about listening to sound is the very often quoted "20-20 human listening range". Thats just like saying "Humans can run a mile in 4 minutes". Yeah, it's true, but can you do it? I doubt even Audiophile "Golden Ears" couldn't hear above 16KHz.

(3) Okay, even if you have golden 20/20 ears and play a custom made instrument that vibrates to 20KHz, I have one last point. Digital may have sampling problems, but *all* analog systems have noise which rides ontop of the signal (Yes, even your studio class-A ref amps).

You can write down as much math as you want, but in the real world, most people can't tell the difference between digital & analog at fairly decent settings.

Re:It is still onboard sound

[Thornkin]

I think the opposite is true. The human ear can hear approximately 20 KHz of sound. Accounting for the Nyquist frequency and the sampling theorem, that means that 40 KHz (or 44 KHz as CD is today) is totally sufficient. Anything higher than that is just bragging rights. On the other hand, higher-bit audio allows for less quantization error and a cleaner sound. It might help improve things. Higher frequency will just make the sound card more appealing to dogs.

Re:It is still onboard sound

[ketamine-bp]

(1) Many instrument in theory could, with harmonics.

(2) I can hear up to 25,000Hz from medical tests, however, whether i want to hear such frequency is doubtful; yet it is always good to have faithful replication of what is happening on the stage.

(3) Digital have sampling problems, but so are analog - that's true - yet the capability of current digital musical appeartus are way behind the tubes are standing now.

but in the real world, most people care about what they think than what they hear.

Re:It is still onboard sound

[]ix[]

Here we go again, I promised myself not to get involved in "audiphile" discussions again but...

If I tried to make a 20khz sound with 40khz sampling, I wouldn't hear a thing. Take two samples of one wave equidistant, and you get the middle and the beginning of the wave, which are always the same value. At 4X the sampling, you get a saw wave (or worse, a muffled trapazoid if you phase shift 45 deg). So, if you do 20Khz at 80Khz, you're still screwed. How many points do you need on a wave to make it smooth? I would say at least 8 at high frequencies (and that has a chance of only getting you about 66% of the power). That's about 160Khz for 20Khz sound.

But if you use a 40.00001 kHz and a brick wall low pass filter at 20 kHz you will get the original wave.

The only real reason to higher than 2X the audible range is that it is difficult to make brick wall filters. In theory a "muffled trapetzoid will still represent the original wave correctly. Nyquist was pretty clear about this. =).

Linear PCM is just such a waste of space above 44kHz sampling frequency. Everytime you double the sampling frequency you double the amount of data but you only get one octave more information. So if the signal is music, frequency shouldnt be linear. Wavelets for instance are much more efficient at storing data but much more complex to implement. 192x32 is just a brute force way of saying: I cant afford to make decent filters.

Your "power-calculation" just shows that you have never taken a signal processing class in your entire life. But thats ok, those classes are hard.

Re:It is still onboard sound

[olman]

(2) I can hear up to 25,000Hz from medical tests, however, whether i want to hear such frequency is doubtful;

No you can't. Perfect hearing on children goes to about 22kHz, that good would be damn rare in adults. Not to say it wouldn't be too common in children either.

You *can* perceive vibrations up to about 100kHz, but that's completely different kettle of fish. So unless you can prove you're a freak of nature..

Re:It is still onboard sound

[olman]

(1) Name a musical instrument that plays frequencies at 20KHz. The highest note on a violin is 3.5KHz. The top of the piano scale is 4.1KHz. In an orchestra, the only instrument faster than 5KHz is the pipe organ.

Most of them. Electro notwithstanding, you don't hear too many pure sine waves. Instruments produce mixed waveforms that have multiple frequency components. So that 3.5kHz violin sound most definitely has 20kHz components.

Of course with proper filtering you can cut it at 18kHz and 99.8% people can't tell the difference. It's very important to *filter* the sound about to be recorded .. You can get crap result with 44kHz samplerate if you don't filter the signal, as you may end with nice harmonics from higher-frequency components withing audible range.

Re:It is still onboard sound

[walt-sjc]

You would think 64 bit addresses are enough... The architecture of the AS/400 was designed to use a 64 bit address once - and never reuse it. Then they had some customers that used up all their addresses in a year. It's apparently a big deal to regenerate the system to reset back to zero.

What's that old saying that 640K ought to be enough for anyone? :-)

Re:It is still onboard sound

[ketamine-bp]

Nope, I can _hear_ it, and that is from a speech-and-hearing-professional.

Re:It is still onboard sound

[distributed]

I totally agree with you. Signal processing classes are definately hard but once you get in you just cant get out.

Another thing how come krappy inaccurate comments like this(8017856) get a score of 5(Informative) while the appropriate replies(8018859) only get 2 !!

Randomness in selection of moderators can only generate noise.

Re:It is still onboard sound

[olman]

Ok, there's always the last 0.1%. Most adults are hard pressed with 18kHz.

Re:It is still onboard sound

[j3110]

You're the highest ranking responder, so you get picked on :)

Just to answer other people in the thread about why you want the full audible range, there are harmonics in sound that if you wipe out, it makes things sound funny... even human voices.

Now onto your arguement.

It's mathematically impossible to represent analog waves in binary without making assumptions that hurt audio unless you over sample.

1) What if you want a triangle wave instead of a sine wave in your example? What about any wave that is skewed like a lot of audio is, and that defines what an instrument sounds like. A violin doesn't make a perfect sine wave when playing a note, and thus the harmonics aren't going to be the same.

2) Phase shifting the wave to fit is the WORST thing you could do to harmonics, which is what you're getting at 20Khz. I would rather have a power loss than a phase shift, because you can adjust for power loss better with an equilizer. In your model with random frequency shifts, there is no way short of DSP the likes of which would require 500Mhz of CPU power to attempt to correct for. The best thing you may be able to do is try to do a PLL on the signal on a lower frequency and recreate the harmonic by hand.

3) Wavelets would be great, but I don't think you want to throw that kinda power at the system rather than clocking it up. Maybe when an OK DSP that supports wavelets becomes cheaper than an ADC running at 192Khz and 16-24 bit audio, things will be different. As for right now, wavelets are practically theoretical. There aren't enough PhD's in math and engineering not working on military projects to spend time developing wavelets for civilian use.

Nyquist's theorem is pretty much trash.

1) It assumes that you can take a measurement wherever you feel instead of where the clock signals.
2) It assumes that the wave is a sine wave.

1 is unrealistic, and 2 is just plain wrong. I bet you can hear a difference between a sine, triangle, and square wave at 20Khz. I gaurantee you can here a difference when it's phase shifted on top of another lower frequency wave of the same note.

So, it's more than a filtering issue.

Re:It is still onboard sound

[Bistromat]

Like the previous respondent said, you need more DSP before you can comment appropriately.

Regarding 2) above:

The whole point of sampling theory is to recreate sine waves. ANY sound source, be it violins or voices, is not a perfect sinusoidal signal. However, these sources can be represented to the limit of human hearing as a sum of sinusoids (a Fourier series). So, by using a sampling rate at least twice the human range, you should be able to adequately reconstruct triangle and sine waves to the limit of your hearing.

IOW, the "harmonics" you're talking about are, in the frequency domain, representations of sine waves. If you can properly reconstruct those harmonic sine waves to the limit of human hearing, you've got a violin, for all intents and purposes.

And that is why sine waves are awesome.

Re:It is still onboard sound

[]ix[]

It's mathematically impossible to represent analog waves in binary without making assumptions that hurt audio unless you over sample.

That is correct. The issue is how much you have to oversample.

1) What if you want a triangle wave instead of a sine wave in your example? What about any wave that is skewed like a lot of audio is, and that defines what an instrument sounds like. A violin doesn't make a perfect sine wave when playing a note, and thus the harmonics aren't going to be the same.

quite correct. But what is the frequency of the highest harmonic? This is what is called the bandwidh of the signal and if you sample the signal at a sampling rate that is higher than twice the bandwith you will be able to recreate the original signal. This is Nyquist theorem and your belif that this is thrash is what leads you astray.

2) Phase shifting the wave to fit is the WORST thing you could do to harmonics.

If you sample according to Nyquist, the recorded signal will contain all necessary information to recreate the original wave, INCLUDING phase. There is no need to phase shift, or to be concerned about phase at all. The ONLY degenerate case is when you sample at EXACTLY twice the frequency of the signal (as in you example), but if you look closely, Nyquist doesent allow that. And you only have to up the sampling frequency by a notch to get it covered.

3) Wavelets would be great...

Yes they would. Its not just a military technology but the "audiophile"-contolled hifi industry wont catch up in a long time. But it is THE optimum, storage method (that can be proven mathematicly). This we agree upon.

1) It assumes that you can take a measurement wherever you feel instead of where the clock signals.

No it does not. I would again recomend a class in either signal processing, fourier analysis or control theory.

2) It assumes that the wave is a sine wave

No, it only assumes that the wave is continuous and has continuous higher derivatives. Wich all waveforms emanating from real life objets satisfy (but not a triangle wave, a square wave, or any other wave with infinite bandwidh, but those doesent exist in nature).

If the waveform is not a sine it can be described as a sum of sines where the last element in the sum has the same frequency as the bandwidh of the highest harmonic. A perfect triangle wave or square wave has infinite high harmonics and can not be stored perfectly in any way digitaly or analog. There will always be residual errors.

Not You nor any one else can hear the difference between a sine and a triangle wave if the period of the base frequency is 20 kHz. Why, because the second harmonic is at 40kHz and that is way beond human hearing. If they sound differently in your hifi set up its because the higher harmonics are reacing havoc in your filters.

It seems that you have missunderstood the concepts of frequency versus period. Only sines that have been going on forever has one frequency (0 bandwidh), all other waves have bandwith and periods (base frequency).

I highly suggest that you read a book or take a class about this because your have missunderstood quite a lot. Do not underestimate the power of the Nyquist theorem. It is the ultimate sampling theorem in the universe.

PS if my spelling is of its because im Swedish DS

Re:It is still onboard sound

[j3110]

You can not distinguish between sine and triangle waves with only 2 points per wave. You pick two points equidistant, and I'll draw a phase shifted sign wave through them. That is BAD! There is not enough information to properly construct a wave with only two points. There is only enough information to construct a sine wave if you already know the wave is a sine wave. I'm not disputing that if you can accurately represent 20KHz that you would have a problem representing 19KHz overlayed on it.

I gaurantee I know more about DSP than either of you know about music and audio equipment. I once was under the delusion that with enough DSP I could recreate an exact sounding digital implementation of a tube using transistors. I was doing this to digitally record my guitar. The reality of the situation set in while researching ADC (It's very expensive to even get a logarithmic comparitive ADC that can follow 192KHz at 16 bit accuracy (18bit internal)). And transistor charicteristics are more skewed than tubes, so it would take a greatter than 16bit accurate sample in order to map to the appropriate position without making generalizations about the shape of waves. I've seen several waves under O-Scope, and know quite well that they are not sine waves. I wrote some wave files with different wave forms and piped them through audio out, and it's more than noticable at any frequency, just like clipping a wave.

Even after you consider all this, you also have to compensate for speaker response, and microphone/pickup recording characteristics. Pickups on a guitar don't match so great with most audio systems, and you need to work on impedance mismatches.

To actually make a good digital guitar, I was looking at pretty much a PhD level project. In fact, I'm saving the project for just that. It combines extensive Math, CS, Art, EE, and a bit of physics. I very much plan on running both through an analog comparator where the speaker should be (with some phase and volume tuning of course) just to see how close it is. If I get 90% accuracy I'll consider the project a success.

I don't know where your assumption that at highest frequency response of a human ear, you can't tell the difference in waveform comes from, but I assume it comes from you piping sound back into a recording device that has a sampling rate. Yeah, you can reconstruct it so that a sampling device can't tell a difference, but we humans don't sample. Our hearing is a very much more complicated than that. Most people don't know that we use the delay of sound from one ear to another to determine the position of a sound's source. We can distinguish a wave from front or behind, with only two ears in a lot of cases. Feeling is also incorperated into our hearing as much as smell is to tasting.

If you stop making assumptions about human hearing from a DSP perspective, maybe you'll start to see yourself.

Re:It is still onboard sound

[nate1138]

You actually can get 192khz/24bit on a sound card right now.

[Raving fan mode on]
The sound blaster audigy 2 is capable of decoding all DVD-audio formats, including the rarely seen but still awesome sounding 192Khz/24bit stereo format. It can also play back the 48Khz/24bit and 96Khz/24bit 5.1 modes. Add to that a 106DB signal to noise ratio and THX certification, and it sounds pretty sweet. I use one in my home-theatre PC to play back DVD-A.

[Raving fan mode off]

Re:It is still onboard sound

[Woody77]

Cymbals (I probably miss-spelled that).

Especially when playing with brushes. The metal-on-metal creates some, very, very, very high-frequency sounds.

And having listened to a good 96Khz 24bit/channel DAT system, and compared against about $30K in quality hardware starting with a Marantz SACD-1 playing a well-recorded Redbook audio signal, there's still no comparison for the detail of the high-frequency sounds.

If nothing else, it lets the high-frequency sound be a LOT less hard on the ears. Ears like sine-waves, not square-waves... Especially since a square-wave contains the fundamental frequency, and then every frequency above it in a logarithmic decary as the frequency goes up.

Re:It is still onboard sound

[j3110]

That's great! It's pretty much perfect. I hadn't looked at the new Creative cards. I've been using Live! for about 5 or more years now, and while I'm happy with it, analog has still been my friend.

I'll have to look at it's recording capabilities. They usually have a lot of options for that kind of thing.

Re:It is still onboard sound

[nate1138]

Unfortunately, the recording is downsampled to 44.1x16. So not as good for recording. Great card for playback though.

Re:It is still onboard sound

[ThePyro]

Speaking of the Audigy 2, perhaps someone with a little more audio experience can answer this for me:

I built a new PC a couple months ago and stuck a Creative Audigy 2 in it. I've also got a good set of 5.1 speakers. Yet when I turn the volume up to moderately high levels, I can hear an obnoxious buzzing sound whenever I scroll a web page or something... even when there's no music playing in the background. What causes that, and how might I fix it?

Re:It is still onboard sound

[nate1138]

I have that exact problem. It only happens when I am moving my mouse or using the scroll wheel. Haven't been able to fix it yet, but I'll keep you posted.

Have you tried the DVD-Audio playback yet? Mine worked fine for a couple months, but recently it has started to play back VERY choppy and garbled.

Re:It is still onboard sound

[Xyde]

Yes - 192khz - because even a fucking dog can't hear sound waves which oscillate at 96,000 times a second - but any good red blooded audiophile will try to convince you they can.

Re:It is still onboard sound

[ottawanker]

32bit DACs are impossible.

Whether they are impossible or not, they are certainly unnecessary. As long as you get the DAC outside the computer case, a decent 24-bit DAC will work fine. I have an ART D/IO (cheap, but very decent external DAC) hooked up to my soundcard, and it makes a huge difference.

Re:It is still onboard sound

[j3110]

I'm really tired of this arguement, so I'll just point you to a site that says it much better than I want to take the time to say.

Ultrasonics

As you'll read, there has been research done, and it is noticable to chop off frequencies above 22KHz. Granted a single wave is not going to show the difference, but we more than likely detect that waves effects on other waves.

Re:It is still onboard sound

[ethanms]

That's a form of surround-virtualization... you may be correct, I haven't checked it out...

I just know that they are the only vendor (codec or southbridge) to take a 6-ch source (be it actual 6-ch from a game, or mono spread out to 6 wacky channels) and encode AC3 from it...

I refuse to buy motherboards that don't come with SPDIF these days :)

Re:It is still onboard sound

[onomatomania]

You are so completely misguided in your attempts to understand the Nyquist theorem that it's almost comical to anyone that has ACTUALLY TAKEN A SIGNAL PROCESSING CLASS IN COLLEGE. Please, stop before you embarrass yourself further.

1. Every signal can be composed of a series of sinusoidal waves. It's a mathematical fact. Yes, even your precious trangle wave can be represented *precisely* by a series of sine waves.

2. Nyquist's theorem states precisely that you sample at regular periods. I don't know where you're getting this notition that it requires you to sample at irregular intervals.

You need to stop thinking about this in terms of the time domain and learn something about the frequency domain. There is a LOT of mathematics that a LOT of very smart people have worked on for a great deal of time before you were even born. And they completely prove your "theories" wrong.

Take any signal. Any arbitrary signal you want. Sine wave, square wave, etc. Now bandlimit it. It's physically impossible for frequencies to extend out to infinity, so every signal that is physically producable in nature is band-limited. Now sample that signal at some frequency. So long as that sample frequency is twice the band limit, you can reconstruct that signal EXACTLY. Not almost, not partially, not with some distortion, but EXACTLY. You may not understand why this is, but you can prove this mathematically.

Finally, about your little triangle wave diatribe. First of all, a perfect triangle wave is impossible in nature. Notice I said perfect. The reason is that to accomplish that would require and infinite number of harmonics, and thus frequency components out to infinity. YOu can see this from its fourier series. Now, a sine wave, on the other hand, represents a single frequency, or an impulse function in the frequency domain. Anyway, the point here is that if you have a triangle wave of frequency 'n' and a sine wave of frequency 'n', the triangle wave is going to have many, many higher harmonics of 'n', whereas the sine wave does not. Therefore OF COURSE you cannot represent a triangle wave of frequency 'n' by sampling it at '2*n'. To think that you can shows a complete understanding of how the frequency domain works. HOWEVER, if you took that triangle wave, and band-limited it at some cutoff frequency (sufficiently high enough to approximate a perfect triangle wave) and then sampled it at twice that frequency, then you would be able to perfectly reconstruct that waveform.

Re:It is still onboard sound

[Short Circuit]

I wouldn't change the input voltage of the DAC...I'd use a _really_ good amplifier to move the signal range to that which the DAC can use. Preferrably spreading the signal out along the full spec of the DAC to minimize low bit errors.

Re:It is still onboard sound

[Short Circuit]

Sounds like the Sound Blaster Extigy/Audigy.

Re:It is still onboard sound

[j3110]

Then from the obvious observation that you can tell a sine and a triangle wave apart, you should come to the obvious conclusion that in your model a human can detect signals up to about 100KHz. I prefer to think of it more as:

There are two theories for you:

1) Recording devices don't clamp signals exactly at the desired frequency, so signals >22Khz that obviously do exist are interfering with the 22KHz signals in a way that distorts the wave when you try reconstruct it.

This theory doesn't hold because experiments involving humans and the higher frequencies showed that even with analog equipment, humans can detect the difference. This theory does show that Nyquist theory doesn't apply, because the Nyquist theory states that there must not exist any frequency signal above the one you are trying to capture.

2) Fourier theorem doesn't apply to human hearing because human's don't hear frequencies at different levels, they hear the shape (for lack of a better word) of the wave as well, so they can detect the frequencies up to 100KHz, but not hear them.

This is the place where you were supposed to introject a counter theory, not attack my knowledge of Nyquist and Fourier theories. They are both mathematical models. Recording devices sampling back at the same rate are not going to tell a difference because they rely on those same theories. Human's aren't going to like it any more than sticking the wire from the DAC on their tounge because the principals, as of yet, have not involved human perception. At least my monitor's mathematical model took into account the frequency response of my eye. I grow very tired of people testing the principals of their design with systems that assume the principals of their design.

Nyquist and Fourier theories are all nice and good when they are applied to the field they were intended, neither of which are human perception of audio. Nyquist theory applies very well to networks, and fourier works great at seperating and recombining waves. I doubt very much that you could apply fourier transforms in a way that you would understand speech, because that's not the way humans hear.

Re:It is still onboard sound

[Cat_Byte]

Dell does has S/PDIF. Even on some of the laptops.

Re:It is still onboard sound

[forkazoo]

What about running the second DAC at 1.5x the voltage. Think about it for a second, and then tell me how many bits it come out to. Yes, you would have to interleave bits between the DACs, but it still ought to work... :)

Impossible? Impossible why? I don't see why this would be the case. In fact, I imagine that with minimum wiring you could run two 16bit DACs in parallel, one handling the top 16 bits at twice the voltage, the other handling the low 16.

You mean the top 16 bits at 65536x the voltage, and the other handling the low 16. Else you've just produced a 17 bit DAC.

OSS drivers?

[cyb97]

Does the royalty free license also imply that we'll see good opensource drivers for a plethora of platforms?

Re:OSS drivers?

[dreamchaser]

Not necessarily. It's still up to the hardware manufacturers to implement it on their hardware, and then either provide drivers for said hardware or publish their specs as well.

Re:OSS drivers?

[Clockwurk]

It depends on how nice intel is feeling. Royalty free doesn't mean that intel doesn't control it. Royalty free only implies free as in beer, not free as in speech.

Re:OSS drivers?

[DrEldarion]

I just hope for good drivers period. I can't tell you how many times I've had problems with onboard audio even in Windows. I've seen computers where the audio will work flawlessly in Win2k but not in XP, where it'll work fine in XP, but not in 2k, where it'll work fine until you reformat and reinstall the exact same OS, then be broken, etc. etc.

I finally got fed up with it and just got a cheapo PCI card and haven't had any problems since. Incidentally, you get better gaming performance when you don't use onboard audio, too.

Re:OSS drivers?

[DrSkwid]

Even better would be if turning it off in the BIOS meant that the OS actually ignored it.

Re:OSS drivers?

[iammaxus]

What do you mean "control"? What could they possibly do? Take down the PDFs with the specs for the standard from their website? Even if they change the specs significantly in some bad way (which i can't really imagine), other companies will stick with the older working version. If its royalty free, its royalty free.

Re:OSS drivers?

[tloh]

OSS drivers?

Forgive me if I missunderstand, but I hope you don't mean Open Sound System. Last I heard that project was long ago superseded by ALSA Those guys have really been on the ball. I don't think we need to worry too much about not having opensource drivers.

Re:OSS drivers?

[cyb97]

oss being an acronym for open source software, whether the drivers are based on OSS or ALSA I don't really care as long as they work.
Of course ALSA being the sound of the future so it would be better to have ALSA drivers, but I guess as long as one is released (and okay specs are avail.) it shan't take too long before somebody ports them to what ever is the modern state of sound-drivers.

Re:OSS drivers?

[ctr2sprt]

Hardware is a fundamentally different beast than software. Software can be copied and modified easily once the initial version has been created. Hardware, on the other hand, continues to bear an associated cost per-copy even after the initial development is finished. Because of the nature of the medium, after-the-fact modifications are extraordinarily difficult. So it's not really valid to compare hardware licensing to software licensing, at least not using the oversimplified "free as in beer/speech" simile.

In any event, if Intel are letting groups take their spec and implement it in hardware that's meant to be sold for profit... It doesn't get much freer than that. "Free as in speech" doesn't mean you have to give away the farm. You're allowed to keep certain rights for yourself, and make certain restrictions on use, just like open source software does. (And just like there are for free speech, in fact.)

Re:OSS drivers?

[John Courtland]

You know, I have that same damn problem with every Creative Labs product I've ever owned/worked on. They are very very picky about the system configuration they are in, and it's a royal pain to tweak it all properly to get a perfect harmony. I've had them restrict boot up, cause the BIOS to freak and need to be wiped, new OS install, drivers don't install 100% and I have to unpack the EXE myself and fish out the missing .ax and .wdm files, etc etc. Needless to say, my next soundcard is a Philips.

Re:OSS drivers?

[jrockway]

Maybe you should try using a real OS. I've had zero problems with my Live! under Linux, and the sound quality is great.

C'mon, what do you expect from Windows, anyway. Ease of use? Hahahahahahahahahahahahahahahaahah.

Re:OSS drivers?

[John Courtland]

Wasn't only the OS though. It hardlocked my system before boot a few times and corrupted my BIOS data once. It wouldn't even POST with the thing in there, and I did enough testing to prove it wasn't the mobo, PCI slot, RAM timings, voltage, etc etc, it was the sound card.

Re:OSS drivers?

[Zooka]

" Needless to say, my next soundcard is a Philips. "

I have a Philips Acoustic Edge PCI card (specs/review) and liked it a lot. It provided good clean sound to my Klipsch speakers. However, I wasn't very pleased with it when I found out there is no linux support for Philips cards. ... :|

Re:OSS drivers?

[0x1337]

Remove the card - reseat it back. Else try a different PCI slot. I had similar problems. At worst - chuck it and try a different Live!

Re:OSS drivers?

[0x1337]

Toss and replace - you have a defective board. As much has I hate Creative, the Live! has some really good support (and good quality) under Linux.

Initial reaction

[Firehawke]

The very first thing I thought when I saw the article itself was, "Please don't let this be as bad as AC'97."

Don't get me wrong, AC97 is cheap, but it really dragged on the CPUs of the timeframe it came out. This one looks like it might be a shot at the Creative Labs end of the market, but with cheaper components (meaning most likely CPU-based)

I'm sure it'll be on pretty much every board before too long-- well, the non-nForce ones, anyway.

Re:Initial reaction

[BoomerSooner]

Agreed, AC97 is a POS. Every computer I've ever seen someone is using it on the driver implementation and quality is pure shit. Just spend the $50 and get an Audigy card.

Re:Initial reaction

[dnoyeb]

Yea, integratedness has fallen out of favor with me. At least those things that are human detectable such as audio and video.

Integrated sound thus far has been a bad failure. It works well if nothing else is taxing the CPU, but otherwise, it can stutter. My nforce stutters when the network is active so no playing mp3s located on my Linux share...

Re:Initial reaction

[A_Non_Moose]

Agreed, but I thought of their video cards.

Like the AC'97, the vid cards were "functional", but just barely. Heck, even compared to the old Rage 128, it was shameful, IIRC.

Tho, I'd rather have the ati rage in a server, and no sound, nor the intel video cards.

Don't get me wrong, integrated sound/vid/net/whatever is ok, but I agree, it has to be at least of some quality, resource friendly, and stable (like the rage vid cards).

Oh, and just in case it is not the case, being able to disable it is always a must.

How is it on nForce mobo's?

Re:Initial reaction

[Anonymous Coward]

Well, that's the fault of your cheap'n'nasty Nforce chipset, not integrated sound per se.
I've built any number of PCs (all Intel-based) over the last 3 years or so with AC'97 onboard audio, and have never noticed the audio "stutter" under any kind of load.
Sorry, but that's the truth. Don't blame AC'97 just because your particular implementation of it is sucky..

Re:Initial reaction

[Clockwurk]

Having built an Nforce based PC (and supported it) I can safely say that nforce pcs are a joy to work with.

You get good drivers and you only need to install one driver (that covers network, sound, chipset, and graphics). The audio is pretty good quality, and the integrated graphics aren't bad.

I would definately go with an Nforce (for an AMD platform) even if I didn't use any of the integrated components. Nvidia makes excellent chipsets and I don't have to deal with VIA.

Re:Initial reaction

[tarquin_fim_bim]

Save yourself some money and buy a calculator.

Re:Initial reaction

[treat]

Could someone please explain exactly what is wrong with AC97? How could the quality be affected if I'm using the SPDIF out? (And why would you complain about quality if you're not?)

Re:Initial reaction

[Anonymous Coward]

Because AC97 resamples everything internally to 48kHz, including 48kHz streams, so it auto-mangles everything you put through it. If that wasn't enough the Windows sound system (many are afflicted by such voodoo) resamples *everything* through its mixer further mangling the sound before AC97.

Unfortunately SPDIF is not bit-perfect by no means, you need ASIO for that. An easy way to tell is to play a Dolby Digital or DTS .wav through a board, if it arrives at the AV reciever unaffected then the computer isn't screwing with it.

Re:Initial reaction

[Anonymous Coward]

I meant to say that SPDIF can be wonderfully bit-perfect but the use SPDIF doesn't automatically mean so, things are often mangled well before it gets to that stage.

Re:Initial reaction

[antiMStroll]

Agreed. I've installed plenty of broadcast machines using embedded audio (usually ACL650) and never had a fault report about skipping audio. For office and most home use - all those systems with $20 Zoltrix speakers - it's fine.

Re:Initial reaction

[TCM]

Where could one read about this stuff? Let's say I care that a decoded MP3 or an AC3 stream ends up at the digital interface 1:1 without any "mangling". What cards are good for this? Does Windows even allow this etc.?

Re:Initial reaction

[Hobophile]

You get good drivers and you only need to install one driver (that covers network, sound, chipset, and graphics). The audio is pretty good quality, and the integrated graphics aren't bad.

Being "not VIA" alone is enough for me, but I've dealt with both the first nforce and the (vastly superior, performance-wise) nforce2 and really liked both of them.

Using the digital out and a set of Creative Inspire 5.1s, I'm pretty much blown away by the quality of the audio. I'm far from an audiophile but I definitely enjoy the sound. Might be a different story using the onboard DACs though.

It sounds like future nforce boards are not going to include the Soundstorm part, and no decision that I know of has been reached on whether or not to sell it standalone. This is sort of a disappointment and has probably done more to keep me from upgrading to an Athlon 64 processor than any other factor. That combined with the Abit/Nvidia falling out means I am not likely to see an Athlon 64 version of my motherboard any time soon.

In my experience the driver package has been fairly pleasant to deal with, albeit with some small oddities every now and then.

Re:Initial reaction

[Hoser McMoose]

Uhh, yeah, because Creative Labs in the pinnical of driver quality? Umm.. right!

I haven't used an Audigy, so I can't comment on that line, but every Creative card I have used, from the old SB Pro up to an SBLive! had TERRIBLE drivers! Creative is one of the WORST companies out there for driver quality! They do NOT play well with any other components in the system. The SBLive blatently violated the PCI spec, and despite this being a known problem, Creative still hadn't fixed their drivers a year later when I finally dumped the card!

Re:Initial reaction

[John Hurliman]

The important feature to look for is AC3 Passthrough. This takes the raw stream from (lets say) a DVD, whether it's Dolby Digital or DTS, and sends it directly to the soundcard. No DAC or ADC is ever used until the signal reaches your receiver. It gets kind of tricky though because some lower end cards will support AC3 passthrough, but you can't send analog audio (an MP3 or game that goes through the DAC on the card) to the SPDIF channel at the same time. You want a card that does everything, that's why people recommend going for an Audigy. The Creative cards are all you need unless you get in to recording or sound engineering.

Re:Initial reaction

[Hoser McMoose]

The Soundstorm chip is expected to become a standalone part, though it looks like they're primarily targetting on-board audio rather than add-in cards. I suspect that it will be just a single chip with little other than audio and a hypertransport connection to the outside world.

As for the whole Abit/nVidia falling out story, I don't know where the heck The Inquirer got that news (or was it The Register, the other techno-tabloid). It came just days before Abit announced plans for several nForce-based products, including Athlon64 boards.

As for me, I've found the nForce platform to be by far the easiest to deal with. Simply beating VIA isn't exactly impressing me, I sometimes feel that trained monkeys could produce better drivers than VIA. However I've found the nForce drivers to actually work well consistantly. That's something that I can't say about any other company, including Intel (Intel usually gets there eventually, but their first revision or two for a new chipset are often VERY bad).

Re:Initial reaction

[Hobophile]

As for the whole Abit/nVidia falling out story, I don't know where the heck The Inquirer got that news (or was it The Register, the other techno-tabloid). It came just days before Abit announced plans for several nForce-based products, including Athlon64 boards.

Heh, that's what I get for not taking an Inquirer story with a hefty grain of salt. I hadn't heard anything about it either way since then, either confirming or denying. Glad to hear that there's at least good reasons to assume that nothing will come of it.

Re:Initial reaction

[Namarrgon]

Works fine on my nForce2 mobo, and my others too. Perhaps a driver issue?

Re:Initial reaction

[dnoyeb]

it works fine if I play files locally, but playing an mp3 from a remote location is horrible.

No question it could be drivers as the lastest for Linux are a horrible joke and have been as such for over 6 months. Had to roll my own there.

but on windows I am having this problem. Perhaps I should revert back to asus drivers instead of nvidia ones. I have an nforce2 400 board from asus.

Progress In Consumer Audio? Yes!

[ten000hzlegend]

True progress from Intel, strange but true

This new system for audio managment is great news for portable devices such as DVD+screen, next-gen PDA devices and even handheld game systems *Gameboy Advance II or PSP?*

I've long been following PC related audio solutions, all the way from Sonarc to the latest 5 and 6 channel set-ups, my normal set-up is bass speaker, left / right and one for routing system alerts etc... this kind of announcement coupled along with the latest cards supporting the new Dolby processing solutions could well make me upgrade

More to post...

All the usual concerns.

[IGnatius T Foobar]

On its face this is a great announcement, but we must have all the usual concerns. Will it work in Linux? Are the hardware API's going to be published, so someone can write Linux drivers? Or is this going to be the next Centrino, needlessly obfuscated to give Intel's friends in Redmond yet another unfair advantage?

I'm also concerned that a new audio hardware API may introduce way too many opportunities for things like Digital Restrictions Management. Long term, doing that is of course futile because someone will find a way around it, but that doesn't stop some hardware makers from setting out the legal minefield anyway.

It's a sad state of affairs when politics and litigation are at the forefront of geeks' minds when technology ought to be.

Isn't this just a bit much?

[UrGeek]

32-bit audio at 192kbps? Why not just stick with 24bit at 96kbps - it is good enough for most studios. And actually 16-bit at 44.1kbps is the most that these old ears are gonna hear anyway - if even that well after sitting front for Jimi Hendrix.

Re:Isn't this just a bit much?

[ten000hzlegend]

With modern audio requirements, getting as close to the fidelity of the original is the "flavour of the month"

Last year, Pink Floyd released Dark Side on SACD, 24-bit audio at 48khz / 96khz, the amount of clarity over a CD, once the benchmark, was remarkable, I attended a launch party at was blown away even in a relatively acoustic poor setting

I for one welcome consumer 32-bit audio

Re:Isn't this just a bit much?

[Shadowlion]

I for one welcome consumer 32-bit audio

I, for one, welcome our consumer 32-bit audio overlords...

(Sorry, had to do it.)

Re:Isn't this just a bit much?

[boa13]

Last year, Pink Floyd released Dark Side on SACD, 24-bit audio at 48khz / 96khz, the amount of clarity over a CD, once the benchmark, was remarkable, I attended a launch party at was blown away even in a relatively acoustic poor setting

How much of that clarity was due to the excellent sound engineers they probably hired? How much was due to the stage setup, and the excellent speakers and amplifiers they probably had? How did you compare the clarity over a CD? If they offered a comparison, how do you know the CD was a good one, and not a voluntarily dirtied version?

I for one am very wary of launch parties.

Re:Isn't this just a bit much?

[Jeff DeMaagd]

The problem is that at 24 bits per channel, it is impossible to fully realize that sort of dynamic range with physical objects.

The extra eight bits to get to 32 bits is simply a waste. The best I can think of is steganography where you can hide data in the least significant byte and few would catch on unless the data was carefully analyzed.

Re:Isn't this just a bit much?

[Anonymous Coward]

"How much of that clarity was due to the excellent sound engineers they probably hired?"

Most of it. The remasters come from the same source as the original: very high quality analog tape.

On the other hand, I can tell a big difference in dynamics, recording piano at 24/96 versus 16/44. Say what you want about "inaudible this" and "overkill that." Better headroom is higher fidelity. It's just not as important for playback as it is for recording. And, sadly for the future of our human rights, more poeple choose to playback than to record.

Re:Isn't this just a bit much?

[ten000hzlegend]

True, we handed Gary Wright who was announcing the various specifications of SACD at the time of play, a 1984 Dark Side CD, a 1993 20th anniversary CD and finally a copy of Echoes which had the latest digital master before the 30th anniversary re-master

Clean, no scratches and if I recall, the Japan import 1984 cd was worth a mint

Anyhow... we played each one and came to the result that the 2 channel 30th anniversary remaster was far superior, even on a great system, and the surround mix was simply amazing to hear

Re:Isn't this just a bit much?

[dabadab]

Well, considering that you specify the acoustic setting as "relatively poor", I would doubt the difference between SACD and CD would not be drowned out by the background noise.

Re:Isn't this just a bit much?

[kelnos]

exactly. that sampling rate is simply overkill. take a look at an application of the nyquist sampling theorem. human hearing maxes out around 20kHz. 44.1kHz is plenty (and with some breathing room) to sample stuff that humans can hear.

now, the increased resolution offered by 24 bits of accuracy per sample could help. but increasing the sampling rate beyond 44.1kHz does nothing: "No information is lost if a signal is sampled at the Nyquist frequency, and no additional information is gained by sampling faster than this rate."

Re:Isn't this just a bit much?

[DustMagnet]

32 bits is only 10 log10(2^32) or 96 dB. Human hearing can works in a range of 120 dB. Of course that goes from hardly noticed sounds to hearing damage. From what I can tell, hearing damage is pretty popular with the kids I hear driving home from high school.

Re:Isn't this just a bit much?

[JebusIsLord]

In double-blind tests, people have been unable to tell the difference between the SACD layer of the new release and the 1992 CD remaster. The cd-layer on the 30th anniversary version is needlessly overcompressed, probably just to make it sound different than the SACD layer. Try it double-blind, you'd be surprised at how much placebo comes into effect.

Re:Isn't this just a bit much?

[MSTCrow5429]

Hey, you do realize we are a representative republic, not a democracy, right?

Re:Isn't this just a bit much?

[Anonymous Coward]

6dB per bit, so 16 (not 32) bits is 96dB. If it were 3dB/bit, the noise on CD would compare to the hiss on cassette tape without Dolby B.

Re:Isn't this just a bit much?

[Tiro]

kHz is not equal to kbps

Re:Isn't this just a bit much?

[gidds]

The theory for high sample rates (AIUI) is that they allow much gentler filtering, giving less distortion in the audible range.

Standard CDs are sampled at 44.1Khz, so the highest frequency they could possibly store is a sound at 22.05kHz. However, this doesn't meant that they will reproduce anything less than that with perfect accuracy. Firstly, the sound needs to be filtered to prevent anything over 22.05kHz hitting the convertors (as they'd cause very nasty artefacts); this filtering has a lower cut-off (usually around 20kHz) for safety, and although the filter has a steep response, it's not infinite; it'll reduce some lower frequencies too, and it'll also cause phase changes at lower frequencies. (I gather current filters are much better than those used for early CDs, which were responsible for much of the early complaints.) Filters are also needed in the player, which also affect the sound.

Greater sample rates would allow much gentler filters to be used, which would have less (or no) effect on audible frequencies, even those above 20kHz.

Secondly, it's claimed that although we can't hear sound at those higher frequencies, we can detect phase changes and timing changes occurring faster than CD can store; the additional timing resolution would help with that.

And thirdly, in the studio (and wherever sound is processed) the tiny changes caused by filtering and slight timing shifts can add up, to the point (it's claimed) where they can have a very audible result. The extra frequency and time resolution, just like the extra sample resolution of 24 or 32 bits, allows mixing and other processing to be done with less loss.

So there are reasons why 96kHz or 192kHz and 24- or 32-bit sound might provide real benefits. I'm unlikely to hear them myself -- I'm a musician, not an audiophile or sound engineer -- but as technology gets more powerful, faster, and cheaper, I'm sure sound quality will only improve.

(If the RIAA doesn't stop it... Oops, a little bit of politics there, yes indeed.)

Re:Isn't this just a bit much?

[chefren]

The new DSOTM is remixed from 1st generation studio masters, while all the old releases were from 3rd generation master tapes (except for an old 4-track tape release which was from 2nd gen. tapes). So the new release can't be compared to any old release. You could compare it to the cd-layer of the same sacd, though. And of course the new vinyl release beats them all.. :)

Re:Isn't this just a bit much?

[DustMagnet]

Of course dB is 20log10, not 10log10 as I said. Maybe that's what the AC wrote. I don't read AC responses anymore. Using the right formula, 24 bits is over 140 dB of dynamic range, which "should be plenty".

Re:Isn't this just a bit much?

[Shadowlion]

Hey, you do realize we are a representative republic, not a democracy, right?

Yes. It's more of a comment for those people who scream that [insert name] political party is preventing or impeding democracy. The United States never was intended to _be_ a democracy in the first place!

Re:Isn't this just a bit much?

[MSTCrow5429]

That's why I always say that so and so in preventing or imepeding freedom, not democracy, as democracy can be such an unfree thing.

I have particulary fine ears...

[Anonymous Coward]

So, I think I'll wait for 42.1 with 0Hz to 1GHz (+/- 0.0000001%) bandwidth and 256 bits samples audio hardware, which shouldn't be to far away :o)

Re:I have particulary fine ears...

[Tyler Eaves]

It's already available!

Although you'll need a CityBlock-ATX motherboard, just buy a new Carnigie Hall module.

Re:I have particulary fine ears...

[tloh]

Jesus! The constant ding of the everyday sound spectrum must drive you nuts then. I'll bet people look at you funny when they hear nothing but see you shouting "make it stop! make it stop!"

Re:I have particulary fine ears...

[imsabbel]

And those audiophiles will still claim that their vinyl sounds better...

Re:I have particulary fine ears...

[The Munger]

Hey... Aren't you the guy who tried to sell me the Acoustic Enhancement Rock with my speakers?

Re:I have particulary fine ears...

[Hoser McMoose]

Don't forget your solid gold cables with a 1m diameter!

A next gen audio specification?

[xankar]

Hear hear!

Pun completely and totally intended.

the way I look at it

[Bubba]

At least they are changing an old standard that has had mixed issues for several years. New input on old (possibly failed in some aspects) standards is always good for sales.

Re:the way I look at it

[jimmsta]

I'm surprised that the standard hasn't changed much in the past few years. I mean, AC'97 was a stepping stone for standardizing audio on PC's, but why wasn't the standard updated to bring us better audio on motherboards sooner? It IS 2004 after all. I see 1997 standards as being ancient. I'm rather surprised that the AC'97 standard has lasted so long -- longer than VESA (ISA/PCI slot)!! Well...anyway... it's about time!

Linux Logo opportunity?

[bstadil]

Any idea what it would take to use this as an opportunuty to establish a sort of Azalia Certified for Linux Logo and a set of requirements that goes with it?

Logo that you could stick on the box and "Journalists" et al could include in the normal fluffy Buzz Word compliance reviews.

Re:Linux Logo opportunity?

[0x0d0a]

Any idea what it would take to use this as an opportunuty to establish a sort of Azalia Certified for Linux Logo and a set of requirements that goes with it?

That would be you, a web server, and GIMP to do up the logo.

The question is what it would take to get journalists to be interested in it.

Re:Linux Logo opportunity?

[Sumocide]

1. Get a license from Dolby. Good luck with that.

2. Implement the specs

3. ???

4. Profit!!!

That's great! ..

[ShadeARG]

.. but when will we see high definition video support with component and dvi i/o?

Re:That's great! ..

[Wesley Felter]

Even a $50 video card has DVI these days and quite a few cards have component adaptors. Sometimes it takes a bit of fiddling with Powerstrip to convince the card to output weird resolutions, but it's not impossible.

That's audio ?

[Rosco P. Coltrane]

The Intel High Definition Audio solution will have increased bandwidth that allows for 192 kHz,

192 kilo-Hertz? that's more longwave radio than audio. Hell, it's like 5 times the frequency of ultrasounds. Who are they kidding? This smells of marketting bull, or deceptive commercial practices targetted at trendy audio posers ...

Re:That's audio ?

[PatrickThomson]

I gather that with 48khz there are ikky problematic sounds if you forget to filter out high frequecies that reach all the way down into the audible domain - 196khz ensures that these artifacts will be well out of the range of hearing and the abilities of most equipment to reproduce.

Re:That's audio ?

[Professeur Shadoko]

Well, actually, 192KHz is the sampling rate.
Even if frequencies that high cannot be heard, using such a sampling rate will decrease the noise added by analog->digital conversion.

Re:That's audio ?

[Anonymous Coward]

I think they are referring to the sampling rate there (maximum). 192kHz is an amazing sampling rate, current cards are usually 44kHz or 48kHz. The way to think of the sampling rate is the number of times per second the amplitude of the wave is taken -- more samples, more accurate sounds.

Re:That's audio ?

[admbws]

192khz refers the the sample rate, how many times per second the sound is sampled, not how many cycles per second. While theoretically, 192khz sample rate does allow frequencies higher than the ear can hear to be recorded, its real purpose is to make the lower frequencies more accurate - for example, a 22050hz sine tone (if you can hear that high!) sampled at 44100hz is only sampled twice per cycle, and would effectively be recorded as a square wave (although, admittedly at that frequency you'd need to be a dog to tell the difference!)

Re:That's audio ?

[bbbl67]

I don't really think they mean 192 kiloHertz but 192 kilobits per second. There is a difference in the case of lossy-compressed audio. The higher the bps, the less lossy the quality of the audio is. And this bitrate also includes all of the channels together, not just one channel.

Re:That's audio ?

[Anonymous Coward]

for example, a 22050hz sine tone (if you can hear that high!) sampled at 44100hz is only sampled twice per cycle, and would effectively be recorded as a square wave (although, admittedly at that frequency you'd need to be a dog to tell the difference!)


This is completely and utterly wrong. I hear this very often though.

At 44100Hz sampling, a 22050Hz signal will be reconstructed as a 22050Hz SINE WAVE. The reconstruction of sampled signals is not as simple as you think it is. This is covered in any elementary DSP book.

With IDEAL equipment sampling at frequency N allows perfect reconstruction of all frequencies N/2 in all cases. The rather = comes about because of the potential of sampling the frequency N/2 at the zero crossings. However, if only two nonzero points are sampled of the N/2 component, it can be reconstructed perfectly.

Using a higher sampling rate has to do more with counteracting clock jitter and the error introduced by non ideal equipment.

Re:That's audio ?

[DarrylM]

192 kilo-Hertz? that's more longwave radio than audio. Hell, it's like 5 times the frequency of ultrasounds.

Yeah, that is pretty high, but it will allow for a flatter frequency response in the human hearing range than what is possible with 44.1kHz or 96kHz. The reason is that the sampling process has a frequency response of a sync function: sin(x) / x. At a sampling rate of 44.1kHz, the amplitude response of the sample at the high end of the human hearing range will be a fair bit lower than at the low end of the human hearing range. This results in less amplitude (volume) range for the higher frequencies - meaning that the sound won't be quite as close to the original.

When you sample at a higher frequency, the sync function is, in effect, stretched out so that the frequencies at the high end of the human hearing range have a much better amplitude response. Translation: the sound output should, theoretically, be closer to the original at higher frequencies.

Other people have also mentioned the benefits of reduced harmonics and such. As for how much of an actual difference to the perceived sound quality this will make, I have no idea. My speakers aren't all that great, anyway. :-)

Re:That's audio ?

[kelnos]

i don't think so. if they mean 192kbps, then this is a huge step _down_ from 48kHz, 16 bit audio. while something on the order of a 192kbit mp3 is fine by my tastes, it is a huge reduction in quality from, say, a 44.1kHz, 16bit pcm stream from a CD. you just cannot (with current algorithms) losslessly compress an audio bitstream down to 192kbps without losing a good measure of quality. for reference, a pcm CD audio stream runs you around 700kbps. now, things like FLAC can drop this number a bit while remaining lossless. but not that much.

Re:That's audio ?

[JebusIsLord]

completely wrong.

Re:That's audio ?

[bain]

Great ... now I can have a dog whistle track on a CD and play it on repeat when I'm at work.

Give the dog a taste of what I go through at night when I want to sleep.

and nobody can hear it so they won't know whats going on with the dogs barking ..

HAHAHAHAHA -- evil laughter of a madman

Re:That's audio ?

[CSharpMinor]

1) Make sure you're running Linux or another POSIX.
2) Install XMMS.
3) Right-click your playlist, select "Add>URL."
4) Type "tone://18000"
5) Play the tone.

That's 18 kHz. At this response rate, you can hear a clear beat frequency (meaning that it sounds like the volume is rising and falling rhythmically).

Re:That's audio ?

[Chris Johnson]

You're thinking of intermodulation distortion.
Actually, a better reason to do 192K is that you get to filter using a much more suitable filter. At 44.1K the filter has to be a brick-wall infinite slope filter, and those have a characteristic sound. Up at 22K this is more felt than heard, for instance it might feel like 'glare' on cymbals or harmonically rich content or bright ambience like a tiled room. At 192K you can still use a brickwall filter but it's well past even what can be vaguely sensed, but you can also use a better-sounding, gentler filter starting at 20K and ending at 96K.
Also, using the really steep filters means that the reconstructed waveform from extremely harmonically rich content (read: almost any commercial CD released in the last few years) will clip the DAC, and that sounds nasty. You're synthesizing a reconstructed, 'stairstepless' waveform using the filter, and the output wave can easily go way beyond the power supply rail voltages even if your samples aren't clipped. If you want to brute-force prevent this by just turning the gain down, you have to turn it down more than 6 db to accomodate all possible sample values, legal and illegal.
Consider a bunch of -FS samples followed by a +FS sample and then a bunch of -FS samples again. Easily created from legal input by simply turning up the gain in the digital domain. Flat-out illegal values that could not have been sampled directly on a ADC with a working anti-aliasing filter. Present on more CDs than you think... (or values like it, if not literally that)

Re:That's audio ?

[xorbe]

That's because the tone generator is generic and crappy. Done properly, that wouldn't happen. If anything, the pitch would wobble, not amplitude modulation.

Re:That's audio ?

[sweetleaf]

Not quite. The first poster is correct - the simple (naive) decoding _is_ a square wave. The second most naive form of reconstruction, linear interpolation, would yield a triangle (first order derivative of square), while more advanced reconstructions such as sync windowing would filter out the higher frequencies via band-limiting.

However, iirc band-limited approachs affect the phase below the eq cutoff, leaving you with a choice:

Either accurately replicate all frequencies N/2 and lower (while distorting the hell out of the f > N/2 range), or filter out the f > N/2 range and distort the phase of the audio region.

Re:That's audio ?

[1,$d]

Using a higher sampling rate has to do more with counteracting clock jitter and the error introduced by non ideal equipment.

Also, a higher sampling rate means music signals are more likely to retain phase information at high frequencies.

I'm still surprised after all these years that people appear to think frequency is the only signal in audio. Phase is pretty important in music, and I think sampling at 2x the sampled frequency makes a hash of phase info near the hi end (think of your favorite band's ride cymbal, or most instruments in an orchestra, or electronic music chorusing effects). I believe I hear this effect, but I've not read nor conducted any double-blind test addressing this, so - grain of salt.

For lots of audio, like telephony, phase is less important, but Intel's new spec will certainly be used to present music. Music may sound better when recorded with high sample rates and played back using this spec.

Re:That's audio ?

[ja]

No, the pitch will not wobble. The amplitude will.

I can't tell if you're joking or not

[roystgnr]

But assuming you aren't, just find a sound card with a digital output (I think all the higher end cards have SPDIF now) and plug it in to your home theater.

Re:I can't tell if you're joking or not

[Anonymous Coward]

Yea , but the AC97 resampled the spdif out stream -

-greg

Re:I can't tell if you're joking or not

[Anonymous Coward]

If you are running some of the Nforce 2 motherboards with onboard sound, they have digital out. And mine runs much stabler than my audigy 1 did.

Re:I can't tell if you're joking or not

[tho 1234]

Unfortunately, i doubt there will ever be a digital output for high-res audio.

Look at any of the commercial DVD-audio or SACD players available- none of them support digital output at 192khz/24bit. If one was available, anyone could bypass the huge amounts of DRM/watermarking on those new discs, and make bit-perfect copies by simply plugging it into your soundcard/dat recorder.

Anyways, the S/PDIF standard doesn't support bit rates high enough for 192/24 audio, so an entirely new format would have to be made, and somehow i doubt the RIAA will allow one to be made.

Of course, digital output at 44.1/16, (well AC'97 resamples it to 48khz) will still be available, and that's more than good enough for 99.999% of the market.

So basically, this high-res stuff is nothing more than a marketing ploy, there is no way you can achieve 24-bit performance on a noisy switching powesupply while blasted by EMI, reproduced by a 99 cent DAC/opamp chip. (well there's no way to achive 24 bit performance period at room temperature, since the johnston resistor noise of any system is greater than the resolution of a 24bit system, but that's another matter altogether)

Re:I can't tell if you're joking or not

[mn2346]

Actually the technology is available. Sony calls it the i-Link and announced it already in 1999. http://products.sel.sony.com/semi/nr5c_dcp.html

It has been recently added to customer high-end products like SACD, Amplifiers and DVD-players. The backbone of the link is firewire cabling (maybe with sone mods, haven't checked). The RIAA monopolists should have no problem with this since DFM is added.

Re:I can't tell if you're joking or not

[stux]

Actually the technology is available.Sony calls it the i-Link and announced it already in 1999

And Apple (who actually invented it in 1996) and the IEEE (who standardized it) call it FireWire.

I agree though. FW Audio should be the way :)

Re:I can't tell if you're joking or not

[stux]

Ooops, I soooo meant 1991 :)

Re:I can't tell if you're joking or not

[Short Circuit]

24, 32-bit performance...it's all good, and it works. If your interference is random, then you can remove it easily: Sample at N times your intended bitrate, then use average every N bits to get a more accurate sample at a lower bitrate.

To be honest, I have to admit I'm looking forward to even higher sample rates.

Not true discrete channels?

[SpookyFish]

This sounds like it could be more smoke and mirrors, though there really isn't enough information to be sure.

ProLogic IIx will "synthesize" multiple channels from a stereo or 5.1 source. I sincerely hope Intel isn't thinking "we can do the same old thing (stereo) and marketing folks can call it 7.1 multichannel because we put this Dolby fake surround processing in the chip!"

Despite how much ProLogic has advanced, it still doesn't hold a candle to true, *discrete* 6+ channel sound (like DD/AC3 or DTS).

Re:Not true discrete channels?

[Alereon]

The main purpose of this is so that you can listen to your old CDs and MP3s and have them come out of every speaker on your 7.1 surround sound system. Users get annoyed when only two speakers of their expensive system are doing anything. For true multi-channel experiences, we've already DTS ES. Unfortunately, Dolby Digital EX relies on Prologic matrix coding to cram the rear channels into the surround channels to make 7.1, but it's the best you can do without a new audio format.

Why onboard

[j_sp_r]

Just wondering, sound is still inside the computer, onboard, thus crap. The only way to filter out the noise is to make it an external device. But I'm speaking about things I can't afford, with my crappy SB live! and disfunctional stereo plug (it crunches) , oh yeah, computer ungrounded isn't that good for quality music also.

DSD Support?

[babymac]

When will we see support for the DSD audio format in computer hardware? I have yet to hear this technology for myself, but friends who have heard it say it's incredible. Like analog, only better. The one bit tech behind it is very compelling...

Re:DSD Support?

[Wesley Felter]

Never, because then you might copy it.

Re:DSD Support?

[tho 1234]

DSD is just sony's proprietary high resolution format (ie they wanted to start yet another format war, dividing up the 100 or so people intersted in high-res audio between DVD-A and SACD).

The difference between DSD and PCM (used in CD's and DVD-A) really isn't as great as the marketing people try to tell you. In fact, virtually all CD/dvd/DVD-A players made in the last 10 years (excluding some extremely expensive audiophile stuff) convert the PCM to a DSD-like format in the digital to analog conversion process. Any player that says "1-bit dac", "delta sigma" do this.

Basically, PCM stores the data at relatively low rates (44.1, 96, or 192 khz) but takes each mesurement to a high resolution (16 bits or 24 bits). In contrast, DSD samples the signal at several megahertz, but the value stored can only be 1 or 0 (ie 1 bit resolution). Mathematically, the two are basically equivilent once the appropriate filters are applied.

The main reason for DSD is that it is much cheaper to manufacturer high speed digital chips than it is to manufacture precise analog chips. So you can get a much better price/performance ratio by using the DSD format, since the 1-bit DAC is an all digital circuit. (on or off)

So basically everyting you use converts the sound to DSD at the DAC stage, the only area where it makes a difference. All of the mathematical conversions between PCM and DSD are essentially lossess, so it doesn't matter what format the data is stored in on the disc.

However, what does matter is the total amount of information stored on the disc. DVD-A holds 4.7 gigs of audio, while Sony's SACD only holds about 1.5 gig. There is basically no way that SACD can beat DVD-A when it stores that much less information.

Anyways, all of this is very academic, only the most expensive audiophile speakers could possibly show a difference between CD, DVD-A, and SACD, and 99% of the poplulation doesn't own/doesn't want to own such equipment. Playing back these formats on a computer is worthless as they will have to pass through the noisy computer environment ( digital outputs not allowed on high-res formats to prevent piracy).

For any EE's out there, calculate the Johnston resistor noise in a typical system, and compare it to the resolution obtained by 24 bit sampling. You'll find that 24 bits is overkill even in an ideal noise-free environment, and inside a computer, the huge amounts of powersupply, EMI and acoustic noise, means that high-res audio in a computer is simply a marketing ploy.

Re:DSD Support?

[Alereon]

SACD audio data is stored on a DVD disc. A single-sided, single-layer SACD disc has a capacity of 4.5GB, just like DVD, while dual-layer discs store just short of 9GB. See this Sony website for more information on SACD discs.

double-blind, controlled test, please?

[bcrowell]

Last year, Pink Floyd released Dark Side on SACD, 24-bit audio at 48khz / 96khz, the amount of clarity over a CD, once the benchmark, was remarkable, I attended a launch party at was blown away even in a relatively acoustic poor setting
I think you're deluding yourself. Audiophiles make a lot of claims that they can hear certain things, but they never test their own claims using double-blind studies in which the other variables are all controlled for.

I teach a physics lab class, and in one of the labs, I have students test their own hearing, to see the highest and lowest frequencie they can hear. There's some individual variation, but basically the top end of everyone's range comes out to be no less than 10 kHz, and no more than 20 kHz. I have never had a single student who could hear frequencies above 20 kHz.

The 44 kHz (IIRC) sampling frequency of a CD means that you can actually record signals with frequencies as high has 22 kHz (half the sampling frequency -- that's a methematical theorem about the discrete Fourier transform). The reason they designed CD audio around that figure was exactly because of the limits of human hearing.

Even if there was a hypothetical human who could hear 30 kHz, there would be many other things preventing it from being useful musically. For instance, your tweeters most likely can't respond well to those frequencies. Furthermore, the music might sound worse to such a person if the 30 kHz stuff was left in. The musician couldn't hear it, and therefore couldn't adjust his tone to make it sound good. The audio engineer also couldn't hear it, and therefore couldn't judge whether it sounded good or not.

Another practical issue is that distortion will always introduce high-frequency harmonics, so that even if you could hear those frequencies, a lot of what you were hearing would probably be spurious stuff coming from distortion.

People who really want to hear good stereo sound should spend their effort on the two things that will make a lot of difference: (1) getting good speakers, and (2) working on the acoustics of the room, the placement of the speakers in the room, and the placement of their own head in the room. Note that all the stuff under #2 is free or cheap. The audio industry would rather have you waste your money on stuff that's expensive, which is why they promote expensive, superstitious ways of improving sound, such as gold monster cable.

Re:double-blind, controlled test, please?

[Amadan-Na-Briona]

I'm told that due to the joys of the analog electronics, the actual frequencies from a cd range up to 11, not 22 kHz (because the system takes for samples, not two). Therefore, there is a real benefit to the new-wave audio standards that are appearing - as the parent mentioned, most people can hear up to about 20kHz, so we're getting an extra 9kHz for our money's worth.

Re:double-blind, controlled test, please?

[tarquin_fim_bim]

The 44 kHz (IIRC) sampling frequency of a CD means that you can actually record signals with frequencies as high has 22 kHz
However the timbre of the reproduced signal at this frequency will be nothing like the original. Anyway use this frequency is only intended for the aural pleasure of dogs.

Re:double-blind, controlled test, please?

[JebusIsLord]

you were told wrong. use Cooledit or something, remove everything below 11khz on a track and then give it a listen.

16khz is usually a pretty good cutoff for music though - most MP3 encoders cut out everything over 16khz. I can hear up to 22khz test tones, but have a really hard time telling if an actual song was lowpass filtered at 16khz or not.

Re:double-blind, controlled test, please?

[Anonymous Coward]

Poster surely meant removing everything ABOVE said limits.

Re:double-blind, controlled test, please?

[kamelkev]

"The 44 kHz (IIRC) sampling frequency of a CD means that you can actually record signals with frequencies as high has 22 kHz (half the sampling frequency -- that's a methematical theorem about the discrete Fourier transform). The reason they designed CD audio around that figure was exactly because of the limits of human hearing."

You are referring to the Nyquist criterion, which states that in order to guarantee you are not losing analog signal information you must sample your source at twice the frequency of the source.

A detailed explanation of the criterion and theory is here

I don't believe it has anything to do with Fourier, or more likely, it can be understood very simply without any knowledge of advanced mathematics (see the link)

I both agree and disagree with you on your above points... it seems unlikely that the average person can hear about 20khz, but that doesn't necessarily mean that sampling at a higher frequency is pointless. It seems somewhat intuitive that the lower ranges would be that much more "correct". I.E. it can't hurt to sample faster, but it probably doesn't help so much.

Re:double-blind, controlled test, please?

[owlstead]

(2) working on the acoustics of the room, the placement of the speakers in the room, and the placement of their own head in the room. Note that all the stuff under #2 is free or cheap.

Maybe in your test setup, but changing the interior of my room does not come cheap. Furthermore, with few exeptions, there is little or no way to check if everything has been set up correctly. Then again, if I can not hear if it is setup ok or not, then it is probably set up well enough :)

Re:double-blind, controlled test, please?

[JamesP]

This is a very interesting topic and I'll add my 2 cents on it...

Testing for haring BW is just fine. But I was wondering if someone has been (double-blind yadayadayada...) tested against distinguishing, let's say a song using 44ksps and anothe one using 96ksps? Because even if they can't hear "the individual" 25kHz, maybe it DOES have an effect when coupled with a lower frequency signal.

As for more bits per sample, I guess this would improve things (as there would be more bits to "cram everything" inside (I'm talking multiple instruments)

Re:double-blind, controlled test, please?

[Basehart]

"People who really want to hear good stereo sound should spend their effort on the two things that will make a lot of difference: (1) getting good speakers, and (2) working on the acoustics of the room"

You forgot (3) an amp with more power than you need.

I got bored with my home stereo system a year ago and decided to put together more of a studio monitor system instead. I now have a pair of Tannoy Reveal studio monitors hooked up to a Hafler TA 1100 power amp, all for less than $1,000, and I'm really enjoying being able to hear things I've never heard before, with a clarity that gives that "being there" feeling.

In my past life I spent quite a while in recording studios and always marvelled at how awesome the music sounded. Then I'd hear it on a regular hi-fi and it would turn into everything else out there, a watered down version. I used to blame this on the many and varied processes the master had to go through to end up on a retail CD, but since setting up more of a studio monitor type enviroment in my home I'm realizing a lot has to do with having a big fat amp powering a set of big fat speakers.

I also decided to use a Mackie 1202-VLZ Pro mixer to route the audio here and there, instead of the usual hi-fi type amp.

The whole things is mounted in a rack mount cabinet.

Re:double-blind, controlled test, please?

[theLOUDroom]

The 44 kHz (IIRC) sampling frequency of a CD means that you can actually record signals with frequencies as high has 22 kHz (half the sampling frequency -- that's a methematical theorem about the discrete Fourier transform).

Yep, you're denfinately a physics teacher, not an EE.

44 KHz sampling rate only lets you record frequencies up to 22KHz if you had a PERFECT d/a convertor and a PERFECT filter. It is provably impossible to implement a perfect filter. (One with a perfect cutoff and a perfectly flat passband.) Sampling at 44 KHz allows someone to design a decent recording setup with compenents that actually exist. Sampling at 96KHz gives the engineer even more breathing room when designing the filter in front of the A/D convertor. Instead of going from H(jw)=1 to H(jw)=0 in the space of 2KHz, he now can do it in 20. This means he can use a filter design with a flatter pass band. This means there is less distortion of all those frequencies that you can actually hear.

Even if there was a hypothetical human who could hear 30 kHz, there would be many other things preventing it from being useful musically. For instance, your tweeters most likely can't respond well to those frequencies. Furthermore, the music might sound worse to such a person if the 30 kHz stuff was left in.

Actually, it's much easier to build a tweeter than can handle 30KHz, than it is to build a subwoofer that can handle 20Hz. There are plenty of tweeters on the market right now which claim to work at 30KHz.
Second, your statement about the 30KHz stuff making the music sound worse doesn't make any sense. The goal of an audiophile-quality setup is to reproduce the original audio exactly. We're not talking about adding in some strange 30KHz waveform, we're talking about preserving the signals that were there in the first place.

People who really want to hear good stereo sound should spend their effort on the two things that will make a lot of difference: (1) getting good speakers, and (2) working on the acoustics of the room, the placement of the speakers in the room, and the placement of their own head in the room. Note that all the stuff under #2 is free or cheap.

Actually, they should buy a good pair of headphones. For $300 they can buy a pair of headphones that would be tough to beat with speakers at 10X the price.

Re:double-blind, controlled test, please?

[JebusIsLord]

no i meant below, then listen to the resulting file. If you can hear anything, then guess what? CDs have content over 11khz.

Re:double-blind, controlled test, please?

[theLOUDroom]

Okay, I just went through the replies to the grandparent post and I was suprised to find that my post was the only one to mention anything about filters. I guess slashdotters could use a little bit more explanation:

The grandparent poster was referring to Nyquist's theorem. Here's a good link on the subject.

The problem is that he doesn't seem to understand how this is applied in practice. You can't just hook a mic up to the D/A convertor. There is a possibility that signals above the nyquist frequency of your setup are present. These would result in an aliasing effect.
Ex:
If your mic is putting out a 23 KHz sinewave. and you're sampling at 44KHz, this sinewave is going to get shifted down to 1KHz. This is bad, because it trashes the real 1KHz signal that you wanted to listen to, and you get to listend to aliased electrical noise instead.

To prevent high frequencies from messing up your recording, you must place a filter before the A/D convertor. This will block those high frequencies from being digitized, but it introduces a new problem:
no filter is perfect. In an ideal world, you want a filter that would pass everything below 22KHz exactly and block everything above it completely. The problem is that sucha filter is impossible to implement. This means that you end up involved in a trade-off situation. That sharper the cutoff, the less smooth the filter response, etc.

It's a pretty complex subject that I've spent a couple years studying and still don't fully grasp (people spend their whole lives studying filters), but the main point I'm trying to get across is that pretty much any A/D converter has a filter in front of it, and the more extra samples above the nyquist rate you can squeeze in, the less demands are placed on this filter.

Re:double-blind, controlled test, please?

[Listen Up]

How entertaining this thread is. :)

EE is simply a field of applied Physics. An EE is an order of magnitude lower on the food chain than a Physics professor. No Physics=No EE. BUT, this particular Physics lab teacher is most likely a TA who has misinterpreted his calculations and misinterpreted his studied materials. Another 10 years of study or so would have fixed his/her post. A true Physics professor would have never made this mistake.

The minute I read the parent post I realized he was mostly correct in what he was saying, but incorrect in the context of why higher fidelity/higher bitrate audio is needed to produce 'superior' sound. You should have known that and held back your caustic flaming of this person's post. The rest of your post was quite informational. Thank you.

Cheers.

Re:double-blind, controlled test, please?

[caseih]

Having 96 kHz sampling isn't about recording pitches above 22 kHz; it's about getting a better, smoother approximation of the sound waves. With 44 kHz sampling, a 22kHz sound wave is very discrete and choppy. Good ears would almost certainly detect this. At 96 kHz sampling, that same sound wave could be more accurately sampled, allowing for the natural rise and fall of the wave form, rather than just the on/off that you'd get with only 22 kHz sampling rate.

Re:double-blind, controlled test, please?

[theLOUDroom]

no i meant below, then listen to the resulting file. If you can hear anything, then guess what? CDs have content over 11khz.

There are two problems with that:
1) What ever filter cooledit is going to apply isn't perfect (because it's impossible to have a perfect filter)
2) That also makes the assumption that all your math is being done with infinite precision (no rounding occurs, no noise is being generated merely by the application of the filter)

The reality is that CDs do have content above 11 KHz, but no content is at 22KHz. This is becuase of the need for filters before the d/a convertor, see my other posts on this topic.

While a recorder might sample a 44KHz, there is a low pass filter inside it which begins blocking signals at some frequency below 22Khz. This filter is going to vary from unit to unit.

Re:double-blind, controlled test, please?

[JebusIsLord]

yeh, that's all fine. I was just trying to prove to the parent that CDs do have content over 11khz, that's all.

Re:double-blind, controlled test, please?

[hankwang]

To prevent high frequencies from messing up your recording, you must place a filter before the A/D convertor. This will block those high frequencies from being digitized, but it introduces a new problem: no filter is perfect.

Yes, 96/192 kHz sampling is a good thing for recording studios for the reason that you explain. Moreover, >=24 bit recording means that you don't get aliasing problems if the signals are amplified or attenuated during the mixing process.

However, this is all on the recording side. After sampling at >=96 kHz, you can apply a digital filter with a perfectly flat passband up to 20 kHz and stopband above 22.05 kHz, and then downsample to 44.1 kHz. In any CD player, the opposite process is performed (the famous "oversampling"): it is hard to filter the noise above 20 kHz in the raw 44.1-kHz signal. Therefore, the DAC converts the signal digitally to a 4 to 16 times higher sampling rate and with a slightly higher bitresolution (e.g. 18 or 20 bits). Then, the DAC digitally filters out everything above 22 kHz while leaving everything below 20 kHz.

The (still digital) signal is now a "smooth line" through the supplied data points at 44 kHz. This signal is converted to a voltage by the true (non-signal-processing part of the) DAC. The part of the spectrum below 20 kHz will be exactly the same independent on whether the original input to the DAC was 44, 96, or 192 kHz. (Note: 1-bit DA convertors use a slightly different approach, but with the same result).

As far as the bit resolution is concerned: in the final signal, 16 bits is enough for a dynamic range of 92 dB. If the hearing treshold is at 0 dB, that means that for peak levels of less than 92 dB, the resolution is fully sufficient to encode even the softest audible sounds. Note that 92 dB is quite loud: about 4 W power to a typical 87 dB/W loudspeaker at 1 m distance. It is defendable to use a bitresolution higher than 16, if you want to hear a ticking watch in the background while the music is playing at the pain treshold of 120 dB. For that, you need 5 more bits: 21 bits. On the consumer end, 24 or 32 bits is a waste of storage space.

Re:double-blind, controlled test, please?

[S.Lemmon]

Really, I think the only time sampling at a higher frequency or with more bits per sample really makes sense is during recording and mixing. When your combining up sounds from several different sources, that extra numeric accuracy makes for less error in the final mix. However, the final result doesn't need to keep it at that level. Really, it's not unlike flatbed scanners that go beyond 24bit color.

Of course you'll find "Audiophiles" who swear they can hear the difference, but then again these are some of the same folks who insist sticking foam mats to their cds makes them sound better. What people fail to remember with all these remastered CD re-releases, is all the equipment used for the remastering process is better - it's not just the CD format.

Re:double-blind, controlled test, please?

[entartete]

http://24bitfaq.org/

since i lacked mod points and didn't want to post a 'yeah, what he said!', here is a link to some info about the benefits of increasing the bit depth, which works similiarly to increasing the sample rate except for dynamic range instead of frequency, the goal isn't to record things beyond human hearing, it's to have more samples/bits to record the stuff in our hearing range with. faint overtones would only get a couple samples of a couple bits each using cd audio quality.

Re:double-blind, controlled test, please?

[S.Lemmon]

Yes - it comes out as a square wave, but you're forgetting two things. First, a "choppy" wave introduces harmonics *above* the base frequency (which is already at the limits of human hearing). Second, the DAC should also have a filter to smooth out these harmonics, turning it back into a sine wave.

Re:double-blind, controlled test, please?

[ajagci]

44 KHz sampling rate only lets you record frequencies up to 22KHz if you had a PERFECT d/a convertor and a PERFECT filter.

Yes, but you don't need to record frequencies up to 22KHz perfectly because they already don't matter. Standard CD audio already gives engineers plenty of room to play with in their filter designs.

Besides, good music is a lot like a good speech: it matters far more what is being said than the acoustic quality. And CD quality audio is already way overkill for getting every bit of meaningful musical information out of a recording.

I get the impression that there is a strong negative correlation between someone's degree of audiophilia and their musical aptitude.

Re:double-blind, controlled test, please?

[Jeff DeMaagd]

I am not an audiophile but I will note these things:

The Nyquist theory is an absolute best-case, and assumed that you sampled at the peaks.

Even with four samples per wavelength you can get pretty weird looking sample data. IIRC, EEs try to get at least eight samples per shortest wavelength to get a decent waveform representations, less than that and you can get some noticable potential frequency and phase shifting errors. On CD audio, that makes it a little over 5kHz.

Re:double-blind, controlled test, please?

[S.Lemmon]

It's not really the same thing - adding extra bits per sample (rather than more samples per second), affects volume representation, not pitch. The pitch ranges of human hearing have been well researched, but I'm not sure what the limits on volume variations are. Add to that the fact that you can turn the volume up on a quiet passage to hear more detail than you normally would (a bit like slowing playback down would let you hear ultrasonic sounds if they were there).

Re:double-blind, controlled test, please?

[Monkelectric]

Wrong wrong wrong... You're assuming the POINT of sampling at higher frequencies is to get a larger frequency response -- its not. It's to REDUCE QUANTIZATION ERRORS and NOISE, and increase DYNAMIC RANGE (the real measure of a sound card).

Quantization errors occur in the less signifigant bits, a high quality ADC will have an uncerainty of about + or - 4 bits. Think of a 10khz signal on the edge of human hearing like a nice china boy cymbal -- a cycle of a 10khz audio signal will be represented by about 4.41 samples :) I know the nyqist limit/shannons theorom says thats enough, but out here in the real world where there's noise and quantization errors its not enough, which leads me to my next point **the nyquist limit is valid only for situations where there is no noise** in other words: THERE IS NO SITUATION FOR WHICH THE NYQUIST LIMIT IS VALID. The Nyquist limit is at best, a guideline.

So now the reason you need higher resolution/bigger samples is because that alters the noise floor. + or - 4 bits in a 24 bit recording is alot less signifigant then + or - 4 bits in a 16 bit recording. Also, imagine at 192khz your 10khz signal is now represented by 19.2 samples -- error and noise is MUCH less destructive with more samples.

I deal with these issues every day in my studio, and the rule with audio is pretty much always, more is better. However, There is a point of diminishing returns -- and IMHO I think that point is 24bit/96khz. It is very difficult to distinguish a 96khz signal from a 192khz signal.

Re:double-blind, controlled test, please?

[Steve Franklin]

And why again, beyond playing computer video games, do I need this on my computer? After my first experience with XP SP1 killing my onboard DVD player, I have decided to put my extra AV cash into my non-computerized, non-windowized, non-BillyGatesized, non-rebootized, instant-on audio/video system. Other than running my computer sound through my stereo, the farther away my LCD TV and audio systems are from my computer, the happier I will be. And the fewer profits Mr. Gates receives from my near-term upgrades, the more ecstactic I will be.

Some of you guys need to wake up from your computer-induced hypnotic states and forget this convergence nonsense. It's all a big mind-freak to let Billyboy take control of an area he hasn't even the intelligence to understand.

Sorry, I usually try not to be this blunt, but there's a point where geekdom sometimes loses track of the bigger techno-picture.

Re:double-blind, controlled test, please?

[sir_cello]

The 44 kHz (IIRC) sampling frequency of a CD means that you can actually record signals with frequencies as high has 22 kHz (half the sampling frequency -- that's a methematical theorem about the discrete Fourier transform).

You mean the Nyquist sampling theorem: it applies as a general principle to the sampling of any signal, you need to sample at twice the rate to be able to provide an accurate reproduction.

Re:double-blind, controlled test, please?

[xorbe]

Duh. If the sample rate is 44.1KHz, obviously there is 11KHz content. All the way up to 22.05KHz, and then you need to filter to stop aliasing. But obviously one desires a smooth roll-off, so the filter point often starts at 16/18/20 KHz.

Re:double-blind, controlled test, please?

[be-fan]

EE is simply a field of applied Physics. An EE is an order of magnitude lower on the food chain than a Physics professor.
---------
That's a retarded comment. That's like saying that a physicist is a better authority on chemical processes, because chemistry is just a byproduct of quantum physics. The EE spends a whole lot more time studying the material concerning filters and audio signals, and its not a strech to believe that he'd be more informed about the subject than a physicist, unless the latter happened to specialize in electric signals.

Re:double-blind, controlled test, please?

[John Hurliman]

1) Lots of people listen to mp3s from their computer, which might be hooked up to a high fidelity audio system or a receiver with SPDIF-in. In either of those situations using an AC97 chipset is hurting your audio quality severely.

2) People who do video editing on their computers who would rather put their money in video equipment than a hifi soundcard will welcome the addition of quality audio the next time they are dubbing the video.

That's in addition to the big market you already mentioned, gamers. I think people who mess around with audio will still go for solutions that use offboard DAC/ADC solutions, 5ms ASIO and have MIDI controllers. You can't please everyone but I'd say this is an important step forward.

Re:double-blind, controlled test, please?

[jorlando]

There was a simmilar testing in the 80's from some audiophile magazine.

It was the time of the boom for CD players, with praces ranging from US$100 and above US$1000

The question: since the CD has a standard wich by itself guarantees a specific bandwidth, a very low harmonic distortion, wow-and-flutter almost immesurable, etc... does make a difference buying a cheap or a expensive player?

From a technical point of view, no, audiophiles said yes... the "better electronics" (whatever it was) would improve the sound.

A test rig was made, many audiophiles invited, double blind, yadda-yadda, the same person hearing the same music from different players votted the quality of sound.

result? whatever you had a US$100 player or a US$1000, the sound was the same... the standard was giving the quality...

maybe today we could have some difference due the very cheap players that are being made, maybe some manufacturers are cutting costs in the analog outputs, dropping signal levels, etc... but I can't be sure, since the components also evolved and are cheaper today and can give better results...

Re:double-blind, controlled test, please?

[Hoser McMoose]

Err, we may not know "everything" about sound, but it's been a VERY well understood concept for quite some time. As compared to basically every other aspect of physics it's extremely well understood. Sure, there are plenty of tricky issues, particularly things like the accoustics of a room, but even those are concepts that humans have been perfecting for 300+ years.

Sure, we'll learn a few new things here and there, but the chances of any sort of really fundamental change in our understanding of sound happening in the next 100 years is pretty darn slim.

Re:double-blind, controlled test, please?

[strange_attract0r]

You are somewhat missing the point. There are more reasons than just raising the maximum frequency for sampling at higher than the Nyquist rate. For a start, errors caused by jitter and misreads are significantly (audibly reduced) - you can even get CD players that oversample normal CDs which decrease error distortion.

CDs are sampled at 44.1kHz. This means, theoretically, frequencies up to 22 kHz could be recorded, in practice the amount of precision required to record this is too difficult. So a higher sampling frequency gives more headroom before the theoretical highest frequency is reached.

Cables do make a difference. Poor cable attenuates the signal more at high frequencies. You need as good a conducting path as possible.

But you are right, in that that person was deluding themself - I don't think you would notice the difference with SACD unless you had just been listening to the same thing on CD. CD is pretty good.

Re:double-blind, controlled test, please?

[sr180]

The digital filtering in Cooledit is not perfect, and it will alter the audio that is passed through it as well as allowing small amounts of frequencies greater than the limit to pass. So unfortunately your test will not prove much.

Re:double-blind, controlled test, please?

[nathanh]

Having 96 kHz sampling isn't about recording pitches above 22 kHz; it's about getting a better, smoother approximation of the sound waves. With 44 kHz sampling, a 22kHz sound wave is very discrete and choppy. Good ears would almost certainly detect this.

No, this is simply wrong. It's hard to explain without the mathematics but basically the "square edges" you see will be completely removed by the obligatory low-pass filter after the DA convertor.

Nyquist's theorem proves (with mathematics and I've done the derivation myself) that sampling at (slightly more than) 2x the highest frequency can EXACTLY reproduce the original wave. Not approximate. Exact.

That's all assuming that you have infinite bit resolution. Of course, in practise you don't have infinite bit resolution, so that's (one reason) why SACD and DVDA sample much higher than the Nyquist rate. You can trade bit resolution for higher sampling rates because they are equivalent. But once again, you need to grok the mathematics to understand why.

Re:double-blind, controlled test, please?

[SmittyTheBold]

This has been explained, but I think maybe I can do better.

Say you have a 44 kHz carrier reproducing a 22 kHz signal. That's all well and good - if the 44 kHz happens to be in sync with the crests and troughs of the 22 kHz wave. What happens when the sampling period is more in sync with the points in teh wave where amplitude nears zero? You will be sampling near silence every time, though much goes on between samples. Now, what happens if you have a signal near the magigal 22 kHz, but not exactly at? Say, 21 kHz or so. You have a signal that drifts in and out of alignment with the sampling frequency, giving the illusion of volume rapidly increasing and decreasing.

That's why much higher sampling rates are worthwhile.

Re:double-blind, controlled test, please?

[Steve Franklin]

You can't burn the mp3s to a CD and play them on your stereo/5.1 system? Even the Niro plays mp3-encoded CDs. Personally, I don't have any mp3s. It's just not worth the hassle, either technically or, now, legally.

People who do serious video editing use really expensive professional equipment. The guys who try to do it on their PCs are mostly amateurs who are just noodling around and are the same subset of computer users who complain that Photoshop is too expensive and therefore use cracked copies.

As for games, I would seriously suspect that as soon as HD LCD TVs come into wide distribution the computer-based game will die a lingering death in favor of the consoles.

As a piece of hobbyist technology, yes, it would be nice if the main board had decent built-in audio. But as a stepping stone toward integration and convergence of audio/video and computers, it is my current personal opinion that putting a few more chips inside the receiver makes a whole lot more sense than trying to turn a PC into something it never was and never will be.

As far as I can see, the only conceivable reason for this process is a social one. There are a lot of guys out there who would rather sit in their den/workshop/computer-room and consume their media there than to come out into the living room and watch a movie with their wives. And Intel ain't gonna solve that problem anytime soon.

Re:double-blind, controlled test, please?

[Listen Up]

What is retarded is your belief that you, as an EE, are of a higher order of magnitude than that of a Ph.D. professor. I will state this now and at the end of my post...You took my comment out of context for your own benefit.

Without Physics and Mathematics, what is EE? Nothing. The same goes for Chemistry. They are both applied Physics fields.

To repeat the original statement, EE is nothing more than an applied field of Physics. That is a simple, factual statement. The parent poster's comment of 'typical of a physics professor' only goes to show his total ignorance.

Engineering is a fantastic applied field, don't get me wrong. The world would be a much different place if applied physics fields did not exist.

And as a side note, your comment is the one which bears the closest resemblance to being 'retarded'.

Have a nice life. And try to learn not to take a person's comments out of context, as you took my comment out of context for your own benefit.

Re:double-blind, controlled test, please?

[Listen Up]

Your comment just proves your ignorance.

Typical, ignorant moron.

Re:double-blind, controlled test, please?

[Jeff DeMaagd]

do you care to elaborate?

Re:double-blind, controlled test, please?

[JebusIsLord]

again, no kidding. we're talking past each other. the original poster just said that he heard CDs only had 11khz of frequency response, and i showed him how to test that it is not so. That's all. Nothing else to see here.

Re:double-blind, controlled test, please?

[mibus]

What people fail to remember with all these remastered CD re-releases, is all the equipment used for the remastering process is better - it's not just the CD format.

And I can personally testify to that... my family has bought a number of SA-CDs in the past months, sometimes of albums we already owned on CD. Even on a regular CD player, they sound absolutely awesome in comparison. The SA- part makes it sound better again in a proper player, of course :-).

Then again, how many people have stereos that can really benefit from this? I do, but my computer rarely goes near it. IMHO the biggest advantage will be for gamers with satellite systems... not for the reproduction, just for the number-of-supported-speakers. Higher supported sampling rates are probably just checkboxes "we must have these or product [X] will be considered superior".

I just hope these babies work better than the AC97 based system on my motherboard :-)

Re:double-blind, controlled test, please?

[cheekyboy]

Yes, but its useless just hearing ONE tone, if you have a complex song that has say 150 intruments, then you loose fidelity per instrument at 44khz, which it all gets mixed, try mixing 72000 wave forms into a 44khz wave, you loose a lot, now do it at 192khz at 32bit and you preserve more waves.

Better yet, take some LSD and you'll hear a LOT more detail both in bitresolution and in bandwidth.

Even if someone 'cant hear' the difference between 44khz and 96khz, there are probably subconcious differences, or they probably arent trained enough to 'notice' the difference even tho they can technically hear it, maybe their memory of the fine detail is poor.

Now even if you ears are crap and can only hear 11000hz, 44khz isnt a waste , or 96khz isnt a waste, as I said, sure ONE wave wont make a diff, but mix 175 diff freq waves and you get more harmonics which get distorted at 44khz.

Surely as a physics teacher you would have known that....

Re:double-blind, controlled test, please?

[cheekyboy]

I bought some high end sony headphones, pitty the cable that was attached to them was as cheap as a $5 headphone, since after a while of bending the wires break and you loose connection, so much for 'high quality sony' they are loosing it, now their connector adaptors too are 'custom plugs' too.

And fixint it your self is so fidly with those tiny wires in the cable to splice/resolder to the plug.

Idiots.

Re:double-blind, controlled test, please?

[Jeff DeMaagd]

Nyquist's theorem proves (with mathematics and I've done the derivation myself) that sampling at (slightly more than) 2x the highest frequency can EXACTLY reproduce the original wave. Not approximate. Exact.

This assumes that the samples are at the peaks, if the sample 180 degrees out of phase of the signal, sampling the valleys, then you have no signal. Not getting a singal when one should expect one is a pretty large error to me.

If an input sine wave is near 1/3rd the sampling rate, you can easily get a bunch of nasty phase and magnitude modulations as part of your output signal. I suppose with spline interpolation a lot of that is fixable with interpolations but that may still result in errors.

Re:double-blind, controlled test, please?

[SuperJames_74]

The 44 kHz (IIRC) sampling frequency of a CD means that you can actually record signals with frequencies as high has 22 kHz (half the sampling frequency -- that's a methematical theorem about the discrete Fourier transform)

Actually, 44.1k allows you to capture up to 20kHz, not 22kHz. What you're talking about is the Nyquist Theorem, and you've got the right idea, but you're leaving out one subtlety.

A low-pass filter is used prior to sampling the audio, and this filter is set to have a ceiling of 20kHz. However, the cuttoff at 20k is not a "brick wall", it's a slope. By the time the audio is completely cut off, it's up around 22.05k, which, x 2, is 44.1k. Geddit? I guess, technically, some sound above 20k gets through, but the highest frequency fully reproduced is 20k.

Re:double-blind, controlled test, please?

[Brandon30X]

Im just second'ing what the parent post said. This is true, and the math does work out. I found in my college days that signal processing was very difficult (both mathematically and conceptually) to understand.

Re:double-blind, controlled test, please?

[be-fan]

What is retarded is your belief that you, as an EE, are of a higher order of magnitude than that of a Ph.D. professor.
--------
I'm not an EE, nor am I the original poster.

I will state this now and at the end of my post...You took my comment out of context for your own benefit.
-------
No I didn't.

Without Physics and Mathematics, what is EE? Nothing. The same goes for Chemistry. They are both applied Physics fields.
---------
While true, that doesn't mean that a physicist necessarily knows jack-shit about chemistry or electrical engineering. We were talking about audio signals, and I am more inclined to take the word of an EE than a physicist, regardless of the taxonomy of the fields. The EE works directly with this sort of thing, while the physicist only understands it indirectly, unless he is a specialist.

Re:double-blind, controlled test, please?

[nathanh]

This assumes that the samples are at the peaks, if the sample 180 degrees out of phase of the signal, sampling the valleys, then you have no signal.

Wrong. 180 degrees will indeed sample at the valleys instead of the peaks but the magnitude is the same, only the sign is different. Perhaps you meant 90 degrees.

If an input sine wave is near 1/3rd the sampling rate, you can easily get a bunch of nasty phase and magnitude modulations as part of your output signal.

Nope. Still wrong.

For your own education, here is Nyquist's Theorem.

Nyquist's theorem: A theorem, developed by H. Nyquist, which states that an analog signal waveform may be uniquely reconstructed, without error, from samples taken at equal time intervals. The sampling rate must be equal to, or greater than, twice the highest frequency component in the analog signal.

Notice the language "without error". There is no error. It's hard to grok, and impossible to believe without doing the maths, but it is 100% true.

Though as I said before, the real world is more fun because sampling is never exact. Errors in the times when samples are taken and errors in the magnitudes of the samples will screw you.

Re:double-blind, controlled test, please?

[Jeff DeMaagd]

If you would explain, I would be happy to read it.

I realize I did mistate it a little bit, but I don't see where my thought was wrong and if you aren't willing to correct me, then I can't learn.

Attacking the least relevant part of my post really looks like a strawman argument anyway.

Re:double-blind, controlled test, please?

[Anonymous Coward]

Funny thing was during a UIUC physics lecture about 5 years ago the instructor took his test system into the auditorium and cranked it up to demonstrate the range of human hearing.

A number of us, myself included, could hear up to 25 KHz on it. Either his system wasn't "in tune" or human hearing is alot better than researchers give it credit for. The professor didn't have anything to say during the class when we protested we could hear frequencies that high.

Anyway, 44.1 KHz sampling does not mean you can get 22 KHz audio out of a CD. In practice there is a transition band right around 17-18 KHz. So in reality you must filter out frequencies around that point to avoid aliasing. So 44 KHz audio definitely does not cover human hearing up to 20 KHz accurately. It also isn't perfect near the bass end of the spectrum.

48 KHz comes closer, but 96 KHz looks alot better when you analyze the frequency response on a scope. And if I remember my DSP class correctly an even higher sampling rate lets you get away with simpler DACs.

Conclusion? Bring on the high sampling rate!

Re:double-blind, controlled test, please?

[nathanh]

If you would explain, I would be happy to read it.

I'm not entirely sure I could explain. I can do the mathematics but putting it into layman's terms is beyond me. My textbooks couldn't state the idea in layman's terms, and those books were written by people who really know the material.

Attacking the least relevant part of my post really looks like a strawman argument anyway.

That's not what a strawman argument is.

Re:double-blind, controlled test, please?

[ja]

> you were told wrong.

Perhaps the poster wasn't told wrong, just didn't remember it completely right :-)

Think about it. If you record a 22Khz sine and get a representation with two nonzero samples for each cycle, then what will you get if you record a cosine instead? Digital clearity? Or "the sound of two zero-crossings clapping?"

If you had used 4x samplerate instead (as the poster suggested) then the phase of the signals individual parts would no longer be such an issue, since there would always be at least two "buckets" ready to catch the nonzero part of the signal.

cheers // Jens M Andreasen

Re:double-blind, controlled test, please?

[S.Lemmon]

Who would want to convert all their MP3s back to plain CDs? Who wants to change CDs every 15-20 minutes? A big reason people convert their library to MP3 to begin with is so they can have it all available at once on their HD.

Far from seeing the computer die out, it's becoming an increasingly common part of people's entertainment systems. That's something companies like Microsoft are banking on by the way.

However, that said, audio quality is a red herring. Few people really care beyond a certain point - otherwise no one would be listening to MP3s at all. Look at how many CD are released these days with the sound compressed to the point of clipping just to make it "louder". It's a complete disaster for audio quality, yet if it didn't sell CDs, the industry wouldn't be doing it.

It's all about convince - remember when industry pundits used to wine about "who would need the audio quality of a CD just to listen to Twisted Sister"? They missed the point - CDs didn't become popular because they sounded better. They became popular because they were easier to use and didn't wear out quickly. This new audio standard is all well and fine, but it's mainly just a game of "my specs are better than your specs - nya, nya, nya, nya"

Re:double-blind, controlled test, please?

[theLOUDroom]

I bought some high end sony headphones, pitty the cable that was attached to them was as cheap as a $5 headphone, since after a while of bending the wires break and you loose connection, so much for 'high quality sony' they are loosing it, now their connector adaptors too are 'custom plugs' too.

Get some Sennheisers. The cord on my HD495's is kevlar reinforced.

Re:double-blind, controlled test, please?

[Steve Franklin]

"Far from seeing the computer die out, it's becoming an increasingly common part of people's entertainment systems."

I didn't say computers would die out. I said computer GAMES would die out.

"A big reason people convert their library to MP3 to begin with is so they can have it all available at once on their HD."

At once? Like my DVD movie (see above) was available "at once"? After an hour and a half of dicking with the drivers? Is that Bill's vision of the future? Nothing works until you reboot the damned computer? Washing machine? Refrigerator? Heart-lung machine? Automobile controls? Hit the brake and you get the message, if you're still alive to read it, "Please release the brake, reboot, and try again."

Not me buddy. And not a lot of other people either.

"CDs didn't become popular because they sounded better."

CDs became popular precisely because there was no needle scratching and no tape hiss. I still have LPs. The sound quality is a bit better but they make this godawful sound between tracks and on quiet passages. As for wearing out, this is also a function of noise. The older they get the noisier they get.

"Who would want to convert all their MP3s back to plain CDs?"

Come now, how many MP3s on how many computers do you think came off of a CD owned by that person? Give me a brake, fellah. I'm not a total idiot.

Re:double-blind, controlled test, please?

[Jeff DeMaagd]

I'm still not finding anything that shows I am wrong.

The Nyquist theorem doesn't state that reproduction is perfect at 2x sampling rate of the highest frequency, it only states that it is the minimum sampling rate needs to be 2x the highest frequency in order to capture the signal. Higher sampling rates do capture more accurate detail particularly on the phase accuracy.

http://members.aol.com/ajaynejr/nyquist.htm

In fact, if sampling a 1kHz sine wave at 2kHz, one has to sample at the peaks or get a magnitude error. If the samples happen at the zero crossing (out of phase sample), one gets no signal samples at all:

http://www.cs.ust.hk/faculty/layers/comp342/wave fo rms/nyquist.html

Re:double-blind, controlled test, please?

[S.Lemmon]

I didn't say computers would die out. I said computer GAMES would die out.

Yes, those record sales are really killing the industry.

After an hour and a half of dicking with the drivers? Is that Bill's vision of the future? Nothing works until you reboot the damned computer?

Yep, better get used to it. As more and more things include a HD, more and more things will need to be rebooted and patched and so on. Otherwise where will they update the DRM once it's been cracked?

CDs became popular precisely because there was no needle scratching and no tape hiss

I really doubt that. Most of the people I knew back then even with the most crappy sounding of K-mart systems wanted a cd. You could go right to the song you wanted with no needle to futz around with or tape to get eaten.

Come now, how many MP3s on how many computers do you think came off of a CD owned by that person? Give me a brake, fellah. I'm not a total idiot.

Ok, I give that to you - you're only 95% an idiot. Where do you think all those ripped songs came from - God in heaven? Plenty of people rip their onw CDs so they can have quick access to them or copy em to a portable player. Don't paint everyone with your own suspect morals.

Re:double-blind, controlled test, please?

[nathanh]

Higher sampling rates do capture more accurate detail particularly on the phase accuracy.

No, you get EXACT reproduction without having to use higher sampling rates.

In fact, if sampling a 1kHz sine wave at 2kHz, one has to sample at the peaks or get a magnitude error. If the samples happen at the zero crossing (out of phase sample), one gets no signal samples at all:

That's because you mistakenly think Nyquist's theorem is Fn = 2Fmax. Nyquist's theorem is Fn > 2Fmax. So what you're seeing is aliasing when Fn = 2Fmax. This causes an attenuation in the amplitude proportional to cosine of the phase difference between the sampling frequency and the signal. If you have Fn < 2Fmax then you get a "beating volume" effect as the phase difference shifts over time.

Don't get all excited. You haven't proven Nyquist wrong. You just didn't understand what Nyquist said.

Re:double-blind, controlled test, please?

[evilviper]

After my first experience with XP SP1 killing my onboard DVD player

How exactly did software kill your hardware. Not to say that it's impossible, but very unlikely, and owning a crappy DVD-ROM is a much much more likely explanation.

I have decided to put my extra AV cash into my non-computerized, non-windowized, non-BillyGatesized, non-rebootized, instant-on audio/video system.

My AV system is a single PC. It is not running Windows, for all intents and purposes, it never needs to reboot, it has better quality than even expensive DVD-players, it's cheaper, more flexible, etc.

And finally, although it technically isn't instant-on, it is up and playing my DVD before my DVD player is. You see my DVD player powers-on for about 1 second before it does anything. After that, it spends 3 seconds or so trying to read the disc that's (not) in the drive before it will let me open the tray... wait... insert disc... close tray... and wait. My computer boots up in about 10 second, it is slot-loading, so no wait, and you can insert the disc within a half second of powering-on. My computer reads the dics faster, renders the DVD faster, seeks many many times quicker, etc.

And most important of all, my computer doesn't give me any crap about forcing me to watch track-0 advertisements and FBI warnings. That alone makes it far faster than a standalone.

Some of you guys need to wake up from your computer-induced hypnotic states and forget this convergence nonsense. It's all a big mind-freak to let Billyboy take control of an area he hasn't even the intelligence to understand.

No, actually it's a great thing when done correctly. Tell a Tivo owner that computers suck as multimedia devices, and you'll be picking your eyes up from off the floor.

I love it, because for the first time, the public can decide exactly what functionality they want in their devices, instead of allowing electronics manufacturers to dictate what you get.

My $250 computer can output to standard TV, a computer monitor, or to an HDTV. It can play Ogg, FLAC, MPC, etc. It can play MPEG1/2/4, it can play WMV, or it can play Theora videos. It can playback HDTV content with no problems, to a HDTV, Computer Monitor, a widescreen projector, or a standard NTSC TV (although I will need to buy an inexpensive HDTV tuner for serious HDTV use). It can downmix 5.1 channels to any arbitrary number of channels I choose, or play a different Ogg file on each of the 6 speakers if I tell it to do so. It will record PAL or NTSC TV from cable or air. It can automatically edit out the commercials or I can manually do that. I can record any content to VCDs, SVCDs, DVDs, or just stick them on a hard drive.

The point, as I already said, is that I really have control, and it will do whatever I want. I am not beholden to electronics or media companies. It can play HDTV movies from the internet, it can play VCDs from a DVD. It can play DVDs stored on a hard drive. It will do exactly what I tell it to do, and will have an interface that I want it to have. I'm a much happier person for having done it.

Now, just try and point-out a reason that what I am doing is somehow bad, wrong, or limiting.

Re:double-blind, controlled test, please?

[walt-sjc]

Dude. Chill. Many of the innovations that make multi-media work better on a PC can migrate towards consumer A/V. Case in point: Tivo. Sometimes it takes a geek to SEE the bigger picture.

Re:double-blind, controlled test, please?

[ja]

> Fourier states that a X Hz signal requires GREATER THAN 2X

Uhm ... How much greater? 4X is greater than 2X >;->

cheers // Jens M Andreasen

Re:double-blind, controlled test, please?

[theLOUDroom]

What is retarded is your belief that you, as an EE, are of a higher order of magnitude than that of a Ph.D. professor.

"What is retarded is your belief" that I think EE is somehow above physics. EE wouldn't exist without the hard work of a lot of physicists. Now let me explain why I said he was "definately a physics teacher, not an EE".

Scientists deal with theory, engineers deal with application.

It's that simple. His post showed that he had never actually tried to implement A/D conversion. I wasn't saying engineers > scientists, I was saying the engineers have a different knowedge set than scientists.
And I don't see why you had to throw the whole PHD thing in there. There are PHD physicists and there are PHD EE's. In neither case does having a PHD mean that you are better/smarter than everyone else.

Re:double-blind, controlled test, please?

[theLOUDroom]

The (still digital) signal is now a "smooth line" through the supplied data points at 44 kHz. This signal is converted to a voltage by the true (non-signal-processing part of the) DAC. The part of the spectrum below 20 kHz will be exactly the same independent on whether the original input to the DAC was 44, 96, or 192 kHz. (Note: 1-bit DA convertors use a slightly different approach, but with the same result).

What you seem to have forgotten is that there is still another filter after the D/A convertor.

Without any filtering the analog for a 5KHz sinewave output would look like a staircase. (Resulting in higher frequency noise.) In order to be able to record and playback matching 5KHz sinewaves, you need filters at both the A/D side and the D/A side.
If you downsample the signal to 44KHZ, your filter requirements on the playback side increase. I'll grant that people don't worry about this so much because speakers act as low pass filters themselves, but the effect on say a 1KHz sinwave, should be detectable.

Note that 92 dB is quite loud: about 4 W power to a typical 87 dB/W loudspeaker at 1 m distance.

92 dB isn't very loud at all. And pretty much anyone who cares about what sampling frequency they're using is going to have more than a 4 watt amplifier. The audio system you just described is that of a $40 boombox, and a human can hear much more than 92 dB of dynamic range. While this may not be a big deal for pop music, classical tends to vary much more in amplitude throught to course of a single piece. (FFF => mF => pp, etc.)

If you want to blast Wagner through my system, with a 175 W amplifier and 96 dB @ 1 W 1M speakers, 24 bits is a much more sensible number.

Now if you want to do some DSP on the playback side (Winamp or XMMS's built-in EQ for example) , 32 bit sounds good because all those rounding errors in your math remain inaudible.

Re:double-blind, controlled test, please?

[hankwang]

What you seem to have forgotten is that there is still another filter after the D/A convertor.

Yes, a soft analog filter (pass for f<20 kHz, stop for f>88 kHz). That's easy enough to construct without introducing significant phase shifts for f<20 kHz. There is no need to fully filter away the range 20-88 kHz, since there is no spectral content in that region (the ADC operates on a 176 kHz PCM stream (4x oversampling)). Moreover, the filter has the same characteristics as the filter that filtered the signal from the microphone during the recording, so if you don't like this filter, then you won't like the filter during recording either. If you want to be safe, use a better DAC that does 16x oversampling, such that the PCM stream becomes 705 kHz and the filter needs to cut off at 352 kHz instead of 88 kHz.

In principle, it is possible to pre-compensate in the DSP for the very small ripples in the analog filter spectrum close to 20 kHz, though I don't think that that is commonly implemented. (DACs are usually not manufactured by the same companies as those who build the filters in the CD players)

92 dB isn't very loud at all.

Well, my comparison with speaker volumes was maybe not so lucky with the current amplifier rating inflation, but let's compare it to typical sound sources (google for decibel levels, all numbers relative to hearing treshold): Large Orchestra (98 dB SPL), heavy traffic (85 dB), tractor (90 dB). Also: Safety limits tell you that 92 dB is permissible for maximum 6 hours per day; moreover. You're probably safe if you listen to music which is mostly soft with short bursts.

Suppose that your Wagner/Mahler music has a typical -20 dB (relative to the maximum that's possible on a CD) level. Then the quantization noise level is at 72 dB below listening level, but you playback at 90 dB, so the absolute level of the quantization noise is at 18 dB SPL. That's comparable to whispering at 5 feet distance. I find it hard to believe that one can hear someone whispering at 5 feet while standing next to a tractor or a passing truck.

However, it is defendable, if you really want to stay on the safe side, to encode in 24 bits, in order to capture the entire range between hearing treshold and instant hearing damage.

I still don't think that it makes sense to use 196 kHz as a sample rate for a transport medium. If you want more realistic sound, then spend the space on surround encoding.

Re:double-blind, controlled test, please?

[hankwang]

One more thing that maybe wasn't clear:

In order to be able to record and playback matching 5KHz sinewaves, you need filters at both the A/D side and the D/A side. If you downsample the signal to 44KHZ, your filter requirements on the playback side increase.

The hard filtering work (cutting off everything beyond 22 kHz and not touching anything below 20 kHz) on the playback side is done by a digital filter on a 176 (4k oversampling) to 705 kHz (16x oversampling) PCM stream. A digital filter can easily satisfy the high requirements that are needed. The imperfect analog filter that follows after the DAC needs to satisfy only very modest requirements.

Re:double-blind, controlled test, please?

[theLOUDroom]

Yeah, I pretty much agree with everything you were saying, but I do have to tak issue with this:

I find it hard to believe that one can hear someone whispering at 5 feet while standing next to a tractor or a passing truck.

The problem with this comparison is that it both events are always happening at the same time. Unless you're going to manually adjust your volume knob for the ppp and FFF sections of the piece, then that 92dB is all the dynamic range you get for an entire symphony. It's already been established that loud sounds have a "masking effect" both in time and frequency, this is what things like mp3 enconding capitalize on. This issue is that this masking effect doesn't last the length of a 74 minute CD.

I still don't think that it makes sense to use 196 kHz as a sample rate for a transport medium. If you want more realistic sound, then spend the space on surround encoding.

I'll agree with this. I think 24 bit/96 KHz, is most likely the limit of human hearing.

My main point is just that there is room for improvement over standard 16 bit/44 KHz sampling.

Re:double-blind, controlled test, please?

[theLOUDroom]

The hard filtering work (cutting off everything beyond 22 kHz and not touching anything below 20 kHz) on the playback side is done by a digital filter on a 176 (4k oversampling) to 705 kHz (16x oversampling) PCM stream. A digital filter can easily satisfy the high requirements that are needed. The imperfect analog filter that follows after the DAC needs to satisfy only very modest requirements.

Thanks for clearing this up.

You are oversampling by 4-16 and keeping 16 bits of resolution? Then you're digitally filtering it and running it out a D/A convertor that can handle 176-705KHz?

Re:double-blind, controlled test, please?

[hankwang]

You are oversampling by 4-16 and keeping 16 bits of resolution? Then you're digitally filtering it and running it out a D/A convertor that can handle 176-705KHz?

Yep. The filter calculations are carried out in more bits in order to prevent roundoff errors, but the output is in 16 or maybe 18 bits resolution at this high sample rate. Google for "dac oversampling bits spectrum", there are plenty of web pages that explain things, e.g. oversampling filters and noise spectra. (I have seen better ones, but I'm not sure what exactly I googled for last time)

It is next to impossible to make a DAC that can create a voltage with more than 18 bits resolution because of the tight tolerances that are required. In a 1-bit DAC, the voltage-resolution problem is translated into a kind of duty-cycle modulation with just 1 bit of resolution. Those DACs oversample with 64x or more and run internally at 3 MHz. It's amusing when people complain that sound cards have too much RF interference with the mother board while the DAC itself is running at 3 MHz together with a DSP processor (pentium-100 performance) integrated on the same chip. This DAC is designed to generate noise in the 200 kHz+ region with a 5 V amplitude... (Cheap-sound-card noise comes from a noisy power supply and a cheap buffer amplifier between the filter and the connector, not from RF interference)

Best soundcard?

[Doc Ruby]

I play all my music from WAVs on my HD, but I don't sacrifice quality for money. The highest-quality DAE from CD to HD (using CDParanoia) gives the same quality as thousands of dollars worth of separate CD transport and data equipment. Then I (losslessly) compress them with Shorten (2:1) to save some money on storage. I often bypass my Onkyo amplifier and KLH speakers to listen with my Sennheiser 600 headphones - all hi-end audio gear. But the bottleneck is the soundcard. Soundblaster Audigy 2 seems really good at $80, but doesn't it have noise from the PC power supply? What's the best way to get all my CD quality from my Debian/i386 HD to my 5.1 surround system, playback only (no ADC)? Let's say my budget is $500 for the "soundcard", which one is the one for me?

Re:Best soundcard?

[0x0d0a]

You have *all playthrough inputs on your sound card muted* and you're still getting audible noise?

I have an emu10k1-based card. Turns out that the OSS/Free drivers (one of the four free drivers available for this card -- there's the OSS/Free drivers, the native kernel driver, Creative's driver (which may be an adaptation of the native kernel driver, not sure), and ALSA. I started using OSS/Free, since Red Hat had defaulted to using OSS/Free with my old card. I kept getting noise -- sort of a buzzing -- when doing things that caused major busses to switch on and off slowly -- moving mice, dragging windows around. I checked my mixer, and all the inputs were indeed muted. It turns out that apparently, the OSS/Free drivers didn't *list* all of the available on-board inputs. There were some still on-board inputs that still were on that I could see after switching to ALSA. I muted the additional inputs with ALSA, and the buzzing went away.

I'm not sure whether there would be any really audible difference, but Creative sells a relatively cheap (compared to professional equipment) external DAC called the Extigy that might be what you want.

Re:Best soundcard?

[afidel]

Basically you have three catagories of solutions that will work for you:
1)Digital out to an external amp
2)PCI card with an external breakout box
3)USB/Firewire sound cards

With solution 1 you can spend as much as you want on the external amp and keep your current sound card. With solution 2 you will need a new soundcard and will need to worry about drivers as not all the cards in the prosumer/pro range have Linux support. With 3 you will probably spend the least but you will REALLY have to be carefull of Linux driver support.

My personal choice would be an M-Audio Delta series with external breakout box. Pretty well supported under Linux and great sound. Though it might be a bit overkill for just output.

Re:Best soundcard?

[haffy]

In order to get proper quality PC audio, you need an external D/A converter and a sound card with digital output.

Personally, I use a high end audio D/A converter and a cheap sound card which supports 44.1kHz digital output.

Most soundcards with digital out always resample to 48kHz on the digital output and thus distorts the sound, which naturally is 44.1kHz coming from the CD in the first place.

memory requirements

[Saville]

Since you can fit ~80minutes of music on a ~700meg CD you have ~146K/sec for your music. That is at 16bit 44.1KHz stereo songs. Now audio data will take 8.7 times as much memory if recorded in stereo, but if recorded with eight (7.1) channels each song will take almost 35x as much memory thanks to the higher sampling rate and the use of 32bit values instead of 16bit. That is 5.08 megs/sec for your audio.

I like that this standard is very future proof, but when can we use it? Already CD sound is good enough for all but maybe 10,000 people on the planet. Most people's audio experience is probaby limited by their audio hardware, not the source sound. Hey, most people are quite happy encoding their mp3s at 128k!

Where will the high quality sound data come from? Audio CDs are still going to be 16bit, stereo, 44KHz. DVDs have compressed audio. Almost all video games use compressed audio of some sort too because we don't have enough memory yet for even CD quality sound.

I love that it is 7.1 and that it is very future proof, but other than making 7.1 standard it seems to be a standard for marketing to use as an advantage, not something consumers will ever use (by the time they can use it they'll have upgraded anyway). It seems that this beyond CD quality audio is just included because they can and we'll never see it in use this decade :)

Better to overbuild than underbuild I guess. But I'm not excited about this promise of higher quality audio.

Re:memory requirements

[damiam]

Now audio data will take 8.7 times as much memory if recorded in stereo

That's what DVD-A/SACD are for.

Re:memory requirements

[Ziviyr]

Bad math, 44100 * 4 bytes is little over 172K/sec. CDs only show as 700 megs due to error cerrection implicitly added for data tracks.

Also, I don't mind cheap overcapable hardware, can turn my mind elsewhere because of it.

(I'd love if they used 6 more bits of this 32 bit video for actual color, since I have yet to see my monitor turn tronsparent...)

Re:memory requirements

[Saville]

That's what DVD-A/SACD are for.
DVD-A is up to 192kHz/24bit, not 192kHz/32bit. Still, Above about 22-24 bits, there is no point in adding bits, as the noise floor of the system is below the threshold of thermal noise and other effects. A true 24-bit converter, for example, would give a noise floor of -144 dB, and apart from the fact that you can't realistically do anything with a -144 dB noise floor, no one can make components that quiet anyway: most 24-bit converters are hard pressed to reach -120dB. So 24 bits is more than enough. So I guess DVD-A is ok, but it is only 5.1, not 7.1.

Re:memory requirements

[damiam]

Still, one could fit a bunch of (losslessly compressed) 32bit, 7.1 audio on a DVD-ROM, even if it wasn't technically DVD-A.

Re:memory requirements

[shadowbearer]

I suspect that most people's listening experience is more probably limited by their *organic* hardware, given how decent quality electronic hardware keeps dropping in price.

But that's just me being pedantic :) (and getting older, with attendant loss of hearing capability from listening to very loud music, sigh)

SB

Re:memory requirements

[evilviper]

I like that this standard is very future proof, but when can we use it?
[...]
I love that it is 7.1 and that it is very future proof,

Hmm... Through my powers of deduction, I'm getting the subtle impression that you like the general idea of this standard being future-proof.

(by the time they can use it they'll have upgraded anyway).

Are you forgetting just how long AC97 has been around? 8 years from now, I wouldn't be surprised if this chip was still used, and that it's specs begin looking a wee bit restrictive.

Already CD sound is good enough for all but maybe 10,000 people on the planet.

As I've said before, how do you know you aren't missing something with CDs when you've never heard anything better?

Besides, these specs don't have to be used in the same way you migh expect. Maybe those 7 channels will be split, so you can have 2 entirely different sounds comming out of the same speaker. Probably not going to happen that way, but just an idea.

7.1?

[Cyno01]

I had this discussion the other day with some friends, none of us are audiophiles, but we all have decent setups. I have 4ch surround for my entertainment center and 4.1 for my sterio in my bedroom, but we all understand that the 5th is a front center, and we all assume, but none of us know that 6.1 has a rear center chanel. But none of us could figure out the arrangment of 7.1 surround. Is there an overhead speaker or no front center speaker and 4 evenly spaced in front. Can anyone shed some light on this?

Re:7.1?

[Firehawke]

Well, the picture I found looks something like a modified 5.1 arrangement. You've still got the three front speakers, two back speakers, and subwoofer, but you also get two true side speakers for a total of seven. I guess this gives you a more distinct frontleft and frontright audio angle, but I doubt I could really hear the difference.

Re:7.1?

[Tyler Eaves]

I think 7.1 is something like:

.1
1 2 3
4 5
6 7

Re:7.1?

[Toxygen]

You have your front, left and right speakers as usual, and 4 surround speakers in a semi-circle behind you.

Re:7.1?

[Rufus211]

Quick google found this review that includes nice pictures.

4.1: Front Left, Right; Mid Left, Right
5.1: Front Left, Right, Center; Mid Left, Right
6.1: Front Left, Right, Center; Mid Left, Right; Back Center
7.1: Front Left, Right, Center; Mid Left, Right; Back Left, Right

I always thought the mids ended up being farther back than shown in the picture though.

Re:7.1?

[Tyler Eaves]

Argh...stupid Slashcode.

Basicaly you take a 5.1 setup and add left and right "side" speakers

Re:7.1?

[DeadMeat (TM)]

6.1 is 5.1 plus a rear center speaker, as you guessed. IIRC 7.1 is 5.1 plus two side speakers.

Re:7.1?

[sirsex]

front center, front left, front right, left, right, rear left, rear right, subwoofer. I guess 8.1 would have a rear center. >8.1 you would be sitting inside one gaint circular speaker box

Re:7.1?

[EulerX07]

Check it out at dolby.

It's basically : Left, Center, Right; SurroundX(left,rear left, rear right, right). Total overkill IMHO, 5.1 is good enough for me.

Re:7.1?

[geirt]

In the movie world, a 7.1 audio mix usually means a 5.1 surround mix plus a conventional 2 channel stereo mix. You can synthesis a conventional stereo mix from a 5.1 surround mix, but the result may vary. That is why some movies are mixed in 7.1, which really is both a 5.1 and a stereo mix.

When the movie is distributed on DVD or used in cinemas they use the 5.1. When the movie is sent on TV (eg. PAL with NICAM), you get the stereo mix.

Re:7.1?

[Jeremy Erwin]

Dolby Digital EX encodes a back surround channel in recordings by distributing in among the two rear surround signals in a Dolby Digital 5.1/5.0 mix. Thus, if the user has the correct equipment, the commonalities between those two signals are reproduced on a speaker set behind the listener.

However, if a normal Dolby Digital recording is played back on such a system in Dolby Digital EX mode, it may sound disorienting. If you use two speakers, offset from the center rear, each reproducing the same signal, it's less so.

The obvious next step was to figure out a way to send a discrete signal to each of the rear surrounds.

There are other experimental designs, which add speakers at various elevations, a second independent LFE channel and so on. The 1950s road show "standard" used 5 front channels and one rear channel. Naturally some of these setups may be wholly incompatible with others.

I prefer OSS

[MarcQuadra]

I still prefer OSS, even on my 2.6 testbox, ALSA is about two-and-a-half more bitches to set up from scratch. I really hate having to do all the module configs when OSS just seems to work.

All I really need is playback from my systems, ALSA is overkill for my needs, and I hate recompiling the alsa-drivers package every time I update my kernel (on 2.4 systems).

Hopefully someone will automate or simplify ALSA for low-end use.

Re:I prefer OSS

[0x0d0a]

Hopefully someone will automate or simplify ALSA for low-end use.

The distros that have shipped ALSA as default, like SuSE, have had pretty good dummy-proof setup of ALSA for a while. Probably every major distro will be using ALSA in 2.6, which means that the remaining OSS/Free holdouts, like Red Hat, will be doing up easy-to-use UIs for ALSA.

I also stopped using ALSA a while ago -- it was just a pain in the ass to recompile the alsa-driver package each time I upgraded the kernel, and all the software I use also supports an OSS interface (and *most* was using ALSA through the OSS compatibility interface). I expect I'll be using it again in 2.6.

Re:I prefer OSS

[hankwang]

it was just a pain in the ass to recompile the alsa-driver package each time I upgraded the kernel

After struggling for a while with the Kernel Versioning Hell (A few major releases ago, compiling a kernel was as simple as make dep; make zImage; make modules, but nowadays that doesn't seem to work without endless tweaking), I found Up-to-date Alsa RPMs for all major Linux distributions.

Re:I prefer OSS

[tloh]

Chances are slim that *New* devices/chipsets such as those based on IHDA will be supported by OSS/Free. Although most existing sound chipsets work fine with OSS/Free, continuing development seems to be non-existant as of 1998. (scroll down to disclaimer section.) That said, if ALSA is considered too much trouble for some, There is an alternative in the commercial version of OSS marketed by 4Front Technologies called OSS/Linux. Cofounded in 1995 by the guy who originally designed OSS/Free, Hannu Savolainen, 4Front has developed many more drivers for Linux through NDAs with hardware makers. More details here. Alas, no free beer from 4Front.

Definitely some fishy Marketing going on here

[codifus]

First off, 32 bit, 192 Khz, wants to appeal to those very serious about audio. 32 bit cards can have a dynamic range ratio of 144 db. That's beyond what normal humans can dfifferentiate, which is 120 db if we're lucky. Not only that, but professional 24 bit cards far exceed the needs and capabilities of most , if not every, user, with aaround 110 db of dynamic range. And they're going to put this mega high tech onboard? Hmm. 2ndly, the inclusion of Dolby. This is to appeal to the movie guys, but the real serious audio guys know that Dolby encoded audio is like an MP3, lossy compression. Serious audio guys will frown on that aspect. Incorporating these 2 aspects seems somewhat contradictory, which marketers always tend to do when trying to appeal to everyone. I, for one, remain highly skeptical. CD

Re:Definitely some fishy Marketing going on here

[theLOUDroom]

32 bit cards can have a dynamic range ratio of 144 db.

Actually, the rule of thumb is about 6 db per bit. (I could look up the real formula if you're interested.)

That means 24bit give you about 144db and 32 bits give you a would-kill-you-if-you-used-all-of-it 196 db of signal to noise ratio. The only real reason to use 32 bits instead of 24 is for speed (computers are more likely to be addressed in blocks of 32 bits) and for DSP applications, where that 10 band eq winamp gives you results in some calculation error in the bottom few bits.

Re:buy yourself an enema

[Doc Ruby]

How do you know I can walk? Are you checking out my ass, asshole Anonymous Coward?

EAX vs Dolby Pro Logic IIx

[BrookHarty]

But isn't Dolby Pro Logic IIx for creating natural surround from stereo for music/movies while EAX allows game developers to create surround sound reflections for 3d enviroments?

And Creative has breakout boxes, multiple inputs, surround emulation software, accelerated audio, EAX# and A3D compatible, support for most games, etc. (And DRM)

I don't see this killing off creative, but will hurt its marketshare from non-gamers.

On the flip side, Creative labs have been quite stale, only minor updates to its Audio card line. They have been adding many other products, they even have mini-pc's, gfx, burners, mice, keyboards, etc..
-
Secondlife

Re:EAX vs Dolby Pro Logic IIx

[kryptkpr]

Creative's drivers suck. One of the few things to ever bring down my windows 2000 machine was sbpci.sys .. bleh, I'm never buying from them again.

This is a step in the right direction

[DaveCBio]

I would have prefered discreet channels instead of 2 channel encoding.

Isn't this all for naught?

[midifarm]

I mean seriously... Professional recording studios at most record in 24-bit 192kHz. So where would this 32-bit recording come from? Hasn't most of the world been dumbed down to where MP3's sound good or at least good enough? I don't know too many people with a sound system worthy of playing anything 32-bit. Besides what is the point of it all?

The hottest selling gadget of the "music" world is the MP3 player and the seemingly hottest article of contention is the online music store. None of these are even close to being prepared for 32-bit let alone the sizes of the files necessary to create such a file.

There are a lot of comments about 6.1 and 7.1 CD's or recordings and it's all rather silly. There's no real precident of a true recording done in surround. Would you really want the lead guitar only coming from the left rear channel? The only time that I would think that it would be cool would be at a live performance, but as far as I know no one has really done anything like this.

So were looking at several GB of needless information to recreate a CD with most likely marginal musical worth, and Intel is leading the charge? I think they're looking at their dwindling x86 market share (AMD is on the upswing, not pushing my Mac-centric views out there) and trying to find a niche by using it's brand recognition. I think Dolby and DTS will have more to say as to whether this proposed solution will have any legs.

Remember most of the manufacturers and broadcasters still haven't totally agreed upon an officially acceptable HD format! DVD took too long. CD was all Sony, but took long enough for acceptance. Where is this leaving the consumer? A 32-bit 192kHz audio card in their computer, decoding 7.1 channels of information so they can play video games using samples that have been resampled from their original 16 or 8-bit formats.

I think the word is overkill and it's needless. Most people can't tell the difference and for those that can, I scoff at you. I've worked with some of the best audio engineers in the world and they wouldn't be able to hear the nuances you claim. There is "air" there in higher fidelity recordings, but most speakers can't play it back any way. Ah well, thoughts?

Peace

Re:Isn't this all for naught?

[cheekyboy]

there is a diference to playing 1 16bit 44khz sample, or playing 187 of them at once in a game, try mixing 100 wav files into one and how much you get back? Its like mixing 100 translucent images on top of each other, you loose your detail.

If you can mix 100 wavs into one and loose zero detail, that is 100:1 compression already.

Either have 32 * 44khz DA converters, or have one big ass 32bit/192khz one. You can post analogue mix the 32 44khz signals into one signal for great output, or do it in software and send it out one DA converter.

Re:Isn't this all for naught?

[midifarm]

Ah play with the latest version of PT HD or even the new MOTU 896HD. All 192kHz capable. Besides, I said the most they'd be at.

Peace

Re:Isn't this all for naught?

[Canar]

A few nitpicks:

32bit MP3s are entirely possible. However, the present psychoacoustics usually limit the effective bitdepth to around 16 anyhow, because we can't perceive much more than that. MP3s have no dependance on bitdepth; you could theoretically make a 1024-bit MP3 and have it entirely compliant; the problem at that point is that the standard MP3 frame sizes are too small to contain all the information.

CD was not all Sony. Philips played a role in there too, so much so that they own the copyright for the "Compact Disc Digital Audio" symbol that appears everywhere on audio CDs. They've even gone so far as to deny the symbol to albums that implement DRM and copy protection.

A sample-rate increase may be nice, as it would limit ringing.

3rd dimension

[owlstead]

When are they going to find out that there is a third dimension out there? When are planes going to fly over my head?

I was wondering if a 4.1 speaker system could not do it all. One straight above the monitor (somewhere along the ceiling), one on either side (a bit further away if possible) and one right behind you. And the subwoofer, well, somewhere. Now you could 'vector' any sound from anywhere.

Or is this too simplistic to get the full 3D experience? Or is a 6.1 needed for this? Audiophiles, attack!

Parent is most likely correct.

[admbws]

Though I can't say I understand how it can possibly represent a perfect sine wave without some form of anti-aliasing. What about other waves? What about more complex waveforms?

Stand by while I do some research...

Re:Parent is most likely correct.

[squid_wrangler]

What about more complex waveforms?

All complex periodic waveforms are constructed from a superposition of sine waves of different amplitudes, frequencies and phases. So, if simple sine waves can be reconstructed correctly, complex waveforms that are sums of sine waves can also be constructed correctly.

Centrino shares some similarities with WinModems

[0x0d0a]

Centrino's wireless Ethernet controller is roughly the WiFi equivalent of a WinModem. Some of the components that are traditionally done in hardware (I'd guess the same stuff as in WinModems, like the DSP work, though I don't know the exact extent of the "softwarization") are done in software. Intel is not holding back on Linux support to secretly help out Microsoft -- I agree with you there. They're just in the same position as the WinModem vendors. If they supply their product's crown jewels -- open source the software that does a lot of the heavy lifting in their hardware product -- they've funded the R&D for what will be promptly snapped up by competitors and produced more cheaply.

So, you are right that there is no plot to help out Microsoft, but the grandparent is right that Intel may be cagey about supporting a platform where users are rabid about having source (and much of the architecture works less reliably without source).

Frankly, I'm frusterated with the whole laptop situation, and I wish, wish, wish that laptop vendors would make some critical mistake in the price wars and accidently commoditize their product, with standard components and form factors, so that things can be built and swapped out a la desktops.

Resampling?

[rudib]

I just hope that these new chips won't resample everything to some native, internal processing frequency, like AC97 does to 48KHz.

Otherwise, just give me HDMI compatible soundcards/DVD(-A) players and surround receivers, and I'll be happy. =)

I'm sorry....

[UrGeek]

...i started so much confusion. But 192 kHz seems wrong when discussion sample rate. But kbps is certainly wrong. We are talking about 192,000 32-bit samples per second, so I guess it should be 192 ksps@32-bits

Protocol vs. controller

[Weaselmancer]

Don't get me wrong, AC97 is cheap, but it really dragged on the CPUs of the timeframe it came out.

Well, that's not really AC97's fault.

AC97 is really nothing more than a 5 wire signal specification. It has more to do with voltages and waveforms on wires. And a register set in the codec that the wires are talking to.

But that's the idea of AC97 - you don't need to know who made the codec, only that it's AC97. Then it's a drop in replacement, pretty much.

But controllers - everybody and their brother has a different idea how to talk to an AC97 codec. And it's the controller that determines the performance. Are you bit banging your codec? Then performance will suck. Are you using interrupts? Performance will improve. Using DMA? Performance will improve again. Does your DMA engine suck? Performance will drop.

If you're having a drag on your cpu due to audio, it isn't AC97 that's at fault. It's someone's lousy idea for a controller. AC97 is a spec, not a gadget.

Weaselmancer

The labels will botch it

[Animats]

Then we'll have the labels compress everything so that it's up near the top of the scale anyway. "Nobody wants to be the softest CD in the changer". Most popular music is compressed so hard it's badly damaged.

The main reason you need more than 16 bits is because, during soft passages, most of the high bits are zero and you may effectively have only six or four bit audio. Classical recordings that aren't compressed really do suffer from this problem.

But really, the number of people who buy classical piano recordings is small.

If the industry can agree that the reference level for popular audio is somewhere well below 100%, this could work out. But that won't happen.

Wow

[ajagci]

The Intel High Definition Audio solution will have increased bandwidth that allows for 192 kHz, 32-bit, multi-channel audio

This is so that my eight-eared mutant pet bat from outer space can finally have a full high-fidelity experience.

For regular humans, of course, CD-quality audio is already overkill.

Re:Wow

[evilviper]

For regular humans, of course, CD-quality audio is already overkill.

I'm not so sure about that. Just because nothing sounds "wrong", does that mean that it wouldn't sound better if it had better range? Maybe it sounds good because you're just used-to not hearing certain things.

I know I've heard plenty of raves from those that have heard SA-CD, and thing it sounds far better. Maybe just wishful thinking, but I'm not willing to assume that I'm hearing everything that's supposed to be there, until I've heard the one with more range and am sure I can't hear the difference.

Re:Wow

[ajagci]

Well, maybe music reproduction would be even "better" if it came with a personal, intimate massage. But that has nothing to do with music either and shouldn't be advertised as improved audio.

Most recordings these days are an entirely artificial product. If they tried to approximate the live experience, what 7.1 would give you is mainly being surrounded by coughing and seat-shifting. Is that better?

And even if you wanted to reproduce sound localization, not for music, but for games, where is the evidence that 7.1 does it well and does it cost effectively?

Re:Wow

[evilviper]

Well, maybe music reproduction would be even "better" if it came with a personal, intimate massage. But that has nothing to do with music either and shouldn't be advertised as improved audio.

What have you been smoking? I never said anything of the sort. Improving the dynamic range certainly is "improved audio". Perhaps you will be able to hear the difference, perhaps not. In any case, how in the hell could anyone claim there is no improvement when they've never even heard it?

Re:Can you recommend some computer speakers?

[KillerHamster]

I'm listening to Mussorgsky's "Pictures at an Exhibition" on my Klipsch ProMedia 2.1 speakers as I type this, and I highly recommend them. Two desktop speakers and a subwoofer; one of the two speakers has a headphones jack, auxiliary plug, main volume control and subwoofer level control. The system has loads of power but sounds great at low volume. It is THX certified. It's built pretty solidly - each desktop speaker is mounted on a metal stand and the subwoofer has a nice, heavy enclosure. Only downside I can see is the price, though it's come down a bit since I bought it. Specs are here.

Where to get them in Canada?

[MichaelCrawford]

The Klipsch 2.1 speakers do look nice.

While I can order them from the U.S., it would be better if I could order them from an eCommerice site in Canada. Do you know of any?

what causes distortion inside the PC?

[violently_ill]

what are the main offenders when it comes to corrupting audio signals inside the PC? i was under the impression that in a purely digital PC setup, only discreet bits of binary code were getting shuffled around with phenomenally low error rates, thus making the signals vastly more resistant to corruption and distortion. apparently, this is not the case.

so why does this happen and is there anything i can do to minimize distortion in my setup? also, will we see a technological solution to this problem in the near future?

Re:what causes distortion inside the PC?

[Alereon]

This is, actually, the case. Assuming a stable source, quality loss starts in the DAC, where the digital signals are converted to analog for playback by your speakers. If your DAC sucks, or is one that poorly resamples everything to 48Khz (Soundblaster Live!, for example), you've already lost the game. After the DAC, the analog signals are subject to interference all the way until they hit the speakers. On a properly configured system, any EMI or other interference should be minimal.

To get the best audio bang for your buck on a modern system, replace first the speakers, then the DAC, then the source. Get damn good headphones or speakers so you can make the best of what you have, get a great soundcard so you can feed the speakers a high fidelity signal, and make sure your CD player is playing in digital mode with error correction on.

On my system, I first purchased Sennheiser HD590 headphones. The improvement was stunning, on all sources and types of audio. I then upgraded my soundcard from a Soundblaster Live! to nVidia Soundstorm (onboard on my new motherboard), which brought a very noticeable improvement in audio quality, even to poor quality MP3s. Finally, I made sure I had the highest accuracy decoder plugins for WinAMP, and used ASIO output to pump my decoded audio directly to the sound chipset, resampling to 48Khz to placate the AC'97 DAC. This brought a smaller improvement in quality, but heck, it was free.

Re:what causes distortion inside the PC?

[Synic]

According to the documentation for the kX Project drivers for emu10k based boards (including the Soundblaster Live!) they offer a front/rear switch because the front speakers are using technology that has greater noise (lower SNR). Just switching from that to any other board with a decent SNR would be an improvement. It's nice to see that some Soundstorm solutions are decent, because my personal experience with AC'97 in the SNR department (and crosstalk interference) has been pretty dismal. At decent volumes driven by the AC'97 output line-out there is quite a bit more hiss on the line than even my SB Live!.

Re:what causes distortion inside the PC?

[Alereon]

The front outputs on a Soundblaster Live! are AC'97, the rear outputs are I2S. Before audio is fed to the DACs, the Live! resamples to 48Khz. This would be fine, except the resampling is done VERY poorly. Resampling 44.1Khz (or any other audio) to 48Khz before feeding it to the soundcard will noticeably improve audio quality. nVidia Soundstorm still uses an AC'97 DAC, but the Realtek ALC650 on my Abit NF7-S v2.0 provides decent quality. It still resamples internally to 48Khz, but the resampling is quite decent. You can do better in software, but the difference is really only noticeable in certain killer samples designed to point out resampling artifacts. I can max out the volume on my system and never hear any hiss or interference. It's pretty nice.

Re:That's how discovery works in litigation

[violently_ill]

the problem is: you can't HEAR the audio quality difference.

Does it have built in DRM ?

[havaloc]

I'm surprised no one has brought this up, but does it have any sort of DRM (Digital Rights) built in to it? If so, no thanks!

They are smoking crack

[melted]

Analog devices allowing full use of 32-bit dynamic range are PHYSICALLY impossible to implement due to thermal noise floor of your DAC and subsequent circuits. Heck, even 24bit dynamic range is impossible to fully exploit for that matter. 192KHz doesn't make sense either. 96KHz offers many benefits when constructing audio equipment (you can use crappy, "cheap" filtering algorithms and DSPs and still get excellent sound), but 192KHz is too much even for your dog.

Re:Can you recommend some computer speakers?

[swordgeek]

You seem to misunderstand the meaning of speaker wattage. This is the MAXIMUM power the speakers can withstand for a short time (I believe 1/4 second) without blowing up. It has no bearing whatsoever on speaker quality, efficiency (how loud they play at a given volume setting), amplifier requirements, or how loud they're "designed" to be played. Ignore that number entirely; it has no relevance for you.

That said, I settled on the Logitech speakers for my computer after a lot of listening--they're the only ones I found that sounded like music. I will admit that I didn't listen to the Klipsches, because they were out of my price range. I expect that they'd be quite good, as they make non-computer speakers which are very nice indeed. (Mind you, Altec-Lansing makes stereo speakers too, and their computer offerings are without exception, unmitigated shite!)

If you have a passable amp, then unpowered stereo speakers are likely to be the best choice. A few years ago, it would have been the only choice, but a few computer speakers are at least considering.

But ignore the 200watt rating. Even if it were valid (it's not), it's completely meaningless and irrelevant to your shopping.

Marketing gimmick

[I_LV_MSFT]

Next year expect 64 bit audio at 384kHz as well.

Bigger - better seems to work for most people, who are easily fooled by the numbers. First of all, there is no 32-bit content available on the market. Even 24-bit is not so common either (DVD-A is less than popular and SACD won't play on PC at all). And even if it calims to be 24-bit does not mean that there are 24-bits of significant information out there - mastering process is an art if its own.

Designing proper 24-bit audio output is not an easy thing to do. The main trouble is isolating the high frequencies which tent to travel everywhere. It is not impossible, but very few people will afford to buy it, so no soundcard manifacturer is doing it right.

If you think about it the lower bit in a 24-bit sample would contribute 1/16 milionth(+ change) part of the signal. There would be far too much other noise in the system for anyone to notice that change.

For any practical purposes (audiophile or not) 16-bits are good enough ant 20-bits are more than fine. Evan at that resolution other factors start to play and need to be properly mitigated, before someone could enjoy true high-fidelity audio.

By the way external soundcard is in theory higher quality, but in practice just a way for Creative and others to charge you more for the same crappy hardware.

Re:Marketing gimmick

[cheekyboy]

yes yes idiot, true for 1 sample 1 instrument, but try encoding 8 channels into 2 wires, you need more khz etc.. crap

try playing 175 instruments/effects doing 7.1 on just 2 wires (ie 2 real channels phsysically).

Thats only 31 samples per instrument at 44.1K

The spec of 192khz is for the final OUT PUT , not for 20000 indiviual instruments at that level.

Dynamic range of ADC?

[Crusty Oldman]

Does anybody know the true dynamic range of this ADC? The best I've been able to find so far is 24 bits at 192 kHz with the Delta44 board. I highly doubt that this unit will have a 32 bit resolution per channel as the article suggests, but I'm willing to be surprised.

I have an application, in software defined radio. See http://antennspecialisten.se/~sm5bsz/linuxdsp/linr oot.htm for an example.

Re:Dynamic range of ADC?

[Anonymous Coward]

The general rule is that you get 6 additional dB of dynamic range per bit. That means that a 32-bit system yields a dynamic range of 192 dB. The range of human hearing is much smaller (it's on the order of 100 dB). If we're dealing with human ears, 32-bit digital recording gives us a lot of superfluous bits taking up disk space. There aren't really any music applications where we need more than 18 or 20 bits. But companies shall do doubt continue to misinform consumuers...

Re:Dynamic range of ADC?

[Crusty Oldman]

The other "advertised 32-bit" boards, are actually only 24 bits of resolution, which is why I asked the question. And regardless the limitations of human hearing, I cited an application that DOES need a higher accuracy, which is why the question is important to me.

Slashdot readers, would any of you know the actual dynamic range of the ADC used in this hardware. Is it really 32 bits?

Improved stability?

[Oestergaard]

Oh, no, please, come on...

I suppose it will make my job more fun and my internet go faster as well, right?

That simple ending statement took all the credibility out of the release note - when will those PR droids learn not to overdo these things...

Stability... Argh! My head hurts.

get a clue, its a combine output. DILL..read......

[cheekyboy]

Maybe you need 192khz to split the signal up into 4 different 48khz channels, so you can do the 7.1 stuff out of 2 channels. ie 2 chans * 4 each = 8 total.

There are many tricky out of phase things you can do to encode 4 6 7 or 8 channels into 2 wire chans. In that case you NEED more khz to do it.

Clue train has left 12 minutes ago dude.

Re:get a clue, its a combine output. DILL..read...

[melted]

Why do you need to "combine" anything when you can simply record the same thing as eight losslessly compressed 20bit 96KHz channels?

More OT: Debian supports flac, not shorten

[Xtifr]

Getting even farther off-topic....

I don't know about the sound card, but as a member of the Debian project, I have to ask why you're using non-free third-party software like shorten when flac provides (slightly) better compression, faster decompression (if you ever need it), a more flexible format overall, and is included with and supported by Debian?

I only keep a copy of shorten around so I can decompress any .shns I get (from, e.g. etree or the internet audio archive), but I never ever use its compression features. The shorten-plugin for XMMS crashes on a fairly regular basis, while the flac plugin seems rock solid. So, I'm highly motivated to convert my .shns to .flacs ASAP. The 5-10% better compression that flac offers is just the icing on the cake. :)

Re:More OT: Debian supports flac, not shorten

[Doc Ruby]

I'd prefer FLAC to Shorten, if only just for FLAC's metadata support. But when I looked at the performance numbers last Summer, FLAC took about twice as long to compress, and offered only about 8% better ratios. I also don't recall FLAC decompressing faster, but of course there's a lot of FLAC development, and not much Shorten. I've also got code that encapsulates a Shorten file and streams it - can FLAC do that yet? I haven't had any problems with xmms-shorten, but I'd take a look at the FLAC plugin. If the features are even comparable now, of course I'd prefer the FOSS FLAC. Got pointers to Debian-specific resources?

BTW, my Slashdot fortune footer on this page reads:
It is the quality rather than the quantity that matters. - Lucius Annaeus Seneca (4 B.C. - A.D. 65)

Re:More OT: Debian supports flac, not shorten

[Xtifr]

>> Flac decompresses faster

> when I looked [...] FLAC took about twice as long to compress

Hmm, according to the Flac comparison page, we're both wrong. FLAC is only about a third again slower at compressing (12:54 vs. 9:44 in their tests), not twice as slow, but it's just a hair slower at decompressing (7:08 vs. 6:31). But that's in the default mode. If you compare "flac -3" vs. "shorten -p8", you'll see that the times are almost reversed, but FLAC still compresses noticably better.

Anyway, they're both slow enough that I would/do run them in the background. :)

I've also got code that encapsulates a Shorten file and streams it

According to the same page, FLAC is already streamable, so you shouldn't need extra code for that. But you can encapsulate it with Ogg if you want. FLAC also has support from a few hardware vendors, if that's relevent.

Latency!

[po8]

I only care about one thing in my audio HW these days, and it ain't bandwidth and it ain't dynamic range. 16 bits + 22KHz is plenty for me.

I want tiny latency. Single sample, if I can have it. Think about it: at 44K samples/sec, that means that you now have about 50K CPU cycles/sample to process stuff. This should be more than enough cycles to get a word in off ADC, process it extensively, and get it back onto the DAC.

PCI adds incredible latency, and IRQ handling adds more. Give me an audio HW standard with a path that lets me grab a sample, process it, and stick it on the output before the next sample is due out, and I assure you Linux-based SW such as JACK will immediately take full advantage!

Why care? Two big reasons: (1) professional studios and musicians view this as a requirement for their audio tasks. As an amateur keyboard player, I couldn't agree more. (2) syncing audio, video, and other events is a billion times easier in this setting.

Big mandatory audio buffers are evil. Please make them go away, without making me buy $2K worth of Hammerfall products. In return, I promise to replace all my audio HW with HW using the new Intel audio standard.

Another Reason Why I Don't Believe in Convergence

[Steve Franklin]

So I turn off the internet connection and plunk in a DVD movie I bought today, "AI" if you really want to know, and SURPRISE! no sound. Fine. So now instead of watching a movie I spend a good hour+ troubleshooting the damned computer. Finally reinstalled the XP update to the Creative driver that wouldn't reinstall properly.

This is precisely why convergence is not going to happen in any form that Bill Gates or Linus Torvalds imagines in terms of multimedia computers or onboard sound chips or video chips or the like. This all comes down to what I call The Washing Machine Metaphor. If you bought a washing machine that was networked with your computer and when you tried to wash your clothes, the water didn't come out because the godforsaken water driver didn't work right, what would you do? You'd bring the damned thing back to the store and buy one with its own onboard circuitry. This is the bottom line. When you want to wash your clothes, you want to wash your clothes, and when you want to watch a movie, you want to watch a movie, not dick around (yeah, I'm angry) with some piece of bovine offal written by some idiot who can't even spell in his native language. This is why it doesn't matter how good the sound from your computer is. That sound just isn't dependable enough for anything anyone in his right mind would call consumer equipment.

Mixing of samples...

[midifarm]

The problem when you start to mix sources, even in various speakers, you inevitably get what's known as phase cancellation. What this phenomenon does is literally cancels out sounds at similar dynamic ranges. This ends up reducing your viable number useable samples plus makes your final product sound muddy as you eluded to with your image analogy.

Plus answer the question, is this for video games or music/movies? If this sudden push from Intel is being driven by the video game industry it seems rather frivilous. I can see wanting to seem like you're there for a live recording or something, but to hear an explosion or a gun fire sample at a higher sample rate? The need just seems unfounded.

Peace

Headphones?

[Alereon]

Have you looked at getting headphones? With a nice pair of headphones, you can get quality that you'd have to spend several times more to get from speakers. Most notably, you gain an unmatched soundstage and incredible immersion in the music. Sennheiser Prestige HD-590 headphones retail for $150, and do not require an amplifier, unlike many high-end headphones. If you want incredible audio quality for a not-so-incredible price, they're your best bet. Check out the Head-Fi Headphone forums for more information and advice.

Re:Headphones?

[ketamine-bp]

I'll suggest Grados (SR-125, SR-225) or Audio-technica (A900 possibly) rather than sennheiser. Sennheiser are good, but Audio-technica often gives a better build and equal sound with equal money, and Grados, though looking rather retro, gives better sound per dollar...

Re:Headphones?

[Alereon]

IME, Grado phones sound rather bright and tinny (except in the very high end), and I don't really find them very comfortable. Audio Technicas are respectable, but all in all I think the Sennheiser HD590 is a very good headphone for someone just getting into high end audio, who wants headphones that are very versatile (good for all types of music, gaming, movies) and don't require an amp.

Sweet! Many purposes!

[Thomas Shaddack]

32-bit depth, 192 kHz samplerate... Too much for audio, but think beyond it. Think about all the signal processing applications - various sensors you can connect to the computer and let it process the incoming data. Think about all the scientific applications.

Think also about the intelligence applications. This depth/samplerate provides lots of redundancy, which is interesting for steganography applications. Think about trusted moderate-speed random number generators for cryptographic applications - just add a white noise generator and cryptographic whitening.

This is a GOOD thing, with many more uses than it may look, regardless of what the detractors say about overkill specs.

Re:That's how discovery works in litigation

[Robobo1]

What? Are you saying you can't hear the difference between 16 and 24 bit? Or 44.1kHz and 192kHz? You're kidding, right?

Now in Googlophone

[MythoBeast]

Does anybody remember the jokes that were going around when they introduced quadrophonic sound systems? They tried to expand speaker number beyond what people thought was reasonable back then, and some comedian made jokes about the "googlophonic sound system", with a separate speaker for every wavelength, coming from every direction imaginable.

How long before someone comes out with a 9.1 sterio system?

Re:Can you recommend some computer speakers?

[aXis100]

Agreed. This has been discussed to death elsewhere, but there are a number of ways to measure speaker wattage.

1) The cheat way (aka PMPO)
This is the absolute peak power of a speaker, ie the peak voltage versus peak current, for an absolute instant in time. It means basically nothing, and can be orders of magniture larger than any usefull output.

2) The right way (RMS)
This is the continous average (root mean square) power that the speaker can handle. Heat is usually overriding factor. If a speaker is not rated in RMS, then it is not worth buying.

"That's pronounced 'azaleas', dear"

[leonbrooks]

This post dedicated to those who can remember the first home-coming scene (ie, not the one featuring Anne Uumellmahaye) from The Man With Two Brains, a Steve Martin classic.

So in theory, two years later...

[leonbrooks]

...you should be able to hook a longwire directly to your audio-out jacks and transmit MW AM directly with unbelievably good fidelity?

Re:Can you recommend some computer speakers?

[CrystalChronicles]

Actually 200 watts is its RMS rating. Its rated 400 watts PMPO. why is 200 watts that hard to believe? its not like its that high in the first place.

people complain that BSDers are elitests

[evilviper]

Forgive me if I missunderstand, but I hope you don't mean Open Sound System. Last I heard that project was long ago superseded by ALSA

... And people complain that BSDers are elitests ...

Linux is hearding torwards ALSA, but the rest of the world is still using OSS. Whatever problems Linux had with OSS were purely implimentation details, because OSS in the BSDs works wonderfully. OSS works nicely on just about every form of Unix.

The switch to ALSA concerns me because it's another significant incompatibility between Linux and the rest of the world. A few more of these nice changes, and software for Linux is going to be just a platform-specific as Windows. You'll need to have a Unix version, and a seperate, completely different Linux version.

latency

[jago25_98]

Once again everyone disregards latency :`(

Re:Can you recommend some computer speakers?

[ketamine-bp]

I agree with you that the 200 watts is its RMS rating - but the true problem lies on:

(1) what is the signal-to-noise ratio at the rated RMS load - Certainly not the quoted value of >100dB - if it was 85dB or above It'll have large names like JMLab or so killed.

(2) What is its frequency? Yes the quoted was 35-20k, yet what is the cutting margin? Whether it is a 3dB bracket or it is an 10dB bracket or its a 0.01 dB bracket (Mark Levinson Amps, eh-huh.) the amplification does matter.

(3) 200 Watts is REALLY a lot. Most audiophile speakers rates at around 100 Watts, and turning it at around 20W makes a really loud sound for music anyway.

(4) I would wonder if its amp's is Class A or Class AB.

True. Music is not just a comm signal

[leftover]

Everyone who is trying to apply Nyquist to a music signal is missing the .. signal in the noise of this thread.

Communications doesn't care about the shape of the waveform, only the frequency. Nyquist gives the absolute minimum sampling rate required to identify the signal frequency and it assumes there are no higher frequencies present in the signal (hence anti-aliasing filters).

I Am Not an Audiophile, but in past data acquisition work I used 10 samples per cycle as a rule-of-thumb to adequately capture the shape of the wave. That criteria happens to fit well with the 192KHz rate, as shown by the parent post.

I do agree withe the comments about 32-bit D/A - that is just silly.

-1, Troll

[Canar]

If I tried to make a 20khz sound with 40khz sampling, I wouldn't hear a thing. Take two samples of one wave equidistant, and you get the middle and the beginning of the wave, which are always the same value. At 4X the sampling, you get a saw wave (or worse, a muffled trapazoid if you phase shift 45 deg). So, if you do 20Khz at 80Khz, you're still screwed. How many points do you need on a wave to make it smooth? I would say at least 8 at high frequencies (and that has a chance of only getting you about 66% of the power). That's about 160Khz for 20Khz sound.

No. You're wrong. At 40kHz, up to a 19.999kHz signal can be recreated with complete accuracy. Just because you think the signal would look trapezoidal if you did a linear interpretation doesn't mean that's what actually happens. If you did a simple linear interpretation, you'd add overtones up to your sampling frequency. To accurately interpolate the signal, you'd use the sinc(x) function, which will not affect the ~20kHz signal at all.

A 44kHz signal reproduces sound up to 22kHz very accurately. The only possible limitation is that when the audio is low-pass limited to 22kHz or whereever they decide to cap, the filtration process can leave theoretically "audible" ringing. Increasing the maximum sampling frequency only decreases this ringing. You won't be able to hear any other changes, as human hearing caps out at around 20kHz.

Please, please, please learn some actual signal processing before littering Slashdot with your half-literate tripe. Thank you.

Poster is wrong on several counts

[MuMart]

If I tried to make a 20khz sound with 40khz sampling, I wouldn't hear a thing. Take two samples of one wave equidistant, and you get the middle and the beginning of the wave, which are always the same value. At 4X the sampling, you get a saw wave (or worse, a muffled trapazoid if you phase shift 45 deg). So, if you do 20Khz at 80Khz, you're still screwed. How many points do you need on a wave to make it smooth? I would say at least 8 at high frequencies (and that has a chance of only getting you about 66% of the power). That's about 160Khz for 20Khz sound. I hate that someone actually thought the whole 2X the audible range was good enough to begin with. You may not hear a difference, but I can. If you don't, check the frequency response of your speakers.

No you can not hear a difference. Maybe you should read up on sampling theory sometime. A 44.1khz sampling rate preserves all phases and frequencies up to the limit of 22050hz without loss.

Yes, you need to reconstruct the signal mathematically, using oversampling (almost all soundcards will do at least 8x oversampling), but a higher source sample rate that 44100hz is pointless. The engineers who devised CDs were not stupid.

Re:what causes distortion inside the PC? - OT

[oddfox]

For about a week now I've been trying to figure out if it's worth my time to switch completely from my SoundBlaster Live! Value to my on-board nforce2 audio. I have a Shuttle AN35N-Ultra mobo that has an MCP southbridge, not an MCP-T, so I gather that it means I can't have all the fancy schmancy add-ons that the Soundstorm technology uses (Tom's Hardware notes that "you have to do without Dolby digital sound and FireWire" due to using the MCP chip). I don't know if it's an actual improvement or not, but it feels like my onboard audio sounds better than the SB card, and it apparently takes a bit less CPU. For the most part I think that the switch looks like a real good idea, but what's keeping me from the switch is a second opinion. Well, that and the whole task of setting up software mixing in my Gentoo setup so that I don't notice the lack of hardware mixing support in the drivers.

Think you could give me an idea as to whether or not the switch is worthwhile?

Re:Can you recommend some computer speakers?

[MoogMan]

Additionally, Wattage can be expressed in three commonly used ways. Two of them are perverted for commercial reasons:

RMS : The "original" standard. This is the only one that really means anything to anyone, if you will.
MPO : "Maximum Power Output". Basically works out MPO = RMS * 2
PMPO : "Peak Music Power Output". PMPO = MPO * 2

I like to think that a good rule of thumb is that if someone markets their product as xxPMPO then dont bother. Similarly with MPO. Although I havent bought any PC speakers (I use hifi seperates), so PC manufacturers may commonly use MPO. Beware!

As the parent noted, the actual "loudness" depends on the sensitivity of the speakers, cone impedance and a few other factors, so dont necessarily believe that higher wattage means higher volume

DVD Audio

[ThePyro]

I haven't experienced anything like the DVD audio problem you describe. But then again, I only have 2 DVD Audio disks and I don't listen to either of them very often.

I have experienced one obnoxious DVD-Audio problem... if I have my Creative Speaker Settings set to 5.1 (and I do have 5.1 speaker) then it comes out sounding like it's just stereo. The rear speakers don't get used at all. But if I set it to 4.1 then everything is fine. Strange.

Re:what causes distortion inside the PC? - OT

[Alereon]

I'm not sure what codec chip that board uses, but most of them use the Realtek ALC650. What you have onboard just provides DAC, ADC, and digital output functionality, all sound processing is done in software. As with all AC'97 sound solutions that I know of (including the SB Live!), it will resample internally to 48Khz. Overall, I'd guess that the onboard will probably sound better than the SB Live!. The only exception might be for gaming, I'm not sure how the 3D positional audio drivers are. Since it's fully software, games may be slower than the SB Live!. All things considered, I'd just use the onboard audio.

Re:what causes distortion inside the PC? - OT

[oddfox]

Thanks for the post, you have no idea how much I appreciate the input. Yes, this mobo utilizes a Realtek ALC650 codec, and does not include the APU that the MCP-T southbridge provides. However, on my Athlon XP 2000+, the burden seems to be quite miniscule, and I don't notice a performance hit in any games. I actually came across a comparison of the nforce and SBLive CPU usage, and it found that the nforce uses less CPU. I thought it was pretty shocking, even though it may have just been because of faulty drivers from Creative.

I think that I'll go ahead and get to finalizing my move over to the onboard audio today. All I have to do is figure out how I'm going to go about getting my Linux audio to work via software mixing. I guess that later on today I may end up removing the SBLive card completely from my case.

P.S. -- Not only am I using an old SB Live!, but it's a Value..

Ummm...

[midifarm]

I'm not trying to evangelize, but doesn't Apple include a dedicated optical S/PDIF I/O standard on all they G5 desktops?

Peace