Slashdot Mirror
Why Distributing Music As 24-bit/192kHz Downloads Is Pointless
An anonymous reader writes "A recent post at Xiph.org provides a long and incredibly detailed explanation of why 24-bit/192kHz music downloads — touted as being of 'uncompromised studio quality' — don't make any sense. The post walks us through some of the basics of ear anatomy, sampling rates, and listening tests, finally concluding that lossless formats and a decent pair of headphones will do a lot more for your audio enjoyment than 24/192 recordings. 'Why push back against 24/192? Because it's a solution to a problem that doesn't exist, a business model based on willful ignorance and scamming people. The more that pseudoscience goes unchecked in the world at large, the harder it is for truth to overcome truthiness... even if this is a small and relatively insignificant example.'"
I know, Stephen Colbert is Reddit's hero and they're starting to infiltrate this site as well, but seriously. Call them lies. That's what they are, that's what they -deserve- to be called. Are people really that passive-aggressive and afraid of expressing themselves that they won't call someone who lies a liar any more?
If you can't hear the difference between cymbals, bells, brass, and other "edgy" instruments at 44KHz/16bit "lossless" and 192KHz/24-bit, you're either deaf or using earbuds.
Idiot.
I do not fail; I succeed at finding out what does not work.
lossless formats and a decent pair of headphones and a set of really expensive MONSTER CABLES will do a lot more for your audio enjoyment than 24/192 recordings.
There, ftfy.
I record my performances at 96 kHz sample rate, I have to say that the music sounds much better at 96 kHz than 48 kHz I think (feel?) because the higher sample rate gives audio effects like reverb a lush, deeper sound.
The more sample units per second give the effects more to work with, in addition, even though you can't hear above and below certain frequencies recording those inaudible frequencies has an effect on the final product.
You may be able to find some scientific proof of this but for me it's an ear thing, higher sample rates sound better.
"If any question why we died, Tell them because our fathers lied."
"I can't hear your rational argument over the impeccably better-than-perfect sound from my 83 trillion dollar sound system. Thank you, Monster!"
/. haven't we had all of our music in FLAC for a decade now? I don't even listen to music much and mine is.
For the rest of us on
I'm not sure why this particular technology is so bizarrely specious in claims. I'm sure in fifty years we'll argue over the best neural interface with its platinum, massaged, better than reality addition is.
I use a pair of shure SE535 which i find to be some of the best earbuds on the market (fit nicely under motorcycle helmet) and I can identify tracks below 192kbps very quickly. 128kbps aac or mp3 sound very poor.
The article may have a lot of study, but reality is that there are those of use that could be considered audiophiles and truly can hear the difference.
Is is not possible that one day 'upgraded' sensory implants could be the norm for humans? Cybernetic generations to come may lament all of the lost audio information in recordings of our era.
Yawn. The point of higher sampling and bit rates is to have the most accurate representation of the original source material. Using a greater number of bits improves the dynamic range and reduces quantization noise. Some real instruments have spectral information above 22kHz, and most stuff created digitally in studios uses a sampling rate much higher than 44.1kHz.
If you want to do more with the music you purchase than just listen (and no doubt some people have the ears the appreciate the higher fidelity anyway) and do things like remix and reprocess, you want the better versions.
Can someone explain to me what KHz "sampling rate" has to do with the frequency range you can sample?
How many more years will slashdot have an off-by-one error on your Score in your profile?
I have a PhD in Digital Music Conservation from the University of Florida. I have to stress that the phenomenon known as "digital dust" is the real problem regarding conservation of music, and any other type of digital file. Digital files are stored in digital filing cabinets called "directories" which are prone to "digital dust" - slight bit alterations that happen now or then. Now, admittedly, in its ideal, pristine condition, a piece of musical work encoded in FLAC format contains more information than the same piece encoded in MP3, however, as the FLAC file is bigger, it accumulates, in fact, MORE digital dust than the MP3 file. Now you might say that the density of dust is the same. That would be a naive view. Since MP3 files are smaller, they can be much more easily stacked together and held in "drawers" called archive files (Zip, Rar, Lha, etc.) ; in such a configuration, their surface-to-volume ratio is minimized. Thus, they accumulate LESS digital dust and thus decay at a much slower rate than FLACs. All this is well-known in academia, alas the ignorant hordes just think that because it's bigger, it must be better.
So over the past months there's been some discussion about the merits of lossy compression and the rotational velocidensity issue. I'm an audiophile myself and posses a vast collection of uncompressed audio files, but I do want to assure the casual low-bitrate users that their music library is quite safe.
Being an audio engineer for over 21 years, I'm going to let you in on a little secret. While rotational velocidensity is indeed responsible for some deterioration of an unanchored file, there's a simple way of preventing this. Better still, there have been some reported cases of damaged files repairing themselves, although marginally so (about 1.7 percent for the .ogg format).
The procedure is, although effective, rather unorthodox. Rotational velocidensity, as known only affects compressed files, i.e. files who's anchoring has been damaged during compression procedures. Simply mounting your hard disk upside down enables centripetal forces to cancel out the rotational ruptures in the disk. As I said, unorthodox, and mainstream manufactures will not approve as it hurts sales (less rotational velocidensity damage means a slighter chance of disk failure.)
I'd still go with uncompressed .wav myself, but there's nothing wrong with compressed formats like flac or mp3 when you treat your hardware right
--
BMO
As a former audio engineer with some ranking success, I can tell you that it's true -- delivering high-sample rate audio as an end format is really pointless. It hardly makes sense in a studio, and definitely is illogical for the distribution of a final mix.
However, there is an increase in quality using 24 bit. Most people just assume increasing the bit depth is the same as increasing the sample rate, but this is incorrect and short-sided. With higher bit depths, you can get your analog components operating a little further away from the noise floor. This also makes dithering much less noticeable (the noise you hear when you crank the volume up as a song fades out). Why? There are more "levels" for each sample to be recorded into. It's like going from 16 to 24 bit color. You would notice this.
For the 192 KHz fans out there, there is direct and proven mathematical reasoning for why 44 KHz audio is plenty. That, and your equipment probably can't produce it. Your converters probably suck at this frequency, and your ears definitely can't vibrate that quickly. More samples doesn't "smooth out" the waveform.
Given how common time-stretched audio is these days — for DJing, looping, etc, high sample rate music files are ideal.
I love the word "truthiness"
my cat is doing the listening? Would 24-bit/192kHz music be better? Seriously. Not kidding.
So I don't subscribe to the $1500 per power cord group that some people do (usually the same folks who claim a $2400 USB cable increases the separation of instruments within a digitally encoded file).
However, I do own some good equipment- not the best, but pretty decent as far as studio setups go. ATM I'm rocking an Apogee Symphony I/O over Apogee's proprietary PCI-e interlink card (a Symphony 64). Yes, Apogee's driver support and customer support is shit, but when their equipment works it works pretty damned well. On the other end of that is a 5.1 setup consisting of four ADAM S2X speakers and a SUB12. The speakers were around $2500/pop and they're self powered (that is, they have the amplifiers built-in) and run over balanced XLR.
I didn't buy this equipment because it sounded "good" or "colourful" or "warm" or any of that bullshit. I bought it, because, when I want to listen to stuff that's in either 24-bit/96kHz or 24-bit/192kHz (which is a bit excessive, I'll admit)- I know that what I'm being audibly blasted with is as accurate as it will ever be. I don't care if the precision is sharp on the ears or unpleasant to some people. If I want to listen to music (when I'm not busy making it), I want to hear it exactly as it was recorded.
And in that regard, there is a huge difference between 44.1kHz/48kHz/96kHz, but lesser of a difference between 96kHz and 192kHz.
The thing about 192kHz is that it's such a high sample rate (a lot of people tend to work at 96kHz professionally), you need the equipment to handle it. Lots of interfaces will happily handle a couple of channels at 192kHz, but forget about streaming 16 channels at that same sample rate over anything that hasn't cost you a few thousand bucks and hooks up to your DAW/recorder over a proprietary high-speed interface.
So there's a lot of junk floating around out there that claims to be 192kHz, but with the right tools (I can't personally tell the difference with my ears) you can quite clearly see that only part of it (or none of it) was recorded at 192kHz. The studio gear used simply didn't support that sample rate, or they didn't opt to use it, or some outboard gear didn't jive well with it, or whatever.
My point here is that a lot of people will try to screw you out of money for 24-bit/192kHz music when in fact you're not getting anything anywhere near that. And a lot of people don't even know what the hell that means- so you get the kind of people trying to listen to that crap through a bog standard HTIB system in a box where the quality is such shit coming out of the speakers that you wouldn't be able to tell the difference between a CD and that stuff anyways.
So yeah, for the majority of people out there- 192kHz/24-bit is pointless unless: A) the entire audio pipeline that produced that tune was running at 192kHz/24-bit, and B) you have actual hardware capable of playing that back properly, and not some HTIB thing you bought from Futureshop that sounds good "because it's really loud".
Frankly, I find it hard to believe that enough people out there want 192kHz/24-bit for legitimate reasons (owning proper hardware for reasonable playback) that there's actually a market for this stuff. So it makes me think that this stuff is being targeted at people with iPods and shitty desktop speakers on their iMac computer. In which case, yeah, it really doesn't matter. You're not going to hear any difference between a lossless FLAC file at CD quality or a 192kHz/24-bit file freshly bounced from the studio masters.
-AC
I see no rational basis for limiting myself to audio intended for those with hearing worse than average. (Nor do I limit what I read because of the poor reading skills of others; limit my choice of where to walk because too many have lost the skill in their desire to drive everywhere; limit who I know because politicians like to divide humanity into them and us; etc)
Limit yourself by personal ethics or by personal physiology, not by pseudoscientific efforts to brand "standard deviations" as deviants.
"Because it's a solution to a problem that doesn't exist, a business model based on willful ignorance and scamming people."
which happens to be a business model that works, unfortunately
intellectual property law is philosophically incoherent. it is your moral duty to ignore it or sabotage it
Ask any GeekSquad or Best Buy salesmen and they will tell you that you need full gold plated $2,000 HDMI cables for professional audio quality and $110 Monster ones for basic audio and video. They are not highly compensated so well for nothing you know
http://saveie6.com/
Two words mate: Libel Laws.
If you live in the UK, be afraid of what you say in public. Be very afraid. One wrong word, and you are screwed.
You mean like, honkies, spics, niggers, dune coons, prairie niggers, kykes, faggots, chinks, canucks, wops, guineas, krauts, and polocks? I think that's everybody anyway, my apologies if I left out any group, I try to be an equal opportunity offender, challenging people to be adults and get over their group identitied. Criticism welcome. Cowardly disapproval spurned.
I know, Stephen Colbert is Reddit's hero and they're starting to infiltrate this site as well, but seriously. Call them lies. That's what they are, that's what they -deserve- to be called. Are people really that passive-aggressive and afraid of expressing themselves that they won't call someone who lies a liar any more?
Okay, everybody, listen up: Anonymous Coward is having a rough day so let's all be extra nice to him!
"I like to lick butts!" by MobileTatsu-NJG (#32700246) (Score:5, Informative)
Is it just me or is this article just a bunch neo whiny cry babies with crappy 16/48 audio cards trying to talk down about well built 24/192 cards
It sure ain't the SCIENCE cause 24/192 is both more samples and more bits than 16/44, so the article is anti-scientific to claim it's a better signal at 16/44 or 16/48.
Finally we get to the meat, distributing music, okay correct for a download distro 24/192 is a stupid format to download, personally I'd rather have 320k mp3, at some point it is easier to just mail a DVD's with those tracks to someone who must be working with 24 bit tracks. It's more of a production format than a buying a CD (in this case DVD) to listen to your favorite band. Most people don't sell this format, just like they don't sell WMA, or .mod files instead of mp3's, or wav ~cdda so the argument's a moot point, and the few people that do sell this format, who gives a shit, is it really bothering you so much you have to tear it down, why don't you take a deep breath and check that your mortgage paperwork isn't signed by linda green?
Creative Audigy 2 ZS platinum pro is now eight years old and on XP taking up one PCI slot, still kicks most of your ass to this very date. Two out of the two I bought, still work and they both have been through several motherboards which fried.
Truthiness refers to a specific kind of lie-- a lie that sounds true, and that a large segment of people really want to be true. The kind of thing that's close enough to true for AM radio talk show hosts.
And now... I'll get off your damned lawn. Don't forget to take your teeth out before falling asleep.
clearly. There users. Surprise niigers everywhere
Duh! It's called political correctness. And if you even dare show a pair, others will kick them till you're blue in the face.
Life is not for the lazy.
than their CD transfers. In the cases where I recall being most disappointed (I've thrown out almost all my vinyl records), it was the dynamic contrast that was missing in the CD versions, for example a pianist striking chords from dropping his hands a foot above the keyboard.
Maybe these were just bad transfers... I don't know.
The Nyquist limit is the highest frequency that can be represented, yes.
But at Nyquist, only one shape of waveform can be represented. Depending on the design of the DAC, it could be a square wave, triangle wave, or sine wave. But only one of those.
With this in mind, I don't understand why Monty says that beneath Nyquist, everything is captured perfectly and completely. That seems plainly untrue to me.
The value of higher sampling frequencies isn't to reproduce frequencies above 20kHz. The value is to preserve the characteristics of waveforms within the range of human hearing, pushing aliasing artifacts into the ultrasonic, where they can be gently filtered out between 20kHz and 30kHz.
That said, to me that means there is some value in 96 kHz distribution.. 192 kHz does seem like vast overkill.
If you live in the UK, be afraid of what you say in public. Be very afraid. One wrong word, and you are screwed.
The whinging pommy bastards can suck my balls.
You my friend, are a stranger to the truth.
If George Carlin were still alive, he would mod you up right now.
for future DMCA kruft
Xiph.org must be talking about elevator Muzak because:
1. High-Frequency Sound Above the Audible Range Affects Brain Electric Activity and Sound Perception
2. High-Frequency Sound Above the Audible Range Affects Sounds Within the Audible Range
http://www.cco.caltech.edu/~boyk/spectra/spectra.htm
That copypasta hasn't been funny for at least five years if ever.
If you wanna troll, let's go... I'll take your side, you take mine and no one under the age of thirty will have any freaking clue what just happened. /g/
>>>
yarbles*
"Political correctness?" More like "cowardice." If you're too afraid to stand up for what you believe in just say so, don't try to hide it behind more useless language like "truthiness."
The article only debunks putting 24/192 and 24/96 audio as the audio that goes on your iPod, because the equipment (eg cheap speakers and headphones) is more likely to damage your hearing from lacking the required fidelity to reproduce the full spectrum but not filter out the dangerous naive mastering.
The problem is not with 24-bit/192kHz music downloads. The problem is some idiot is touting them "as being of 'uncompromised studio quality'". Who said they're even close to lossless...?
There is NO problem with the technology.
Back to content...
Educating people is fine, but the elitists will always say swear that x is better than y, even if it is provably otherwise. Just like some people will swear they saw Elvis working as a hooker at the Rt. 97 truck stop blowing Jesus.
Silence is a state of mime.
I think Truthiness covers half truths too. A half truth is that 24-bit/192kHz audio is higher quality than 24-bit/96KHz audio.
The whole truth is that only your house cat would be annoyed at 96KHz, or an audiophile dog.
I know, Stephen Colbert is Reddit's hero and they're starting to infiltrate this site as well, but seriously. Call them lies. That's what they are, that's what they -deserve- to be called. Are people really that passive-aggressive and afraid of expressing themselves that they won't call someone who lies a liar any more?
Okay, everybody, listen up: Anonymous Coward is having a rough day so let's all be extra nice to him!
You're a filthy, stinky fucking nigger-face who thinks he;s funny. That's the tragedy of it all. Deep down past that, you're a patheticly sad little man who craves attention.
There is a huge problem with file sizes (so both hard drive space and download bandwidth) with lossless files
Not any more, pumpkin.
We hit the terabyte size in drives a couple of years ago.
Grandparent said "and download bandwidth". Filling one of those drives takes four months or more over an urban home Internet connection capped at 250 GB per month, and that's if you don't do any Facebook, Slashdot, Cracked, deviantART, Netflix, YouTube, or everyone else's favorite bandwidth hogs. Over rural broadband, it takes 16 years due to 5 GB/mo caps imposed by satellite providers.
A couple of my own notes about ruining the dynamic range...
- today it's taken to a new level, I see more and more songs where not only DR is destroyed but sound is also harshly clipped
- not only mastering, but the mixing step is also to blame, where every track volume might be cranked up or, too many tracks recorded where the song becomes an (easily-compressible) mush
- it sometimes appears that a band might have the first one or two records with good sound quality, but when they make it big, they "bend over and take it up the ass", thus the rest ones being crap
He mentions in the article that 24/192 source would be a benefit to those that need to reprocess the audio, such as in remixing/effects/etc.
Therefore, using his own logic, you can deduce that being able to download and/or purchase music in those formats would be a benefit to anyone wishing to use it as a source in their mixes.
So he proves himself wrong that distributing 24/192 is pointless, as im sure there are plenty of remix artists out there looking for high quality samples, and many of whom use sound processing hardware designed specifically for a 24/192 source.
-HasHie
And don't forget to put yours back in after you're finished sucking Colbert's dick. Then go back to reddit with the rest of the pedophiles.
I made a rule in our facilities a few years ago that if it wasn't at least 256 Kilobits, we wouldn't air it.
That'd make it impossible to play mono recordings, such as 8-bit game soundtracks or old Beatles songs, because in many codecs, mono typically maxes out at half the bitrate of stereo. MP3, for example, doesn't go over 160 kbps per channel if I remember correctly. Or by "256 kbps" did you mean "128 kbps per channel, which for most recordings that people in the studio will deal with means 256 kbps"?
Inaudible (super hi/low) resonant harmonic frequencies sustain a waveform over time = less volume needed - fidelity improved. But A/D D/A complicates things
Even TFA states that mastering comes into play very much, and I've noticed that 24/192 music usually is way better mastered than 16/44. It kind of makes sense, since why would you bother releasing 24/192 through a crappy analog chain, while 16/44 is so ubiquitous that resulting CDs run the gamut on mastering quality. While I agree you will not hear the difference between _perfectly mastered_ 16/44 and 24/192, I think there is a greater point which is missed, and that is mastering tends to run better with higher fidelity formats since crappy mastering is more obvious with 24/192. Maybe 24 bits will lower the noise floor more so high dynamic instruments (drums, etc) will come across a bit better due to less compression usually applied. Not sure.
Of course this is ludicrous.
No one can see X-rays (or infrared, or ultraviolet, or microwaves). It doesn't matter how much a person believes he can. Retinas simply don't have the sensory hardware.
I wouldn't be so sure... $10 IR filter goggles. The human senses do have limits, but they're rather soft and fuzzy. First, there's genetic variation in the exact sensitivity range (e.g. some people can perceive further into the "infrared" spectrum than others, it's a common high school & college lab experiment). Plus, pedantically, everyone can detect IR up to 3,000 nm at least, cooking would be highly impractical otherwise, and Beethoven felt for vibrations so he could continue composing/performing despite his deafness (IOW, our senses overlap, very important for concert goers that like to feel the bass).
Second, and more importantly, the raw signals are integrated by the brain in a semi-predictable pattern (obviously it's a self-teaching neural network, so people process things differently, although there are common trends). An insect has a compound eye with dozens or hundreds of photoreceptor units. Individually, they're not terribly sensitive, but when integrated provide a much clearer picture. It's akin to how photographers can merge multiple overlapping images to create gigapixel-level quality.
Given harmonics, pinna distortion and such, it wouldn't surprise me if hair cells do not impose an absolute limit on hearing, as the article states. OTOH, I doubt that 192 kHz offers any real sound improvement, but I don't think you can argue that with just biology, as there are few, if any, definites in that subject.
I'll buy that a higher sampling rate makes certain reverb techniques algorithmically simpler. But if the original signal doesn't have any audible energy over 20 kHz, why store energy over 20 kHz in the mastered file? You can downsample from 176 or 192 kHz to 44 kHz during mastering. Then on playback, the sound card resamples it back up to 88, 96, 176, or 192 kHz to filter out ultrasonic images before handing it to the DAC.
As I'm very interested in dating that girl.
There are entities I'm sure who could do marvelous things with the bandwidth above 44KHz.
These same entities also have plans for the extra 8 bits of color in your 32-bit colordepth capable devices.
Just curious - in the same way higher quality imaging allows for larger scaling, does higher quality audio allow, for instance, louder music to be heard more clearly?
Than an "audiophile"? I think not.
Yup, it's one in the same.
Life is not for the lazy.
This article is of particular interest to me because just recently I dropped over $1000 on a pair of high-end Sennheiser HD-800 headphones, but now I'm finding the amplifier's background noise is a lot more noticeable. Before, with cheap headphones that didn't have the same dynamic range, it didn't matter, but now it's the limiting factor.
I've got a reasonably decent FLAC collection, some of it classical music in 24/96Khz format, but the background hiss is detracting quite a bit from the potential quality.
What's a good amplifier for headphones that's optimised for a low noise floor instead of power, but isn't over-priced? I don't need a receiver with hundreds of inputs, I need something that takes a single digital input, and outputs the highest possible quality headphone output. Is there anything like that out there?
That might hurt if it didn't come from a guy who is angrily whining about a pop-culture reference on Slashdot.
"I like to lick butts!" by MobileTatsu-NJG (#32700246) (Score:5, Informative)
forgot "abo".
The correct word is "verisimilitude."
English is already diverse enough that we don't need to invent stupid synonyms for useful words that already exist.
Difficult to explain, but it reminds me of how some people say that there's no point having a frame rate higher than 30 fps. No, your eyes can't actually see the screen flickering above that frame rate, but that doesn't mean it looks perfectly fluid. The author is assuming the point of diminishing returns is actually a point of no returns, which may be far from the truth.
If it weren't for the fact that all popular music has its dynamic range compressed to provide maximum loudness for the entire song, dynamic range would be be a problem.
The problem is that, on soft passages, where the high 8 or 10 bits are zero, you're listening to 8 or 6 bit audio. That quantization can be heard. This is a problem for classical recordings made without any dynamic range compression. Of which there are very few.
This is an issue only if you listen to classical music in a very quiet environment. It doesn't matter for car audio. It doesn't matter for Apple's trendy crap earbuds. So almost nobody cares.
^ So says the article...too bad MPEG audio (including MP3) wasn't finalized until November 1992, with a public release in 1993, and formal specification in 1994...(first software mp3 encoder wasn't released until July 1994) :P
Unfortunately, such a gross overstatement kinda makes me doubt everything else in the article.
"A Goddess rarely smiles for she is forced by others to be an island unto herself." - Zephiris
I think I can find a compromise that should work for everyone: Why not just run the needlessly good 24 bit 192 hHz music file though a lossy compressor that does psychoacoustics well - something like AAC or maybe even OGG? Everyone agrees that the vast majority of the data in 24/192 can be thrown away with zero perceptible loss. Fine, let's do it. But let's do the bit discarding in some principled way, guided by a reasonable psychoacoustic model. Isn't that a lot better than indiscriminately downsampling to 16/44.1? By anyone's lights, a 16/44.1 FLAC at 1100 kbps will not sound better than a 24/192 OGG at 1100 kbps - or even 700 kbps, for that matter. The nice thing about this plan is that we have good models for the human threshhold of detection. Scientists claim that 16/44.1 is so good that any improvements on it will not be detected. Maybe, but what if they're wrong? Why not start with the data rich source and apply our acoustic models to throw out only the data that we know is FAR FAR FAR BEYOND our threshhold of detection? It would still be most of it, but at least we'd know we're throwing out the RIGHT data.
I've done double blind studies on my dog and he can tell the difference.
Article is full of crap that is just plain wrong and misguided and the analogies suck. I'm not sayin 192KHz 24 bit is needed but the article is weird and says things that just are not true.
For example, saying that a stairstepped sine wave is mathematically the same is wrong --- stairsteps are impulses convolved with a square wave "impulse". This creates a roll-off at high frequencies. Basic signal processing. If you don't understand it, don't worry. Many don't. But the resulting sine wave will be the wrong amplitude. Sampling theory is based on infinitely fast impulses at each sample point, not stairsteps. A subtle point, but he misses many subtle points.
As for 192K 24Bit, there are reasons it is useful as opposed to 48Khz or especially 44.1KHz 16Bit:
1. Dynamic range. 20 bits gives 120dB +_ a few. But 16 bits (96dB) is not enough. 24 bits is way overkill, but doesn't hurt anything except storage space. Home theatre systems with 16 bits make audible noise when you turn them up. Put you ear next to the speaker when it is quiet and you will hear hiss. It may hurt your ears if they are there when when some sound comes through, but that depth is audible. His assertion that 16 bits is enough is not science, it's his opinion. (maybe even 18 bits is enough, but 18 is on the edge)
2. Simplicity of DAC - 192KHz means that dac filters can easily remove images. 96Khz is high enough to make the filter job simple, but 192Khz is simpler yet. His assertion that doing steep filters in digital is no issue means he doesn't really understand digital filters. Steeper slopes means higher lobes and more passband ripple.
3. All his talk about ultra sonics is laughable. Design a bad amp and it will sound bad. So what? Oh --- put in a bad signal so it will sound better?
4. My only point of full agreement is that you need good equipment first, 192/24 second. And I partially agree in that 192/24 is overkill.
There's a whole lot of snake oil in the audio business that needed some serious debunking.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
This article is bunk. I wasted 15 minutes reading through it. And they didn't even cover multi-tone and complex waveforms (which would have shown it to be bunk). Pure sine waves actually do well with digital sampling. But as you reach the edge of the Nyquist limit, you reach a point where the number of waveform states (how many sines waves of various values can be mixed) that can be rendered by the sample converges to unity. E.g. it can only support ONE sine wave at that point. Raise the sample rate and then you have the capacity to render multiple sine waves at the same frequency and many others.
A higher sample rate at say 192kHz is NOT done for the purpose of being able to encode sinusoid components up to 96kHz. It can do that (with that one sine wave limit that point). But is is appropriate to sample after a low pass filter (for example at 18 kHz) that limits the signals to only what you want. And then after conversion back to analog, clean it up with the low pass filter (again, at 18 kHz).
Listen to speech filtered with a 4kHz lowpass filter in an all-analog path. You will be able to tell it is filtered if your hearing is normal. Now digitize that filtered speech with an 8kHz sample rate. Convert it back to analog, and filter it again. The highs (up to 4kHz) will still be there (Nyquist says so, and this is valid). But there will also be new intermodulation products all over the place, especially among the high frequency components. It will give the audio a tinny or metallic sound quality.
Looking at it as combinations, a 44100 Hz sample rate at 16 bits is enough to render a 22050 Hz tone at any of 32768 intensity levels. However, if you have a 2nd tone of 22000 Hz, with each at 16384 intensity levels to avoid an overload, there are now 268435456 level combinations to be encoded. Now the 16 bits isn't enough. You need to double it. That can be done by either 32 bit sampling (hard to do) or doubling the sample rate (still 32 bits but now done as a pair of 16 bit samples). Fortunately you won't have mixed signals that high very often. However, you can easily have many signal components at lower frequencies. You will need plenty of bits for each. Even 192 kHz sampling is not enough to render 4 full range sine components at around 4 kHz. One or even a few levels of inaccuracy won't be heard. But these combinations rise very rapdily with just a few components.
For wider band audio with a higher sample rate, because most people hear weakly at higher frequencies, the effects will be less perceived, if at all. But they will be there, and a small portion of the population (including myself) can hear it.
Personally I'd rather they would go with 32 bits and 480kHz sampling.
now we need to go OSS in diesel cars
All this talk about 44 and 48 and 192 is interesting.
But the question is, does it go to 11?
As someone of kraut, dago, limey, paddy, and prairie-nigger descent, I'm highly offended. You forgot almost half my family. Thanks a lot.
This is what you would see from the apple store:
Crappy album, $8.99
Crappy album lossless audio $14.99
Crappy Album 24/192 $22.99
Even though the album was probably recorded at 24bit 96k or less.
People would buy the 24/192 version, play it on their computer who's audio driver is set at 16bit 44.1 over some cheap speakers who have a range of 60hz-12khz, that cant even be stressed by a low bitrate lossy audio files or on their 6.99 ear buds from their iphone.
While we could use an improvement in audio files for the end user, let keep it somewhat realistic.
While a higher sampling frequency and bit depth is a big help during recording and editing, much of the sound quality can be kept during down sampling when done correctly before distribution to the consumers. If people were downloading 24/192 files, they would use some free audio convertor to encode them to mp3 or m4a to put on their portable device, then we would be right back where we are at now.
I once had the same idea that 192kHz is overkill, but I've been following a course on digital audio processing and I'm not so sure any more: it's not about the frequencies per se, but about the shape of the waveform. This has its influence on the timbre of the sound. Not that I'm convinced that 192k quality can be heard either, but it's not simply a matter of the Fourier spectrum and frequency response of the ear. That said, a lot of it is probably just nitpicking, if you want better sound I suggest investing in better speakers just like the article said. When playing high quality music through laptop speakers, the sampling rate isn't the reason why it sounds shitty.
"It's too bad that stupidity isn't painful." - Anton LaVey
If you record sound with the Bass guitar at live gig levels, the bass goes squeaky with most forms of compression. Severe wave forms break during most forms of compression. With a straight WAVE file it sounds fine. That's what I have found.
The purpose of existence is to make money.
Because we should be long-sighted. Sound technology is evolving just like all other technology - in a decade, we may have drastically better speakers, for example. And at that time, I don't want to have to redo all my music.
If we can stay ahead of the curve, we'll be better off (just look at the betamax>blu-ray progression, didn't anticipate better quality TV's)
The reason the studios use it is because then often modify the music/samples. For an end listener its pointless.
What? Dogs can't enjoy music now?
-- no sig today
Linear quantization never made sense to me as far as encoding audio. Human ears, like our other senses, are logarithmic. The difference in linear intensity between two soft sounds is far more detectable than the same difference between two loud sounds. Linear quantization is thus wasteful in one end of the absolute intensity scale, and possibly insufficient in the other end. Why use an encoding so far from the optimal? Hardware considerations are not a good excuse because the same digital processing circuitry that the average delta-sigma DAC chip in every piece of consumer gear uses to convert the audio into a high bitrate/low bit depth stream before actual conversion to an analog signal can be trivially modified to handle nonlinearly quantized inputs.
"Politicians and diapers must be changed often, and for the same reason."
Excuse me, sir, I don't believe you did a peer-reviewed study to determine if he was a troll or not. Until you can show me some data in a proper scientific journal that he is a worthless troll, I think it's an open question still.
Plus, if he is a troll they have those big pointy ears, so that's clearly how he got his great hearing. You know they live under bridges for the acoustics, right?
Big apple, new Yorik, undig it, something's unrotting in Edenmark.
24 bit is a potentially significant upgrade. 192khz is not. What really makes a difference is surround sound. 5.1 music sounds amazing and I'd take a 44khz 5.1 channel recording over a stereo (2.0) 192khz recording any day. Grab an SACD or DVD-A disc if you can find one and check it out (surround sound only, as I think most agree the increased resolution is pointless). It's the best sound upgrade you can make (if you can find something you like in the pathetically small amount of surround sound releases). There is a version of Blu-ray (3.0) that is audio only that I hope takes off as SACD and DVD-A are all but dead.
The mighty wiki disagrees: "The reported completion date of the MPEG-1 standard, varies greatly: a largely complete draft standard was produced in September 1990, and from that point on, only minor changes were introduced.[2] The draft standard was publicly available for purchase.[14]"
Analogies don't equal equalities, they are merely somewhat analogous.
There is one aspect that my rusty EE signal theory can no longer reproduce: what about phase accuracy of frequencies near the sampling frequency? Isn't it true that for a simple S&H quantizer, the amplitude information captured by the filter varies with cos(d), where d is the phase difference between the sampler and the signal?
I understand that higher-order filters are less susceptible to this quantization loss, but don't they do so at the expense of phase accuracy? Even if you sample at 192kHz or higher, how do you maintain both phase and amplitude accuracy if the reconstruction rate is only 44kHz?
Many people think a "factoid" is a small fact. Actually a factoid is something that sounds true, but is actually false.
You were mistaken. Which is odd, since memory shouldn't be a problem for you
You mean like, honkies, spics, niggers, dune coons, prairie niggers, kykes, faggots, chinks, canucks, wops, guineas, krauts, and polocks? I think that's everybody anyway, my apologies if I left out any group, I try to be an equal opportunity offender, challenging people to be adults and get over their group identitied. Criticism welcome. Cowardly disapproval spurned.
No no no... Porch Monkey. It's okay, we're taking it back.
You missed the "soulless" AKA Gingers.
If george carlin were still alive, i bet he'd want out of that grave and not really care about someones post on an internet forum.
This article blows. Its rife with errors and assumptions the author doesn't understand. I'm going to go kill myself now.
Sincerely,
Every AES Member
96KHz isn't the audio frequency. It doesn't mean that the audio contains a 90Khz tone. It's the sampling rate. The higher the sampling rate smoother the signal.
Human perception wise a audio signal recorded at 96KHz sampling rate might well be indistinguishable from one sampled at 192Khz, but so is the file size between these files for practical purposes. I don't deceive my self thinking that I'm hearing better sound from a 192Khz file, specially considering that I'm using a basic pair of headphones on a my basic phone to listen to them. But my thinking is that future technologies might let you do interesting things with the extra bit of data which is useless to us right now. So given the choice I opt to get the higher sampled versions. Kind of like with digital pictures which are too noisy or blurred, but which might be cleaned up with future algorithms to give us a slightly more useful picture.
I may not be able to hear the difference between the quality increase, but, since the overall size of music files is so small relative to bandwidth and storage, I prefer FLAC et al simply for the sake of having a true archival version.
You forgot tundra nigger. Mustn't ignore the northern peoples.
It is also the lone insult in a hilariously offensive five minute tirade that got a guy beat down by a cop up here. Everything else rolled right off, until he came out with that one.
Oversampling (i.e. 192kHz) allows much more room to develop a good anti-aliasing filter. [... filters] designed for the higher sampling rate can have more linear phase.
This is an especially serious issue for percussive sounds, which have both a very broad spectrum and a strong sensitivity to phase errors in reconstruction.
The broad spectrum means there's a lot of energy in the high frequencies that map into the audible range due to sampling aliases. Oversampling lets filters have greater image attenuation.
Percussive sounds are very short and the phase relationship between the harmonics must be maintained to keep them short when reconstructed. So a non-flat phase response in the antialiasing filters lengthens the time of the reconstructed sound. This is VERY audible, making the sounds "muddy" rather than "crisp". Phase distortion also interferes with reconstructing the apparent location of the sound source in stereo and other multi-channel audio systems. So the flatter phase response of the anti-aliasing filters that are possible with higher sampling rates produces a very noticeable improvement in the sound quality.
= = = =
I learned that last from Steve Eberbach, designer of the DCM Time Window loudspeakers. These had a very flat phase response, good enough to allow a listener to track thechanging location of the "veep" sound of a recorded accoustic guitarist's fingers sliding on the wound strings. In addition to not distorting the sound (thus not producing an acoustic image of the enclosure), the speakers also had a hack to cancel the reflection from the wall behind them, resulting in the acoustic effiect of the room's wall going away, becoming a window on the recorded performance. Thus the name: Time (because the response was flat in the time domain (phase) as well as frequency), Window (for the "window on the performance" effect).
CDs began to come out shortly after the introduction of these superb loudspeakers. And Steve had a lot to say about them. The low sampling rate chosen and resulting rotten filter phase response wiped out much of his speakers' advantage over the competition. (The choice of a linear, rather than a floating-point-like compressed, encoding also limited dynamic range, making quantization error audible as noise and intermodulation distortion in quiet passages.) Only listeners playing vinyl disks or dolby tapes could really appreciate the difference between his product and other high-end speakers.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
In the grand scheme of things, we're all pretty much blind and deaf.
Ydco co
"It's true enough that a properly encoded Ogg file (or MP3, or AAC file) will be indistinguishable from the original at a moderate bitrate." Rubbish. Any lossy format but particularly mp3 sounds GRUESOME to anyone with a trained ear. And untrained ears can certainly tell the difference once it's pointed out, usually on a good system. If you want to know how to get 11:1 compression ratio on a pseudorandom source like sound, it's simple - they throw away most of the information ,particularly spatial information in the upper frequency ranges. You can't "hear" some of those frequencies, but you can certainly perceive when they are absent. I personally cannot stand mp3s and never use them. FLAC all the way.
There was already a perfectly good word for that.
That's not equal opportunity offence: anyone who matches two of the labels is offended twice as much as someone who only matches one. You should apologise to canuck faggots for double-counting them as well as to all those (e.g. limeys, frogs, lipstick-wearing pigs) you forgot.
I gotta go with the AC on this one, as that stupid ass word is okay in politics where every damned thing is just a different degree of spin so "truthiness' can be appropriate but here I doubt its being done for some sort of spin, more likely its a classic case of "biggerer is betterer" and Lord knows we've seen that enough in society, everything from SUVs to Hollywood boomfests, so I have to say that stupid ass word just don't fit in this situation.
As for TFA? Meh I suppose its all pretty much relative and how much abuse your ears have taken on what sounds good or not to you anyway. Most here would probably gag if they picked up my MP3 player as its all 64k but after 30 years of playing rock bass with big ass amps combined with all the outside noise frankly when i'm out and about I can't tell any difference. Now of course inside is a different matter but i don't have a bunch of noise but even then anything from 192k through 320k sounds fine to me and i'm sure if you gave me a blind listening test i doubt i could tell the difference between lossless and 192k.
So why not just let folks choose from whatever size they want? Its not like the old days when we had to squeeze every bit of room out of our 10Gb HDDs, just give us 192k, 320k, and the 24bit 192khz and let us listen and decide for ourselves.
ACs don't waste your time replying, your posts are never seen by me.
Suppose you two sound sources producing a 94KHz sound, and a 90KHz sound (say a sine wave). Individually the human ear cannot hear them. But played together interference will create harmonics, one of which is at an audible 4KHz.
196KHz is unnecessary as a final product, but if the sources are to be mixed then the extra range could create audible sounds.
The point of 24bit/96khz is that it allows for better processing of the data.
It's meant for music post-production (applying effects/plugins, remixing, etc).
Who says that 24bit/96khz downloads are made just for listening?
There are no "final mixes".
Those audio qualities are meant to allow remixing of the songs.
The problem with everybody commenting in this topic/news/thread is the premise that 24bit/96khz recordings are meant only to be heard. Yeah, there is no point in hearing music at that quality. But there is a lot of people that enjoy remixing music, and that is definitely what they need.
Having uncompressed, lossless audio of YOUR music (yes, you buy it, it's yours, and you own it, and you can do what you want with it, et cetera, et cetera, et cetera) allows you to do post-processing that you otherwise would not be able to do with a shitty compressed AAC.
Let's say I wanted to dub a song I own over a home video I took of my kid sledding. Let's say I wanted to add some effects to it. I could do this if I had high-quality sampling of the original. It would sound like shit if my source was a 128kbit MP3.
The higher the sampling rate smoother the signal.
Well... no. There's enough information in a low sampled curve. As TFA explains it, the output isn't "jagged" when played back in analog.
Human perception wise a audio signal recorded at 96KHz sampling rate might well be indistinguishable from one sampled at 192Khz
as explained in the article:
- Yup the human ear won't hear anything aboe 20kHz sounds, because it doesn't have any receptors for that.
But there are some real-world problems that come into the mix. No audio installation is perfect. You always get distortions.
- Thus, a 192kHz sampled file could contain frequencies up to 96kHz. These are sound which can't be heard in theory. In practice if you throw 96kHz frequencies to a sub-optimal speaker, the speaker can barf a lot of distortions, including distortion below the the 20kHz. So not only are you trying to output a sound that can be heard, but you force the speaker to produce bad noise *which* is audible.
But my thinking is that future technologies might let you do interesting things with the extra bit of data which is useless to us right now.
Hard to do anything with those bits at all. We simply lack the anatomic feature to do anything with them. Unless you do something like transpose everything at lower frequencie (slow down everything 2x = move everything 1 octave lower). At which point you aren't really outputing the original sound anymore. You're simply using the data to produce new sounds that weren't here to begin with.
The only practical use-case for this would be zoologist studying animals whose sound are beyond the human hear range. In that case "moving everything a couple of octave down" would help the scientist have an approximation with which he can work (to find rythms or other variation that are inaudible in the original frequency range). But that has nothing to do with hearing music made by human, for humans, with instruments designed for human hearing ranges.
Kind of like with digital pictures which are too noisy or blurred, but which might be cleaned up with future algorithms to give us a slightly more useful picture.
The situation with pictures is slightly different. What you're speaking about is spacial frequency. I.e.: resolution.
And human eyes can percieve way much more than some blurry low-res pictures. And in addition to that, there's this thing called zooming which makes perfectly sense to record picture at higher resolution. Because looking at details is simply looking at the same picture at another scale.
The "visual equivalent" to 192kHz sounds would be recording colours outside the human range. Like recording also infra-reds, microwaves, ultraviolets, and X-Rays.
Things that can't never been seen, because human lack the corresponding apparatus. The only way to get someting out of this extra data would be to transpose it into the visible domain. Thus use pseudo-colours to display levels of low infrared (heat), etc.
Just like the "zoologist" use-case above, there are a lot of scientific use-case where that could actually make sense (as an exemple, think about all the data collected by astronomers).
But in no way is it useful to record X-Rays to enjoy a painting by some known artist. The painting was done by a human painter, for human public, using colours chosen for their effect on an un-aided human visual system, disposed on a canvas in a way which is pleasing to the eyes.
(Well, okay. I know that some scientist use infra-red or X-ray image of paintings to analyse how they were done, what are the layers underneath or if there's even another picture over which the current one was painted. But these are scientist analysing the paint, so we're agin on the "scientific analysis" use-case).
24/192 makes sense as an intermediate format to avoid rounding errors, aliasing during filtering, etc.
There could be also some scientific value to keeping
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
What's pointless is any further debate about moving to 20MHz samples at 64 bits when music distribution has a much more serious (and actually real) problem. So much of our music is being destroyed beyond recovery before it even leaves the production desk.
No music I produce will succumb to this trick, ever. Perhaps that's why I don't get as much radio play these days.
"Nine times out of ten, starting a fire is not the best way to solve the problem." - my wife
Well lies are not the opposite of truth. But truth is the opposite of lies.
Lies are intentional falsification of the truth. But you can stick to believing non truths as truths and not really know that we are spreading falsehoods thus we are not lieing.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
In a perfect world, 16bit/44.1kHz might be enough. But we are living in a real world, that means that we and electronic devices are not perfect in implementing 16bit/44.1kHz audio. For example, we do not have a perfect electronic brick wall low pass filter. For mixing and mastering a recording, it is much better to have higher resolution materials than 16bit/44.1kHz. In fact, almost every audio studio does that in higher resolution all the time.
As for 192kHz -- it's not going to make anything worse, but it's not going to make anything better either
According to the article, it can make things worse.
Real-world playback hardware can make distortions.
A 96kHz sound is inaudible (for humans, at least).
But a 96kHz signal thrown on a real-world speaker might get distorted. And some of the these distortions can end up in the audible spectrum.
So instead of hearing nothing, you end-up hearing noises caused by something which shouldn't be heared and thus has nothing to do here in the first place.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
When human perception is fine with 8bits per channel?
Because when you fiddle with the signal you need that invisible signal to retain your fidelity.
Reasons for audio fiddling include:
1) Sound balance within the room
2) Sound quality for listener presence
3) Sound quality for room acoustics
4) Sound quality for different speaker systems (including headphone)
And that sampling frequency only gives you the correct frequency replication up to the Nyquist limit. It doesn't replicate phase or amplitude correctly, you need oversampled source for that. To get the high C of a flute to sound different from the high C of a piccolo, you need to include more than just a sample at twice the frequency, since the overtones are at different apmlitudes compared to the main note.
So you do need 92kHz sampling. Or limit your ability to distinguish real-life instruments with a main frequency over 11kHz.
You got marked flamebait and yet I can prove the same thing double-blind using Blu-Rays and uncompressed audio as wel {...} I've flipped between audio inputs for several people while watching movies without telling them
No sorry. That's single-blind. They don't know it (they are blind), but you (the experimenter) are doing the flipping so you know (you're not blind).
Double blind would be giving both sample to a machine choosing randomly which signal to produce (A-B-X tests for example. You, the experimenter, give 2 samples to a machine. The machine plays A, then B, then chooses one of the two randomly and the audience has to pick up if it was A or B. Neither you or they know it).
Also, you're home made experiment fail to take into account:
- The switching between the 2 source is audible because the equipment switchs modes.
- There's no guarantee that the sound recorded in the 2 sources is exactly the same. Specially regarding the volume. Our brains are wired in a way that we think that anything louder is always better. If the 24/96 track is a few fractions of dB louder, the audience will find it inherently better.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
No, there is no audible difference between 44.1kHz and 192kHz if all you want to do is listen. However, if the intent is to do any post-production work, re-mixing, mash-ups, whatever - then the quality makes a big difference.
Try running time-shifting or pitch-bending (not dumb-resample where time and pitch both change), and I assure you, you'll get much cleaner results starting with the 192kHz file.
A single photo receptor might not be able to see a transition shorter than X ms.
BUT
Your eyes and your head move around. Or objects themeselves can move around.
- Have a laser pointer.
- Have the laser light blinking, even at some ridiculously fast rate (200Hz).
- Move the laser point around, fast enough.
- You'll get the impression of a dottet line, not the impression of a moving point.
Your retina can notice things blinking at more than 200fps, even if single receptors can, just due to the relative momtion of the object inside the field of view.
Hearing frenquency range is fixed (well, mostly. I know /.ers can think of corner case, like when doppler effect comes into play). Your ear hears noises up to ~20kHz and nothing beyond due to physics and mechanical constrains. A 30kHz sound will always be a 30kHz sound (well minus the doppler corner case) and will never be heard. A 5kHz is a 5kHz sound no matter what and should be heard by anyone with an ear still able to detect 5kHz noises.
The video equivalent of this isn't the FPS question, but the wavelenght. An eye can only see visible light. You cannot see deep IR or microwave, nor can you see high -UV or X-rays (well again, corner cases: the repectors in the retina *should* be able to detect some near UV light, but the eye len blocks this light. And rightly so, because otherwise the UV will fry the retina. But some people with replaced artificial lens could see a little bit of UV).
insisting that 192kHz sampling is better, is like insisting that you need to be able to record from microwaves all the way up to X-rays in order to enjoy classical paints. sorry, no. You won't be able to see any difference in a reproduction of Monnet with and without the x-rays.
The fps situation is closer to the problem of number of speakers in a positionnal audio system.
In theory we have only 2 ears and can should only need 2 channels.
In practice humans move their head around. For a 2 channel audio to be positionnally perfect, you would need to track the motion of the head and vary the channels accordingly.
It's simply cheaper and easier to put a greater number of speaker and channels, and let the ears hear the difference caused by the motion of the head.
even if it's an technnical overkill, it's simpler that way.
For the same reason (specially with older CRT which could actually output it) its simpler to output at 150fps, rather than try to deal with and compensate for artifacts due to thing moving in the field of view at 30fps.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
What are the bitrates/quality of the audiotracks that are sold online today? thnx
afaik, with my midclass Sennheiser phones I've heard quite a difference between 192kbps and 320kbs (i know you talk extreme quality at the moment, not the strong compression on CD quality)
Only if your definition of "perfectly good" is "so convoluted that nobody EVER uses it". ;)
Let's be honest here, verisimilitude exhibits a superlative and ostentatious preponderance of syllables.
"Mind, as manifested by the capacity to make choices, is to some extent present in every electron." -Freeman Dyson
Seems you got lucky with your onboard audio. My experience with onboard audio over the last three mainboards is as follows:
-Abit IC-7 from 2004: Lots of background noise. Scrolling the screen was audible as crosstalk on the headphones. Buying a 20 Euro Soundblaster Live (PCI) was quite an improvement.
-Asus M2N from 2007: Supposedly 24 bit high definition, which I don't quite buy in terms of actual quality. But good enough that I didn't bother to get a discrete sound card for this PC.
-Asus M4A78LT from 2011: OK (but not great) with walkman headphones at low volume. Unable to provide more than low volume to said headphones without clipping. Upgraded that one with an old Soundblaster Audigy I picked from someone else's discarded PC. Sound quality improved at all volumes and high volumes were now possible, as opposed to the onboard audio.
C - the footgun of programming languages
The way to go is to use lossy compression formats based on 24 bit raw data with at least 96kHz sample rate.
Reducing the file size drastically from that starting point is possible without any reduction in perceived quality. But doing that by the way the CD does (e.g. removing half the samples and cutting of the lower bits) does a really bad job of distributing the error.
Especially a dynamic of more than 16 bits is important for classical music or movie audio tracks. If you have a 60dB dynamic in a track, the silent parts will be quantized to 6 bits on a CD. A dolby audio stream will at a medium data rate will have much better signal quality than the CD in cases like that.
Of course the worst thing to do is to convert it to CD format first and then add lossy compression later, as you get the worst of both worlds.
He claims: "No peer-reviewed paper that has stood the test of time disagrees substantially with these results."
How about this one?
http://www.physics.sc.edu/~kunchur/Acoustics-papers.htm
My feeling is, I *know* I hate the sound of dither. And I *know* I hate the flat sound of stuff missing. And I *know* I hate the audiophile super-precise 16- and 32-bit mods/chiptunes and so on, even when they're made by producers with huge experience in audio processing and studio work and... musical theory, and so on. And those digital songs are created by people who should, optimally, be producing the best possible stuff to listen to. But instead the only people who like it are apparently called Seapunks.
Where do I stand? I don't really know. I don't care so much as long as the song I'm listening to sounds as good as an analog recording. In my experience, that happens around 24 bit, 192khz. I don't know *precisely* where it happens, but I know the next step down the digital compression staircase, (164 isn't it? I don't remember) has noticeable losses, and 128 is intolerable for most music. And we're talking about, hmm, almost doubling in size. And we're still talking about megabytes, not anything huge. So I make the sacrifice, and I don't hear any of the things I hate: *dither*; digital conch-shell effect (great now I sound like a Seapunk); "something's not there"; no bass; none of the high-end distortions or hisses I know should be there from experience listening to that synthesizer; etc.
The author gripes a little about "training" the ear making people think they have better hearing. He also goes on about how the wider range is needed in the studio to have more room to work in, but from experience I know if you screw up a recording, once two layers of sound are mixed you aren't going to take that mix and magically move one of them around without also moving the other. But he's talking about side-effects and so on. What it sounds like to me is he's saying "well, if everybody was using transparent oversample filters both in the analog-to-digital and digital-to-analog transition, and if everybody had really fine and precise playback and speaker equipment, and if everybody was a perfect sound engineer and producer and everybody was a perfectly trained listener, there'd be no reason to go to 24 bit 192khz."
And yet there are all these little indications along the way of how the wider range and higher frequency are useful for correcting errors. So it sort of dawns on me, he's asking for *more* effort out of the world in order to justify staying at a width and frequency that have *less* to offer, and his major argument is the amount of space it will all take up. So it sort of fails Occam's razor in a way.
So am I wrong about my reasons to keep using 24 bit 192 khz? I've been doing that for years, and I only go into all this because people are starting to ask questions. Like the other day I was reading an article that bewailed our fates at the hands of "all these people who are producing music for the iPod-headphone crowd".
I had to stop, like, wtf? What's an iPod headphone got to do with it? Then I realized, I make music for the JVC marshmallow earbud. The original ones that still cost around $20, not the new trashy model (which I have, now, and which I hate) that only cost $14. Am I some kind of culprit of some kind of some shit or other? What am I doing wrong? I mean, arguably, my digital tracks are equally for people who buy really low-range response giant speakers for their cars, and I do that on purpose because it's funny. So I have a reason.
But where are all these people suddenly coming from who have these really huge bones to pick with entire industries and crap? What does it all MEAN?
((If you wonder what I'm talking about when I mention 16/32-bit mod music.... fine, if you want to force me to do it, there's a bunch of stuff you could dredge up from the 90s but here's basically THE top result for searching for such stuff: http://modarchive.org/index.php?request=view_by_moduleid&query=34414 ... if you want me to rip my own dick off, force me to listen to that cymbal crash on constant repeat))
"Stratigraphically the origin of agriculture and thermonuclear destruction will appear essentially simultaneous" -- Lee
Even the best D/A converters are inaccurate beyond 11 or 12 bits. That bascially means that 16 bit lossless is really 11-16 bit lossless.
Just go to a good headphone forum and you can quickly grab a $80 pair of headphones that are insanely good. Just don't buy those POS Dr Dre things that are worth $20.
1) 192 khz sampling -> up to 96 khz frequency.
2) Subliminal advertising for cats !!
3) PROFIT !!!
How about this one? http://www.physics.sc.edu/~kunchur/Acoustics-papers.htm
Abstract is:
"Many misconceptions and mysteries surround the perception and reproduction of musical sounds. Specifications such as frequency response and certain common distortions provide an inadequate indication of the sound quality, whereas accuracy in the time domain is known to significantly influence audio transparency. While the upper frequency cutoff of human hearing is around 18 kHz (or even lower in older individuals) a much higher bandwidth and temporal resolution can influence the perception of sound. Non-linearities and temporal complexities in the auditory system negate the simple f ~ 1/t reciprocal relationship between frequency and time. In our group's research -- which lies at the intersection of psychophysics, human hearing, and high-end audio -- we measure the limits of human hearing and relate them to the neurophysiology of the auditory system. These experiments also help to define the criteria for perfect fidelity in a sound-reproduction system. Our recent behavioral studies on human subjects proved that humans can discern timing alterations on a 5 microsecond time scale, indicating that that digital sampling rates used in common consumer audio (such as CD) are insufficient for fully preserving transparency."
Hey! You missed 'mick'. This is discrimination! Also I think you meant 'polack', you merkin wanker.
Even if the human ear could not tell the difference at normal playback, a higher rate will allow it to be played back slower with a quality the human ear still can't detect. This is important if you do a lot of editing.
You can see this in high speed video, when it gets played back at a slower rate than recorded, it still seems very smooth, but if you slow down something recorded at the normal rate, it is clearly not as smooth; Audio works the same way.
If you are plannign to do any kind of audio editing it is even more important to get it in a higher rate format.
You willfully leave out nerds, geeks, dorks, and spazzes? Obvious /. bias! ;)
I8-D
Damn, I hate getting to these threads late, especially when it's a subject that interest me so much. Always some clown with an offtopic first post (modded up of course) followed by an answer to the offtopic post that's modded offtopic when it isn't. I'd have to wade through hundreds of responses to find any real insight or information.
TFA is exactly right and exactly wrong.
If you're listening to modern, popular music, a 16 bit sample is more than sufficient, because popular music has no dynamics. Even when they digitize the old analog music that was engineered to give the best dynamics physics would allow the medium to have (think Boston's first album) they compress the dynamics to make it "loud." I mention Boston because the band's leader was really pissed off at how bad the CD sounded.
But if you're listening to classical, with its very soft passages, loud passages, and especially when there are cannons in the recording, you want as large a dynamic range as you can get -- and with digital sampling, that means as high a bit rate as you can get. The very soft (compared to the loudest) sounds will have the same as an eight bit rate or lower -- the highest crest of these waves will take fewer than eight bits to render.
As to sampling rate, that depends on your output transducers, whether speakers or headphones. If you have a boom-box type setup with a four inch midrange and a subwoofer (most common these days), the sampling rate doesn't matter much because your speakers aren't going to be able to accurately reproduce the 15+kHz tones accurately anyway. However, if you have good (read: expensive) speakers, with each one having say an eighteen inch woofer, two midrange drivers (squawkers) of different sizes, a good tweeeter that will go up to nearly 20 kHz and what they used to call a "supertweeter" with a range of 17-30kHz, those expensive speakers are wasted on a 44k sample rate.
At that sample rate a 15kHz tone has only three samples. With only three samples there's no way to accurately draw the waveform. With three samples there's no way to discern between a sine wave, a square wave, or a sawtooth wave.
We now return you to your regularly scheduled offtopic jokefest.
Free Martian Whores!
I will take truthiness over the mental masturbation that is this article. The sampling rate should be adjusted for each and every track. But putting the idea out there that its a crappy choice is a lie too. I will dump compression any day for the original WAV. Then this argument truly is utterly pointless.
After reading a considerable amount of this growing debate, I have this to address to the people who staunchly support the article's premise:
I get tired of all this "probably" assumption. How probable is it that everybody in the world is going to grab the best possible equipment for recording, conversion, amplification, reconversion and playback and make sure the entire chain from creation in studio to recording to distribution to downloading decompressing and playback is going to involve all of this fucking equipment and that everybody's going to use it properly? Give! Fucking! Up! You fucking... all you autistic chart-wizards make LESS sense than the people you accuse of being fucking "audiophiles"! Your ear, for example, isn't a fucking test tube with a formula written on it! People like you remind me of this one "mentally superior" moron who really did think that a circle was just a 360-sided polygon. You'll cite all your expertise, but just listen to the shit music that gets recorded in 16 bit and 44.1 khz: it's a bunch of fucking chiptunes and weird ass math-audiophile .MOD tunes from the 90s, that sound like exquisite dogturd. Frankly, I'd rather have this hugely "unnecessary" range, frequency and sampling rate that do nothing but TAKE UP A FEW MORE MEGABYTES, and listen to the world's IMPERFECTLY recorded music produced on ANALOG instruments and catch all those imperfections than worry about the seemingly autistic insistences of a handful of overanalysers like your camp.
"Stratigraphically the origin of agriculture and thermonuclear destruction will appear essentially simultaneous" -- Lee
I wish you guys would get this right. There is absolutely no way you can tell the difference between a 15kHz sine wave, square wave, or sawtooth wave (apart from amplitude, perhaps).
Sawtooth waves have even and odd harmonics, and square waves only have odd ones. This means that the first harmonic of a 15kHz sawtooth wave would be at 30kHz, and the square's 3rd harmonic would be at 45kHz. As you pointed out, even if you could hear them, you'd have to have damn good speakers to reproduce.
Three samples is enough to reproduce the 15kHz fundamental per Nyquist.
The main article cites this: "Audibility of a CD-Standard A/D/A Loop Inserted into High-Resolution Audio Playback," as a representative confirming source that there is essentially no perceptual audio difference between CD and SACD bitrates. However, it is clearly not a scientifically done test nor are the authors in any way scientifically trained. Moreover, they ignored several important factors in doing listening tests.
First, they do not define what the expected outcome should be, that is, they seem to state it as 50% from
"were the same as chance:49.82%"
but then in the next paragraph they start making comments like
"Females got 18 in 48, for 37.5% correct"
If women are getting 37.5% correct when the statistical expected outcome is 50%, then there is a correlation.
If the expected outcome is 37% then the authors should have explained the reasoning for different criteria.
Second, one of the most important factors in listening tests is whether or not the music was very familiar to the human testers. For example, I might be able to pick a particular version of the "1812 overture" from 16 vs 24 bit but I would not be able to do that for "Eminem" - When I know a particular piece well, my ability to discern differences increases. The article did not mention this at all.
Lots more anyway the upshot is just that it is not a scientifically done test nor writeup.
For most people, there is no place where sounds above 20 kHz will irritate a nerve ending enough to send an impulse to your brain. Thus, no sound higher than 20 kHz is audible, and 20 kHz corresponds to a 40 kHz sampling rate. (One sample at the low point on the wave, the next sample at the next high point, etc.
The problem in your analysis is that a "sound higher than 20 kHz" may be inaudible, in the sense that you don't detect a sustained sine wave at such a high frequency. But the Nyquist theorem applies to Fourier components---infinitely long unmodulated sine waves---rather than intuitive "sounds." Modulated sine waves at audible frequencies have Fourier components above audible frequencies with audible effects on the modulation.
Mike O'Donnell http://people.cs.uchicago.edu/~odonnell/
It almost feels too easy doing this, like beating a 5 year old at chess but..
U mad bro?
http://en.wikipedia.org/wiki/Factoid
"A factoid is a questionable or spurious (unverified, false, or fabricated) statement presented as a fact, but with no veracity."
we had some costumer that we had to put on satellite but that is extreme cases (people living over 8KM cable distance from the co servicing the line or with extremely bad cables)
Long distances to the DSLAM and undermaintained cables are the reality in the more thinly populated parts of the United States.
Is the 5GB/month the absolute max or the one 80+% of people chose as the next tier is significantly more expensive?
The latter. Providers of big downloads or streams have to plan for the tier that customers actually have. But because agricultural technology has shrunk the fraction of people who need to live in rural areas to grow food, providers of big downloads or streams appear to ignore satellite users and target urban and suburban demographics, optimizing their PC- and TV-targeted offerings for DSL, cable, and fiber.
Years ago, SUN microsystems promoted a nonlinear quantization, called "mu-law." A key problem is that the nonlinear function has to be applied to the sum of many frequency components, so it causes cross-modulations between them. A particular example: a low amplitude high frequency signal component may appear and disappear as a high amplitude low frequency component varies between 0 and its maximum. Since a high frequency component is much louder than a low frequency component of the same amplitude, the effect can be quite dramatic.
Mike O'Donnell http://people.cs.uchicago.edu/~odonnell/
Maybe its just me or perhaps the Canadian disposition, but I don't think Canuck is really all that offensive (compared to some listed).
I can think of a few more not worth mentioning. There is a few on the list I have heard of, but really don't know what they mean, which I am OK with really. Some of the war ones, seem quite mundane, though perhaps they started out as code or something, like Jerry or Charlie, etc... Which actually reminds me of The Cryptonomicon and using the word nip, as a shortened Nipppon.
It seems many slurs probably came out of wars, I wonder how many were specifically contrived purposely to try and dehumanize a group simply to make it easier psychologically for soldiers to kill them. Which really if you think about it, makes it even more offensive to use such language. Anyway as my grandma told me, sticks and stones may break my bones, but names can never hurt me.
You must not be using Monster uranium tipped, cables with platinum mesh shielding. The casing is made up ground up of unicorn hooves, and leprechaun tears. A Native American Indian shaman then did a special secret ceremony than imbues the cable with special supernatural powers.
I can go way beyond the mere mortals 192khz, 320khz is the absolute lowest that I use. Only my cables let me fit that large a sound file down it, as the fatter the file, the more cable you need!
Most of the comments here are covered in the fine article. It makes a hell of a lot more sense than most of the magical thinking being espoused in these comments. Most of you are regurgitating the same myths that the article dispels in detail. You just end up looking stupid posting comments on an article you clearly did not read.
The conclusion "lossless formats and a decent pair of headphones will do a lot more for your audio enjoyment than 24/192 recordings" does not prove the headline "24-bit/192kHz Downloads Is Pointless"
The article points to the visible light spectrum as analogous to the audio spectrum. This makes more clear the faulty reasoning. Light is not sound. Light is quantum at its source not analog. The best analogy for quantum (to explain that mysterious atomic effect in human perception terms) is.... digital! Analog does not resemble it so much.
What is "Fluorescence"? How do overtones and ultrasonic noises interact with audible noises? The answer is simply: They are. Do we understand it fully? No. Sorry, but that is not what Science means. If you think that Science means we know all these things for certain then you don't understand the word science. For you science has become a religion of certainty and false security
What organ causes hearing and where is "sound" created? The Brain.
To me this argument is like a teenager trying to say that only certain drugs will get you legitimately high. That someone 'couldn't really" have narcotic effects in their brain because they weren't using the "real" stuff. (Near beer vs real beer, or lotsa vodka vs only beer...) But one is conflating the mechanism with the final effect. If the final effect is psychological then ALL TRICKS TO ACHIEVE THAT END ARE VALID. Yes, there a physical limits that are known and are important and it is important to debunk pseudoscience that is glossing over that stuff. However, sometimes that is simply not the important question. Sometimes the important human effect is not in the realm of the known or is in the realm of the psychologically subjective. In that case, don't discredit the whole branch of human knowledge called science by applying it to something for which it is not suited. Like the question of "what kind of art is scientifically best" you are getting into angels on a pinhead territory and then look to the company you keep.
Stupidity is its own reward.
The author may be correct that 24/192 offers no advantages, he is wrong in saying that it is slightly worse.
While he's correct that frequencies much greater than >20khz can cause problems downstream, the problem is at least as bad with 16/44. The sampling theorem says that 44khz sampling is enough to correctly reproduce frequencies in the audio range (half the sample frequency). However, 16/44 reconstruction requires prefiltering and I believe can also introduce spurious high frequency components (the Nyquist theorem says nothing about frequencies higher than half the sample rate), so a brick wall filter is needed to remove any frequencies over 22khz, so you need to filter out high frequencies in either case. At least with a higher sample frequency you can use a more gradual filter, which is better in theory (though probably no different in practice). In particular, 24/192 will not sound any worse than 16/44
Such an elementary error calls the value of the whole article into doubt
There are reasons why this bitrate and sampling frequency are used and it can be heard. It is not futile and it is not just big numbers for the sake of it. I speak as a technical director for an AV company and as man who built several home and project studios and was part of 3 major studios migration from analog to digital technologies. I once was a teacher and technical supervisor in a sound design school.
24bits:
-you can and you WILL hear the difference with 16bits. It basically record finer amplitude variations than 16bits, therefore the dynamic range is increased and there is less approximation of values when the sound is digitized. the end result is that stereo spacialization is usually better as the right and left channel amplitude differences are closer to reality and very fine variation will lead to audible different result. Less quatization noise is heard (the low amplitude pop corn noise and "8bit feel" you get when listening to low amplitude digital recording) as lower amplitude values are represented with more bits and therefore are less coarse. More importantly any processing playing with amplitude is rendered much more accurately with finer detail. A digital compressor/limiter won't screw you stereo image, an expander won't bring more distortion to your mix by amplifying quantization noize for example. Echos and especially reverb will be MUCH finer and accurate, the tails won't cut off and won't sound like white noise.
192KHz:
- this one is tricky as it has a lot of use in the studio but it can barely be heard even on the best systems. Basically pretty much all AD/DA system uses brickwall filters to filter frequencies above 20KHz, the limit of a very healthy ear, so as to prevent foldback frequencies. The higher the sampling frequency the softer the slope of that filter is because the foldback won't happen until 96KHz is reached compared to 22KHz on 44.1KHz. the brickwall filter at 44.1KHz is harsh and many people with good sound system were complaining (me included) that it could be heard and was annoying. At 192KHz it is softer, enough to not be a disturbance. On the other hand the most important reason and use for 192KHz is latency. When recording someone in the digital world you have to deal with the fact that as a certain number of samples will have to be created before what goes in, goes out, at 44.1KHz there was an audible, annoying delay, if audio was processed it was unlivable for most musician. at 192KHz this delay is essentially eliminated and only the most discerning musician will be annoyed by it. So in that sense 192KHz is not really needed for most people and indeed very few people have systems that will indeed let them hear the difference with 44.1KHz but it is there.
I guess we all like to believe there is a big evil industry in all domain that make us buy stuff we don't need, I like to believe my i5 750 is as good a an i7 960 for what I do but the reality is the i7 960 IS better and with the right application the difference means a lot. Same goes for cars, a 2001 Toyota echo will get me around but a more expensive cars will get me around in more comfort will less issues. Same goes for audio, most people using gaming headphone or 5.1 gaming audio setup and cheap all-in-one sound system will never ever hear the difference between 16bit 44.1KHz and 24bit 192KHz but it doesn't mean it is not there and it doesn't mean it is not significant and that it's a lie. It might not be a necessity but for professionals like me (as in "it is how a make a living" not as in "I am an expert, listen to me") it is significant. For audiophile it is significant also and for people who listen to music all day (ear fatigue will come in much later with 24bit 192KHz than with 16bits 44.1KHz).
You do realise that the phase is already smeared by the microphone, the preamplifier and probably a channel EQ? To say nothing of room acoustics in the playback environment.
The majority of vinyl cut since the mid '70s utilised a digital delay. I'm sure the surface noise and scratches really benefitted from these speakers though, especially records cut on lathes using non-oversampling 32kHz 12 bit delays. As for the phase accuracy of consumer grade dolby decoding... HA!
I sent this article to my friend, who owns a recording studio, and has been doing Audio Engineering for 15+ years. This is his response by text:
I find a couple problems with this article. For one: hardly any studio records at 24/192. 24/48 is more common along with 24/96. Recording at a higher sample rate decreases the risk of aliasing as they have pointed out but then they say to use over sampling to fix this problem. Which is somewhat right, but a lot of plugins being made namely compressors introduce a shit ton of aliasing that if recorded at a higher sample rate wouldn't affect it as bad. Over sampling helps this if the plugin is designed to have over sampling. Most do not. The benefits of recording at 24/192 is mainly for people who work in film. When using pitch and time manipulation the quality of the sound has less artifacts when using lower sample rates. Still, it would be better to not have to convert sample rate and bit rate down to consumer level. Think of the sound of a WAV compared to an mp3. But also brings me to my last gripe: Consumer products such as cd players, iPhones, stereos and such have very crappy Digital-to-Analog converters which would defeat the whole purpose of having higher quality audio anyway. The converters in these devices are probably worth a dollar or so. It would be like running a blu-ray movie thru a 1960s television. No point. Although the article brought up good points, overall I would have to disagree with it and find their argument full of shit. My 2 cents. Lol
Judging by the modding I guess Louis C.K got mod points.
"If you are going through hell, keep going." - Winston Churchill
I'm a proud Hillbilly, lived my whole life in the WVa hills, even managed to have a career as a software developer, only moved out while I was in the service/drafted.
Not that we don't like NYC, Caribbean Islands, the different hills and mountains in Colorado, WY, AZ, NM, etc.
But you can call me Hillbilly and be accurate. I think it's illegal to discriminate against Hillbillies in Cincinnati, where lots of us have gone looking for good jobs.
Think of the Irony!
Guess how long ago 1992 was? That's not exactly a gross overstatement - rounding off by 2 years?
missed crackers
The post walks us through some of the basics of ear anatomy, sampling rates, and listening tests, finally concluding that lossless formats and a decent pair of headphones will do a lot more for your audio enjoyment than 24/192 recordings.
Right--because if *you* can't see a use for it at the moment, there must not be one... ...or maybe you're just plain wrong.
There's no place like
Thank you. You took the words right out of my mouth. That statement (about 15 KHz sine, square, sawtooth) perfectly summed up how poorly the poster understands digital signal theory.
I looooove the melodious distortion I get from using single ended WE 300Bs as my output stage. Ella never sounded better!
Your hard drives last longer than a year?
Just because it CAN be done, doesn't mean it should!
Wouldn't this only be an issue with processing? I was talking about encoding for storage and transmission only.
"Politicians and diapers must be changed often, and for the same reason."
Umm, TFA talks about 16 vs 24 bit and addresses the question of dynamic range. It's the mastering that's the problem, not inherent issues with 16 bit audio. Somebody else already addressed your comment about sample rates. Did you actually read TFA?
and a "regular" CD.
But we have paragon amps and proper monitors.
And a proper listening space.
The assumption that it's just your ears that contribute to perception of sound, the assumption that people only perceive sound up to 20kHz, these and other assumptions and statements made by so many self proclaimed experts here are demonstrably incorrect.
You'll never really experience the kind of audio reproduction that is possible with $15k worth of high end audio equipment, and that's fine, it's not something that's "worth it" to you. I'm sure most posters don't even have a proper place TO listen to high-end audio. It's not for everybody. It isn't being a snob, it's just an interest you don't share, or really know very much about.
the interwebs are srs bsns
Well... no. There's enough information in a low sampled curve. As TFA explains it, the output isn't "jagged" when played back in analog.
No, but it is *aliased*. The waveform between two samples is a simple interpolation. It is probably pretty close to the original sound, but there will always be some error too.
Simple math problem:
- take this *aliased* waveform. (the result of a join-the-dot interpolation).
- compute the "error" (i.e.: substract the original perfect waveform from what you consider an aliased thing)
- do a fourrier transform on this error (i.e: look at the harmonics).
- all the frequencies which compose the error will be above the audible frequency range
- i.e.: you won't be able to hear the difference. i.e.: the aliasing isn't audible
that means your ears don't give a damn fuck about the aliasing.
And that's using the "join-the-dot" misconception, which doesn't even exist when playing back on real-world equipment.
Linear interpolation (actual "join-the-dot") did make problems back in the module-tracker era, when 8kHz instruments samples were interpolated into a 44.1kHz soundcard output.
because then, some of the "error" was in the audible range (4kHz to 20kHz).
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
I think most of the folks in this argument aren't realizing they are arguing about the wrong thing.
My "CD Quality" 192 kbps MP3 rips are ripped at 44kHz. 192 kHz in this context IS overkill for any human.
Wouldn't this only be an issue with processing? I was talking about encoding for storage and transmission only.
No, this is precisely a problem for playback. With sublinear encodings, there is no way to present a low amplitude component of a sound accurately in the presence of a high amplitude component. Since perception is sensitive to components at different frequencies fairly independently, the loss of accuracy in the smaller component can be quite perceptible.
In fact, nonlinear representations, such as floating point and phasor representations, are often good for certain parts of processing, but not for playback.
Mike O'Donnell http://people.cs.uchicago.edu/~odonnell/
You're actually wrong. Human ears are relatively good at hearing phase relationships and volume relationships between sounds, as these are key components in determining a sound's direction. Thus, even though you cannot hear the fact that it has turned into a sawtooth wave, you can at least potentially hear that the peak is at the wrong point in time, and you can almost certainly hear that the amplitude is reduced inconsistently from wave to wave.
This paper is also wrong in its claim that 20 kHz is "generous". It isn't. I've done listening tests and have successfully heard high-pitched whines up to... it was either 22 or 23 kHz (which was where I stopped trying, not where I stopped being able to hear), and I'm not even all that young. Admittedly, this is at relatively high amplitude, but the notion that most people can't hear 20 kHz is just plain wrong, and if you start out with that fundamentally wrong premise, you pretty much have to question all the other assumptions, too.
They also make the fundamentally incorrect claim that everything below the nyquist limit is sampled perfectly. This is also provably and trivially false. The Nyquist theorem says no such thing. It merely says that signals above that limit will result in "folding", causing aliased frequencies below the limit, which means that any frequency below the Nyquist limit can be captured without aliasing. However, music is not a single frequency in isolation; it is a bunch of frequencies interacting in complex ways. The Nyquist theorem says nothing about the phase of a signal near the Nyquist limit being consistent relative to other signals at lower frequencies, and in fact, it is not. Nor does the Nyquist theorem state that the frequency will be captured in a way that maintains consistent amplitude as you approach the limit; indeed, it isn't.
Read the Wikipedia article about the Kell factor in display technology, and you'll understand why this is a problem. Notice that with display technology, there is no anti-aliasing filtering involved (because the signal is a known signal that is entirely below the Nyquist limit), so this roughly maps onto what would happen if you could magically create a perfect anti-aliasing filter on the input side. You don't become nearly artifact-free until the frequency you are sampling is about 2/3rds of the Nyquist limit. This is an indisputable fact.
Admittedly, these artifacts are less objectionable in audio because of the anti-aliasing filtering that occurs (both on input and output), but no filter can magically "fix" that inconsistent amplitude. It represents actual information loss—the signal is equally likely to be a constant 15 kHz tone with constant amplitude as it is to be a signal that varies on either side of 15 kHz with a variable amplitude—and once that precise phase and amplitude information is lost, it is impossible to definitively reconstruct it.
In other words, this article is just plain wrong, almost top to bottom.
Besides, the real question is not whether 44.1 kHz is "good enough". It provably isn't, if you care about faithful reproduction over the entire human hearing range. The question is whether the information in the top octave of human hearing is in any way useful or important, to which the answer is "probably not". That's not the same thing as saying that 44.1 kHz or even 48 kHz sampling rate faithfully reproduces the entire range of human hearing, though, but rather it is merely saying that most people don't care about its deficiencies. A 48 kHz sampling rate is "close enough" up to about 16 kHz, which is a broad enough frequency range to be "good enough" for all practical purposes.
Check out my sci-fi/humor trilogy at PatriotsBooks.
Truthiness... Like needing a gold plated HDMI cable because it is better than one without. People want to believe it, it sounds good, but it's absolutely false. Anyone that understands that HDMI is digital, it transmits 1's and 0's, will know that it isn’t dependent on the conductiveness or noncorrosive of a material. RF on the other hand flows over the surface of the metal without penetrating deep into the cable. Hence, you want a metal that is non-corrosive as well as good conductor. If you used something other than gold for RF you can run into a problem of the RF signal reflecting from impurities in the connector and having part of that signal travel back towards the cable. This in turn will cause interference in the wave and cause problems. This has absolutely nothing to do with how HDMI transmits data, which is in the cable. But people are still thinking that to watch TV you must need gold plated connectors because gold plated connectors were used before... It has to be worth that extra $20 to $1000, right?! Lol! Truthiness....
...I can see at least one bogosity and a couple of omissions. The author claims that the "phase doesn't matter" with the Nyquist criterion, when it can easily be shown that, for instance, sampling a 20KHz sign wave at exactly at 40KHz can result in a zero signal if the input and the sampling are synchronized such that the sampling points all occur as the input waveform crosses zero. If they're slightly out of sync, something will get through but it'll be greatly attenuated. More importantly is the issue of "aliasing"--if there's any component to the input that's of a higher frequency than the sampler, the digital result will contain a "difference" component somewhere in the audible spectrum. For an idea of what this might sound like, listen to Don Ellis playing his trumpet through a ring modulator at the beginning of "Hey Jude" from the "Live At Fillmore" album. In practice, the sampling rate is placed somewhat higher than the maximum input frequency, to compensate for the analog input filter's cut-off being less than perfect. The 44.1 KHz rate for CD audio was the lowest rate at the time that allowed the recording industry to be able to claim "high-fidelity" i. e. reproduction of a 20-20KHz bandwidth. 48KHz is probably safer. Admittedly 192KHz is overkill, but perhaps not for mastering, assuming the amount of post-processing that's likely to happen between the original recording and the listener. Typical "webcasting" software, for example, contains multiple layers of digital filters, compression and whatnot, so it helps to start with something that's not already compromised.
The sampling rate doesn't mean the signal you hear is "smoother." That claim is total garbage and shows immediately that you don't understand what analog-digital or digital-analog conversion is about.
A signal of finite duration can be expressed as the sum of sinusoids ("pure tones"). It takes only two pieces of information to reproduce a perfectly smooth sinusoid: its amplitude and its frequency. Sampling at discrete intervals gives us enough information to reproduce exactly all the sinusoids present in the signal up to half the sampling frequency. That's called the sampling theorem.
Furthermore, you quite assuredly do quite literally deceive yourself thinking you're hearing better sound from a 192 kHz file. This is no insult to you, nor am I saying you're being disingenuous with your claim; it's just that part of being human is that our cognitive biases are often stronger than our sensory perception. Do an ABX double-blind test between your 192kHz file and a version correctly downsampled to 44.1kHz and there's no way you'll tell the difference. Your ears are physically incapable of hearing any frequency anywhere close to the missing frequencies.
Please read the linked article, as Monty does a great job of explaining all of this and more.
My biggest problem with this is the fact that if i want to get a digital file from the Itunes store, I'm not going to get 16/44, I'm getting a lossy format, and If I rip a CD, I'll rip to mp3 at 320kbps. why? because I have the CD, I can put it in anytime I want, but my Iphone doesn't do FLAC
So anybody telling you that you don't need 24/96 or higher because we already have 16/44 is clearly in the way of people getting even that. I bet if audio DVD was ever the standard, we wouldn't get more music per disc, but higher quality music, and along with it higher quality equipment needed to play that music. I wouldn't be surprised if in this ideal scenario, 16/44 would be considered "good enough" for people.
So I'm all for progress, I'm sure today's music wont benefit from higher bit/sampling rates, but you never know what people in the future can do with it. there's been many other times when people have stood in the way of progress and declared the status quo "good enough"
They did this for a while. Maybe still? About 10-12 years ago. I forget what the market-speak trade name for it was. But they sampled at 1 Ghz. The trade magazines were divided in their opinions, and it must have died a fine death in the marketplace since no one here has referenced it This was definitely a recording format. .
I think we have to differentiate between mobile and home/studio listening. Considering the average playback hardware, for listening, 16/44 is fine for the 99.999 percent of listeners. For mobile listening 192kbps MP3 will exceed the needs of most. I prefer 320 kbps because it makes percussion sound better.
Most people listen in a car, bus, at work, on the job, etc. Low noise floor and dynamic range are moot. the reproduction amplifiers in cheap phone/pods aren't up to the task anyway, much less the average headphone/earbud.
I hesitate to use the term "audiophile" because of its pejorative connotation, but for people with above average sensitivity in hearing and training in sound artifacts, I think high resolution files are a good thing. Not only for private listening, but for a possible future when we regain a public domain and the remixing/sampling world takes off again.
For an analogy, think of DVD compilations of old TV shows that were encoded from tapes of television broadcasts. They look...ok...but when they go back to the original masters and re-release them there is an appreciable difference. Strangely though, consumer television/video playback formats are increasing in resolution, while common audio formats have been regressing.
The author of the article completely misses the secondary benefits of 24/192 delivery: Such fidelity (if delivered uncompressed) allows the recording to be considered archival and can be used for audio research, or as a historical record of the production methods used at the time the music was mastered. The format is indistinguishable from 24/48 *for most people*, obviously, but that doesn't make it worthless. If I could own 24/192 downmixes from the studio masters, I would consider it a unique opportunity.
"If the overtones of a flute high C and a piccolo high C are both under 22Khz, then sampling at twice that will catch all the overtones, and replaying the sample at the same rate will perfectly reproduce them."
Only possibly if the flutist plays forever - if the flutist ever stops playing, then the signal can't be both time-limited and band-limited.
Y so srs?
Yes, you can hear the difference in amplitude, but AFAIK that won't affect the music in a qualitative way.
And IIRC, human ears are not good at hearing phase differences unless the phase is changing. Again, you won't hear a qualitative difference between the fundamental of a square wave or sawtooth if you can't hear the harmonics.
You have a few inaccuracies in your post. No sampling rate will ever perfectly capture a square wave or sawtooth wave, unless you use the exact frequency of that wave (or multiples of that frequency), and happen to match the phase of the wave with the sampling point. So given your example of a 44k sample rate capturing a 15kHz square or sawtooth wave, you are correct that you can't reproduce the waveform from these samples. You can't perfectly reproduce those waveforms even if you used a 196kHz, or 196000kHz, sampling rate.
According to the Nyquist-Shannon sampling theorem, you can perfectly reproduce a sinusoidal waveform (phase and amplitude included), if the frequency components of that waveform are less than half the sampling rate. Therefore when we talk about sampling signals, we are always talking about sampling sinusoidal waveforms. Square or sawtooth waves cannot be sampled, because their sinusoidal frequency components extend to the infinite.
Hope that makes sense to you.
I can't comment too much on the sufficiency of 16-bit levels to a sample, but the article does say that the noise introduced at this level is below human hearing, so if correct, seems to me it'll do the job. That's 65536 different levels of amplitude. Should be enough to capture the quietest oboe and loudest trumpet, at the same time. If pop music recording studios are compressing the dynamics to the upper range of the bit level, that doesn't stop a classical recording studio from using the whole 16-bit range.
Hence, you want a metal that is non-corrosive
That is why your cheap "digital" USB cables have gold plated connectors.
Great spirits have always encountered violent opposition from mediocre minds -- Albert Einstein
When two or more instruments play a loud chord, the interference of the inaudible overtones from each instrument produce a distinct "ring" of audible difference tones, audible only at live gigs and on well reproduced SACD recordings. I've seldom heard the same effect to the same degree from a CD. Don't be fooled, this is a real and reliable enough effect for us classical musicians to use it to tune chords. This "ring" should be reproducible in 24/192 when these HF overtones in the stereo or surround channels interfere, which a CD cannot reproduce since there's nothing > 20kHz.
Granted, as mentioned in TA, the amp and speakers need to not be so rubbish as to introduce distortion > 20kHz.
Whilst I can tell in a blind test between the CD and SACD mode of the same disc of a recent BIS recording of Carmina Burana, it's only during certain passages of music where I am listening out for the difference tone "ring". Most of the rest of the time, I can't tell, and 16/44 CDs sound great. I don't think the fact that I am a classically trained musician matters.
That said, I think it's important NOT to be under the illusion that, just because you can't hear anything over 20 kHz (actually, ~16 kHz for most people), that there are no audible consequences when there is more than one channel.
In fact, given that well mastered vinyl played on good cartridges can reproduce fequencies to 60 kHz and beyond, this live "ring" may help explain why some folks still prefer vinyl recordings of classical music to the CD.
You can't perfectly sample any waveform at any frequency, but the more samples per crest, the more accurately the waveform will be reproduced. At CD sampling rates you can indeed reproduce a 300 Hz waveform of any shape very accurately; there are 146 samples in its crest. That's plenty to accurately describe a sawtooth or square wave with subaudible aliasing. Not so at 15kHz with only three samples.
According to the Nyquist-Shannon sampling theorem, you can perfectly reproduce a sinusoidal waveform (phase and amplitude included), if the frequency components of that waveform are less than half the sampling rate.
Remove the word "perfectly" and that is accurate.
If pop music recording studios are compressing the dynamics to the upper range of the bit level, that doesn't stop a classical recording studio from using the whole 16-bit range.
I didn't say that was the case. I said with pop music it doesn't matter since dynamics don't seem to matter any more in pop music. But if your cannon in the 1812 Overture are at the highest level and the soft flute is 1/100th of that, your flute only has a range of 0 to 500. That's only a few bits.
Free Martian Whores!
Quoting from TFA:
In our hypothetical Wide Spectrum Video craze, consider a fervent group of Spectrophiles who believe these limits aren't generous enough. They propose that video represent not only the visible spectrum, but also infrared and ultraviolet. Continuing the comparison, there's an even more hardcore [and proud of it!] faction that insists this expanded range is yet insufficient, and that video feels so much more natural when it also includes microwaves and some of the X-ray spectrum. To a Golden Eye, they insist, the difference is night and day!
Of course this is ludicrous.
No one can see X-rays (or infrared, or ultraviolet, or microwaves). It doesn't matter how much a person believes he can. Retinas simply don't have the sensory hardware.
I beg to differ.
As opposed to a lie that sounds like a lie? :/
I sell short stories in audio form (mp3) and I never sample over a configuration of a mono channel at 96 kbps with a sample rate of 44100 Hz. It works superb for me. I can't afford to waste web hosting space with 128 kbps.
For example, this one: http://sathyaish.net/stories/thelastleaf.aspx is at that configuration. It sounds great!
Also, all of these are mostly at 96 kbps with some exceptions at 128 kbps. http://sathyaish.net/voice
I don't see the point for general distribution. However, just because a human cannot hear the differences in an audio sample like this doesn't mean its not useful. If you process audio a lot, having more headroom results in fewer errors and side effects.
The same is true for my photographs. I record at far higher resolutions and colors than I really need, because I lose less when manipulating the images. Its not my final output that needs the headroom, its the steps before.
I expect that probably far fewer people manipulate audio, and that's the more accurate reason why a format like this has little value in the distribution case. There is a sweet spot somewhere that allows some fiddling without artifacts without also being too large to be generally useful, or so it seems to me.
I don't buy music online because the quality is so bad. If you know what it sounds like on CD and then you listen to in courtesy of an iTunes download, whole parts of the range and timber are just awol and it's all you can think about.
noise canceling headphones are a horror if what you're interested in is getting as close as you can to "being there". Ditto stuff like DOLBY. All those switches stay in the OFF position.
I have to RTFA, but in general i will testify that there are a lot of us out here who love CDs because of extremely high fidelity of the music and loathe iTunes and Amazon downloads because it sounds like shit. We have money to spend, but we're not spending it on that.
So if someone is trying to rectify this and sell to this part of the missing consumer base , all I can say is "of course I want all my music to be in lossless digital format, stored on some device that fits in my pocket and with my entire collection readily available to me at any time.
Why?
Is this going to happen any time soon? "
The point of this post is that distributing audio at 192/24 is pointless. This is correct. There are lots of reasons to record audio at high bit rates/depths. Among these is the fact the higher harmonics greatly affect processing and is big part f the reason many engineers will still run their audio through analog equipment for "that sound". In post production for tv and film high sample rates allow for greater flexibility in time stretching and pitch shifting. Also for archival purposes it is always better to have more and be able to give out only what is necessary. There are various tests with trained listeners being able to discern between high sample rate (96,192) and regular(44,48) sample rates. This was in a controlled in environment on proper equipment able to reproduce the high frequencies up to to 100k. As per the nyquist theorem digital audio limits the highest frequency to half of the sampling rate, thus the highest reproducible frequency with 192khz audio is a 96khz tone. This is nearly 4 times the highest frequency we can hear but the harmonics ainteract with the frequencies we can hear.
OK I see what he's saying and he's right. This is similar to the claims made for monstrously thick cables years ago. The claim then was the impedance of VERY thick cable , the resistance, was lower so more hi fi sound made it to the speaker.
The only problem was- no one could identify the difference in double blind studies. So much for that you might think but never think reality will interfere with marketing, and these fat cable companies are all doing OK even today.
Still doesn't make the schlock that Amazon and iTunes give you any better.
For anyone who isn't aware, the Hi Fi world is chock full of people who are basically insane and who not only will, but LOVE paying astronomical sums for any technology that promises higher hi fi. Thus the $150,000 home speakers and the $2500 cables and the ads that claim that their master craftsmen know *just exactly* how many times to wind some solid gold wire around some speaker part in order to get the highest high fidelity.
A similar situation exists with wines. Astronomical prices for nothing but the marketing around a bottle.
What can anyone say? Stupid people's money eventually drains away and the money of vain glorious stupid snobs gets hoovered out with an elephants trunk.
Who said humans were the only listeners? There's dolphins and dogs .. and of course, digital editing requires gross oversampling for frequency shifting, shortening or any other resampling technique. The fact is the sample rate is probably too low for that.
The assumption of the article is rubbish. It assumes the only processes involved are converting digital data to analogue and then human listening of the analog.
I heartily agree and have done the same. It was a revelation when I first got my Carver power amp and discovered with its power meters that most of the time it was pulling less than one watt per channel. Of course, I had reasonably sensitive speakers too, but nothing extravagant.
Read the Nyquist-Shannon sampling theorem. It's perfect reproduction. Mind you, there are caveats. Sample length has to be infinite. So, it's not practical, but what I said was actually perfectly accurate.
In any case, what I was trying to get at, is that sampling a sinusoidal waveform, even a 15kHz wave at a 44kHz sampling rate, reproduction is going to be very accurate. The mathematics show it. Certainly orders of magnitude more accurate than capturing a square or sawtooth.
If a soft flute has a range up to 500, that's still quite an accurate capture of amplitude. It's greater than 10, and even 11, so Nigel would approve. Besides, the practicality of listening to a cannon and soft flute, at their natural volume level, in the same piece, is rather bewildering.
But but stegasaurous err... stenography depends on it.
No one can hear 24 bit audio so the lower bits can be
written almost in the clear to pack a message down the
road.
The low bits are also magical and can be used for keys and other cryptography
values hither and thither....
Mind you, there are caveats. Sample length has to be infinite.
Exactly! Note that the closer you get to "infinite" the closer you get to "perfect". The higher the sampling rate, the closer to infinite and the closer to perfect.
In any case, what I was trying to get at, is that sampling a sinusoidal waveform, even a 15kHz wave at a 44kHz sampling rate, reproduction is going to be very accurate. The mathematics show it. Certainly orders of magnitude more accurate than capturing a square or sawtooth.
Yet there are more than sine waves in sound. A rock guitar fuzzbox changes the guitar's sine wave to a square or sawtooth (most fuzzboxes and wah wah pedals have a switch to select between square and sawtooth). With three samples (do the math!) it is impossible to discern those three entirely different waveforms. They will be distorted into a sine wave.
Have you ever studied sound with an oscilloscope? One of my undergrad physics classes was about this very thing, although it was in the late '70s and there were no digital samples back then.
Have you seen rock or blues bands with the guitar feeding into a small tube amp, with a mic in fron of it feeding a transistor amp? That's because if you overdrive a transistor amp to clipping levels, you get a perfectly square square wave, but with a tube amp the wave's corners are rounded (as seen in an oscilloscope) at clipping levels.
It is mathematically impossible to do that with three samples.
Besides, the practicality of listening to a cannon and soft flute, at their natural volume level, in the same piece, is rather bewildering.
Those pieces are ancient, and are usually performed outdoors. You would have to have an incredibly good setup to get anywhere close to accuracy with those pieces.
Free Martian Whores!
I beg to differ in this regard. "A Fourier analysis of a sine wave is the sine wave itself. A Fourier analysis of a square wave or saw-tooth wave shows harmonics and subharmonics." We can hear those subharmonics.
So, then the next thing to look at are the Fletcher - Munson curves. These curves are averages over a population, without stipulating, by frequency the standard deviation. While I can hear to 15780hz, my wife can hear to 16,200hz, and my father, to 12,500hz. What I have as a threshold, such as hearing the ticking of my watch, my father is unable to hear his watch, even when pressed against his ear.
FM curves should be broken down by age groups, with groups being 1 year apart for people over age 60.
No, I believe that anything over 60hz sampling is a waste of bandwidth. My high quality earphone diaphrams are resonant at 20 cycles, so in theory they should vibrate at 20khz, and they do, but with tremendous mechanical loss. And my ears as well have tremendous loss above 15khz. If I need to hear above 15khz, I should have electrical connections directly to nerve endings in my body, with the belief that nerves transmit messages at the speed of light.
Leslie Satenstein Montreal Quebec Canada
God, I wish you would shut up and go read something about acoustics and sampling theory.
Sample _length_, not sample rate. The longer you sample a sine wave, the closer to perfect you will be able to reproduce it, provided the sample rate is over twice as high as the frequency.
While I'm no expert on audio, only having studied undergrad signal theory for an electrical engineering degree, seems like the question here is: can the human ear discern between hearing a square or sawtooth wave, compared to hearing their sinusoidal waveforms bandwidth limited to the audible frequency range. If the answer is no, then distorting a square or sawtooth wave into sinusoidal components is not a problem. Hence there is no need to perfectly reproduce a square or sawtooth wave, because our ears would not be able to tell the difference.
Sample _length_, not sample rate.
Exactly what do you mean by "sample length?" If by it you mean that there are three samples in a 15kHz tone and hundreds in a 300Hz tone, then that is accurate. Your 300 Hz tone wil be more accurate than the sample of a 15kHz tone. But its is because of the number of samples collected per wavecrest.
Nyquist can be overly simplified to say that you need more than two samples to reproduce a wave.
can the human ear discern between hearing a square or sawtooth wave, compared to hearing their sinusoidal waveforms bandwidth limited to the audible frequency range.
That is exectly the right question, and the answer is a clear "yes". If you can hear a tone you can discern different wave shapes for that tone. It's the main reason people say that LPs sound "warmer" than CDs; it has to do with CD's aliasing distortion, which analog recordings don't have (even though there are other forms of distortion).
Raise the sample rate where there are enough samples to accurately render a 20kHz waveform of any shape and your digital sample will sound "warmer" than the LP while lacking the LP's inherent noise problems.
Nyquist doesn't apply to analog recordings because there are no samples per se, it is continuous. LPs had a fantastic frequency range. The way quadraphonic LPs worked was the rear channels were modulated with a 40kHz tone and added to the front channels, then subtracted on playback by phasing. That 40kHz tone that held the rear channels is twice as high as the best human ear can discern.
Free Martian Whores!
Sample _length_, not sample rate.
Exactly what do you mean by "sample length?" If by it you mean that there are three samples in a 15kHz tone and hundreds in a 300Hz tone, then that is accurate. Your 300 Hz tone wil be more accurate than the sample of a 15kHz tone. But its is because of the number of samples collected per wavecrest.
No, I think he's talking about the window of time over which the system is evaluated. IIRC, there's some stuff in Shannon-Nyquist about theoretically perfect reconstruction requiring an infinite time window (regardless of the number of samples taken per cycle).
That is exectly the right question, and the answer is a clear "yes". If you can hear a tone you can discern different wave shapes for that tone. It's the main reason people say that LPs sound "warmer" than CDs; it has to do with CD's aliasing distortion, which analog recordings don't have (even though there are other forms of distortion).
The reason (some) people like the vinyl sound is that it actually distorts the signal much, much more than a CD does, in a way which some find pleasant (it's responsible for the "warmth" you describe). The reason for this is easy to understand, if you spend some time investigating how LP recording actually works. Look up RIAA equalization sometime. The TLDR version is that just to play back a LP without having it sound like ass, the playback system has to implement circuitry which un-does some signal mangling done during LP mastering, and the un-mangle process is never perfect (and can't recover everything which was lost in mastering anyways).
Also, from your comments about CD distortion, you seem to actually believe in the "stairstep" mental model of sampling systems. They don't work like that. The stairsteps, or quantization noise, do not actually appear in the final output. Proper sampling and playback requires two key "brick wall" low-pass filters in the analog domain. One goes before the ADC, the other after the DAC. The purpose of the first filter is to remove all analog signal components which could cause "aliasing", i.e. those above 1/2 the sampling frequency (AKA the Nyquist frequency). The second filter is called a "reconstruction" filter, because it literally reconstructs the original bandlimited analog waveform from the stairstepped waveform. Basically, all the quantization noise is composed of frequencies greater than 1/2 the sampling frequency, so if you once again filter out everything above the Nyquist frequency, you're left with the original analog signal.
The big deal about Shannon-Nyquist is that they Did The Math, and proved (beyond a shadow of a doubt) that a sampling system consisting of an input filter, a sampler, a 'desampler', and an output filter really does more or less perfectly reconstruct the original analog waveform, minus all frequency components above the Nyquist frequency. And yes, this really does mean you can perfectly reconstruct a sine wave with only slightly more than 2 samples per cycle, no matter how impossible that might seem by intuition.
Raise the sample rate where there are enough samples to accurately render a 20kHz waveform of any shape and your digital sample will sound "warmer" than the LP while lacking the LP's inherent noise problems.
Nope.
This "20 KHz waveform of any shape" stuff (especially with respect to triangle waves) is a classic way that people fool themselves about how this works. Here's a neat web page showing what's actually going on with a triangle wave:
http://www.bsharp.org/physics/guitar
Basically, in signal theory, all waves of any shape are composed by summing sinusoids. A 20 KHz triangle wave is actually a 20 KHz sinusoid fundamental plus an infinite series of lower-magnitude sinusoidal harmonics at multiples of 20K.
The thing is, your ear is basically an array of small amplitude sensors, each of which has a narrow bandpass filter so it only responds to
You're wasting your time. He's been told over and over again, even studied the subject, and still comes away clueless. He doesn't get it, and he never will.
mcgrew thinks "waveforms" are magical strings of sound that sail through the air clinging together in a "square shape" or "sawtooth shape."
He can't grasp that it's possible to not hear part of a square wave because the harmonics are too high for the eardrum or hair cells to respond to - or comprehend that the harmonics above 20kHz have almost no energy (0.02% for a high piano note) and won't budge your eardrum anyway.
Instead, he prefers to think that 20kHz roll-off filters "distort square waves into sine waves." If you could just keep that darn wave square, by God, you'd hear what a different quality it has from a 20kHz sine wave!
Want to know what's funny? The author of the article claims the maths are being misunderstood. LOL!!!!!!!!!!!!!!!!
Does the author understand the fact the Nyquist theorem only applies to continuous time signals, not discrete time signals? No.
Does the author undestand the difference between a continuous time signal and a discrete time signal?
Probably no.
Does the author understand the fact the continuous time signal in the theorem is represented as a Real Function (http://www.proofwiki.org/wiki/Definition:Real_Function)?
Probably no.
Does the author know what the Domain of a Real Function is if it isn't specified (this is explained by the article I linked above)?
Probably no.
Does the author have any idea what a Real Number (http://www.proofwiki.org/wiki/Definition:Real_Number) actually is?
I guess that's another no.
Does he know something about interpolation errors and quantization errors?
Does he know something about audiophile equipment?
Let me think very, very hard now...
No.
Wait a minute, please...
Still no.
about flac for years
So what do you call it when people refuse to believe the truth because it doesn't fit their perception of reality?
There are damn good reasons why studio work is done in 24/192, and while I agree that most playback devices cannot produce the uber-high frequencies, nor would we want them to, I think the arguments against distributing 24/192 are pretty weak.
For one, his argument about digital sampling is bullshit, and demonstrates a poor understanding of the Nyquist-Shannon theorem. In his idealized case of a pure sine wave, yes you only need to sample at the Nyquist frequency. Once you start mixing different sounds together, that all falls apart since the sound wave is no longer a stable shape but rather an additive-subtractive mess of several frequencies, which do reconstruct in such a predictable fashion. Heck, even a simple square wave at f/2 will result in audible distortion on the DAC side, because it simply cannot recreate the infinite "slope". It's not even a matter of hearing up to 22khz, I know I can't anymore, but the harmonics of a true square wave cover the entire range down to 0hz, and that's what you can actually hear. If that square wave gets tapered or rounded by inadequate sampling, you end up with a triangle or sine wave which sounds radically different.
Does the average ear need 24/192 to be satisfied ? No. Does it mean we should entirely stop distributing such content ? Fuck no. I have a pretty decent studio setup, cheap but decent, and good enough ears with the technical training to notice those distortion artifacts. Okay, I'm a freak of hearing with perfect pitch and damn near digital memory for audio - hell I can identify a few dozen vocal mics just by listening to a mixed and mastered CD. Those high resolution recordings are for ME! They provide me with some geeky audio entertainment, which makes it worth the extra download time and minor expense of 24/192 capable equipment. My wife, who is a trained opera singer, cannot hear those details; she doesn't listen in such analytical fashion. Hell she can't even tell if I subtly pitch or time-stretch a track for DJ mixing... For her uses, 44khz is more than enough. So what's so wrong in providing different files for different listeners, and why does this Monty guy think his opinion trumps anyone elses ?
-Billco, Fnarg.com
24-bit is a vast improvement over 16-bit, both technically, and audibly. Also 48k is a significant improvement over that, just ask any experienced audio engineer. Anything above 48k/24bit is a marginal improvement in most cases, if any improvement can be detected at all. But do not confuse the issue by giving the impression that the current audio fidelity is sufficient, it's pretty atrocious, and it is a problem, despite what you may think. Most notably lossy data compression is proven to cause ear fatigue, so the most important step is providing lossless audio, which is available in many cases. The next step is to improve fidelity over the archaic 44.1k/16bit standard, to at least 48k/24bit. If you doubt this, ask any professional that works with audio and prepare to be schooled.
Actually, you can't discern a wave shape at all. If you have a wave with the exact same harmonic ratios as a square wave, but the higher harmonics are out of phase, the wave will not look like a square wave, but it WILL sound like one.
It's only when the phases are changing that you can hear the difference - with a constant tone, a square wave is indistinguishable from an infinite number of waveshapes with the same harmonic content.
The hair cells in your ears respond to ranges of frequencies, and they ignore anything outside their range ("ignore" is probably a bad way of putting it - they are unresponsive).
Only if your definition of "perfectly good" is "so convoluted that nobody EVER uses it". ;)
Let's be honest here, verisimilitude exhibits a superlative and ostentatious preponderance of syllables.
Indubitably.
What about playing that music on a very large sound system at some sort of concert or outdoor festival by, I don't know, a DJ. Just because you won't need it, doesn't mean others won't.
Never say never. Ah!! I did it again!