Slashdot Mirror
Stereo Component for Digital Audio
An anonymous reader writes "The techs over at Tech Radio interviewed Lydstrom, Inc. about their recently announced Songbank MZ3-5000, a digital audio player meant for the home that uses compression technology from Lucent to store up to 5000 songs on one player. Apparently the quality of the output is better than MP3, but the device is also MP3 compatible. They posted the interview on their Archives page. "
My guess is that a huge number of us have hacked together stereo
components to do this already (I use xmms & the IRMan plugin for
mine- although I wish it had better handling of large numbers
of playlists). Interesting tho.
i dunno..it worked on my linus.
Sometimes you gotta draw a line.
Loads fine here in Netscrape, also in Opera and even Lynx...
Forgive me my ignorance but what the heck is irman???
Mmm, crispy sound files.
The xing encoder is hosed, and it's the source of your low high-end loss and distortion. Try bladeenc instead.
I have one of those Packard Bell remotes...how did you get yours to work with Linux? (assuming you did, sionce you mention xaudio and perl...)
Yup, I've been thinking about getting into the whole MP3 thing (what with about 220 CDs to get on-line), but am concerned about quality too: In my mispent youth I blew a small fortune on a Beosystem 5500 with the 5500 CD player. It drives a Carver TFM-22 Amp and thense a paid of Carver Silver Amazing Loudspeakers. Best US$10k I ever spent in 1986 (though a frind's all McIntosh *Drool* system blows mine away, I wasn't exactly a millionaire like he was and couldn't spend $65k on a system).
I've been thinking of selling of the B&O, and replacing it with a decent DAC and jukebox for local play in the Living room, with on-line MP3 compressed versions for piping into other rooms in the house.
- Rene S. Hollan, posting as an AC at home
He mearly optimized an existing Bell Labs algorithm... (wich was unusably slow)
That's old news... the new Xing encoder (AudioCatalyst 1.5) works above 16khz. Mind you, it's not as good as Fraunhofer, but it works OK. Definitely sounds better than the old Xing encoding.
they are working on one
You know, I've been reading up on all this (as a gina owner), and it seems as if the OSS package doesn't support any of the popular multi port audio cards (event, aardvark, motu, gadget labs, etc.) are supported. Apparently the manufacturers have been less than forthcoming with adaquate technical data for anyone to develop a driver.
Which irks me to no end. I'd think by now someone would have been able to get a semi-functional event driver patched together from just dumping data from a windows box, but I guess not.
But most multi-port card makers are going for for the multimedia (semi)pro-audio market... in which linux is simply not a contender. Event and I think aardvark have said they'll support Be, though.
So yeah, spending $30 on 3 cheap soundcards would be the only option under linux if you have slots / interrupts available.
Different manufacturer's codecs will handle the problem differently, but the "problem" as it were lies strictly in the compression method itself.
As I guess you guys know, MP3 works on a psycho-acoustic masking scheme... but cymbal hits are really dense harmonically, and so the encoder will get really confused and start dropping the wrong packets (to mix metahpors), yielding that warbly sound.
The best way to get around this is to simply cut out the really high harmonics (over say, 16kHZ) -- xing's encoder doesn't do this, which is why it sounds inferior. The cymbals under under other codecs might lose some of their "richness" compared to the original, but at least you don't get that damn warbling.
So you're saying that the web designer probably also has something to do with the engineering of the product? Probably not.
lirc support is really a must have (especially since you can use the irman with lirc to :o) since it would allow for cohabitation between xmms and other ir-controlled proggies
You guys must be like, 13 years old. Come back to /. when you're old enough to realise what it's here for. Why should you do this? Because Stone Cold says so, apparently.
I wish when I was 13, I'd been as cool as you. Me? I just did stuff with my friends - you know, play football, watch a game...
Oh yeah, I see, you're the ones who all the others hate, becuase you just sit there, talking shit - even your teachers hate you. You don't actually have many friends do you? Guess they got scared off by your habit - the one were you foam at the mouth and then bite people.
Nasty tick you got there.
Hmm.
One of the 'founders' (initials AKP) has done this sort of promo-only website over and over again. I've met and spoken with him, this is first hand knowledge.
Check out the crazy web of made up hype surrounding them. For instance, their press page has an article (http://www.siliconindia.com/Jan99radar.html) written by someone from Spydre.com... Hmmm. the address for spydre is the same as lydstrom's boston office. There is more evidence than just this. I think what they do is, put some ideas out there and see which ones get the most interest.
The slashdot community should do some due diligence on this.
_IFF_ you use windows, the combination of Ampapod and Winamp work very well. The playlist problem is solved by Winamp quite nicely, and Ampapod lets you type song numbers on the remote. New hardware support for Ampapod should be available soon, but right now it only lets you use the X10 MouseREMOTE. If you've got video on the box, there's even an On-Screen display.
There's an article in an old Wired mag (3.08) about Ken Thompson's development of this system. It makes amusing reading when you think about the mp3 revolution which has occurred since the article was written (Aug 1995). His compression method also appears to have a much higher sound quality than mp3, if the listening experiments in the article are to be believed. Unlike mp3, his method of lossy compression is almost impossible to distinguish from a CD.
IRMan is a piece of hardware that allows you to use standard remote controls to control your PC. There's software available that works under both Linux and Windows. The webpage is http://www.evation.com/irman/
Even with techo you can really hear a difference with the crappy sound cards in most computers these days. A cheap card in a noisy PC environment simply cannot compete with a quality DAC in an external device like a CD player. The really cheezy amplifiers and attempt at tone control make up the rest of the crappy sound. The best think for listening to MP3's would be something like a SB Live/AWE64 that has an external digital out and then buy a decent external DAC. But if you are listening to music on computer speakers then you are just hosed anyhow :>
I too thought MP3 was good enough. Then I heard MPEG4 on the Windows Media player - it is a big improvement.
The Ensoniq AudioPCI has 2 DACs onboard and apparently there is a way to route one of them to the "line in" jack for 4-speaker sound (I am not kidding). I don't know whether it's any good for what you want to do, since it might have too much crosstalk between outputs, though.
Multiple cheap PCI sound cards would be pretty easy and cheap, though, if you can spare 3 slots. Also, USB audio might eventually be good for this, if it weren't for the tendency for manufacturers to always bundle DACs inside speakers, as opposed to separate products.
According to the interview, the system's "reccomendations" are essentially push advertising, trying to get you to buy new songs online. I mean, we've all suffered suggestions from these "altavista reccomends" and "ebay reccomends" type things. Do I want to be assulted with advertising while I'm trying to dial up a playlist?
Second, to do this they have to profile my musical tastes. That's demographic research, and there's little to no reason for them not to turn around and sell it. Privacy paranoids take note.
Third, as the guy also mentioned, lucent is a big name in the ongoing SDMI debate. The unit's OS is flash ROM updated, so when phase II of SDMI gets implemented, no more MP3 support.
It's just an embedded system with a propriarity UI and a propriarity codec designed to get people to buy more (music / hardware).
Thus, I urge people to keep up work on the open source hardware front.
If you're interested in this kind of thing, you should consider subscribing. --Aaron
About that mute button: the reason why its not :) Mute
part of the plugin is that I have my prologic
receiver connected to my computer and I'm using
that remote with IRman. The first version of the
plugin didn't even have volume control!
was just one of those things I overlooked, it will
be in the next update along with a few more
features to keep Rob a bit more happy.
While I cant comment on ampaPOD as I've never
heard of it before, I can tell you that I've
been quite happy with my IRman and the people
responsible for creating it. I asked them to
create an x11amp (before the name change) plugin,
and they said that they didn't have the time to
do it but would support anyone who did. So I took
that offer up and they shipped me a free IRman
for creating the plugin. If there's any features
you'd like to see in the next update please dont
hesitate to ask. Already planned: mute, 100's
(or 1000's) of playlists, next/prev playlist
option to beep on button presses, option for
tv-like display on button presses, and maybe lirc
support.
Get a good pair of headphones, and then encode some nine inch nails music. The mp3 lossy compression really messes up nine inch nails, especially stuff from The Downward Spiral. Stuff like white/grey noise, fuzzy synthesizers, harmonics, and large volume changes all get messed up with mp3 at 128kbps. 192kbps sounds pretty good, however. I've noticed similar problems with a lot of other industrial music, but even with normal rock music I can tell the difference between 128kbps and 192kbps with a good pair of headphones.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Sounds like a problem with either Netscape or your particular setup. A browser should not crash under any circumstances, no matter how badly formed the HTML is.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Yeah, but the problem is that vinyl doesn't keep its sound quality. My 15-year-old CD, as long as I don't scratch it, will sound the same as a new CD. A 15-year-old vinyl will not. My parents' 35-year-old vinyls sound absolutely terrible (warped surfaces from heat changes and deepened needle grooves from excessive play).
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Xing's encoder discards everything above 16 kHz, which is great for speed but TERRIBLE for sound quality! blade enc is great for > 128kbit encoding, and the Fraunhofer codec is cool for 128kbit... you can get the original source distribution for free in many places, and somebody wrote a patch for it to make it work without a time limit. I'd say there are plenty of quality mp3 encoders for linux out there.
I built an mp3 player myself out of a motherboard with built in audio and video, a hard drive, a small power supply, and a packard bell IR serial remote. Total investment: about $200, and a few hours of my time to code up a front end to xaudio using perl. It doesn't have a display currently, but I can select from 100 different playlists (arbitrary, could be 10,000 but I don't think i'll need more than 100 for now), and eventually support a CD-ROM changer attached to the machine.
Check it out:
http://www.veedub.nu/mp3player/
I use multiple (3) Esoniq 1371 AudioPCI cards with the ALSA drivers for 'em. They were the "quietest" cheap cards I could get my hands on for this exact purpose.
I'm intending on building a simple fibre optic distribution system from my MP3 machine to rooms of my house. basically sawtooth waveform at some unbelievably high frequency (10MHz?) with a comparator to turn each channel into PWM and then send it through the fibre (I am using digital transmitters and receivers) At the other end I can smooth it out and (hopefully) recover the audio.
I'm still trying to think of a way to simply get both channels into one fibre with one transmitter... if anyone has any ideas, please email me.
Andrew
Depends on your music. Simple stuff like techno
and normally "noisy" stuff like alternapoprock
tends to sound about the same. But if you listen
to something clean and complex (jazz, stuff with
a lot of horns or strings, clear vocals) you can
definitely hear the difference.
I've kicked the bitrate up to 256 and gotten
pretty good results, but 256 kinda sucks as far
as space goes.
software should not crash just because the inputs are malformed.
this use of the term `stereo components'
that thing Ken Thompson was talking about when he was bashing linux and mp3?
What good is it if it isn't Free (an open standard)? After the shit people have tried to pull with gif, I don't think anyone is willing to go with a proprietary "standard" anymore.
support gun control: take guns from cops
This is hard, anyone got any thoughts?
IRman works with linux, but ampapod and the X11 remote can do cool stuff like web surfing, and I guess I could always just pop VMware into linux and use ampapod after all.
But then again, I've heard that X11 makes poor quality stuff, and Rob uses IRMan, which looks high quality but doesn't have the same flexibility, tho it looks higher quality.
But the X11 remote costs less.
But its supported by a company and the IRman looks only loosly supported by a comany (not exactly a global corporation, looks like a buncha students).
Geez, someone got some thoughts on this? I gotta get one of these, like NOW!
Check my Go-related blog for beginners: DGD
I use the xmms and irman like Rob, and it works great for me. I also have a couple X-10 units and use the X-10 remote with the irman. Only thing that doesn't make sense to me, there is no mute for the xmms plugin, wierd. I don't have anything to do with the irman, but best 40 bucks I spent. Your milage may vary.
All I can say is this thing is sweet! Here are the details from the site, for those too lazy:
:)
o approx 250 hours of songs
o IEEE-1394 connector for extra storage units
o auto detect song title, artist, etc (if stored on CD)
o smooth fade between songs and beat matching for DJ-like transitions between songs
o "CustomDJ": "...an intelligent agent that watches how you listen to music. As it learns about your listening habits, it will automatically play the songs you like at the right time and day of the week."
o touchscreen remote
o 3 stereo (analog) outs - play different selections in different rooms at the same time!
Plus its good looking. Costs $699 - $999 depending on when you want this thing and which promotional offer you take.
Are there any software mp3 players that have the beat matching and DJ abilities this has? Those are the kind of features I would put in a design of my own - can't do that on a CD player (without having your own personal DJ
Complexity Happens
For the features you get the price on this thing is actually right on. The machine I built has ended up costing me about 500 bucks so far total - between lcds, keypad ir stuff etc...and the computer. If I added some of the more advanced features this thing has Id spend days programming and still have to fork out another few hundred bucks for a multichannel soundcard. And It still wouldnt have the beatmatching feature which I think is really cool-If I can figure anyway to pull this one off on my box Its going in ther.
Ther is a difference but you really need a nice sound system to notice it. The biggest differences are really in extreme upper and lower spectrum, so unless you have some really tricked out speakers you really wont hear any difference at all.
Another thing to consider is the quality of the ripping/encoding software used to actually make the mp3's. Not all rippers/encoders are created equal (see NexEncode Studio version 1.0 - ick! - do they still make that thing BTW? I dont see how they GIVE it away). Some of the (free)linux software is lacking in this respect.Sure thers the Xing encoder....now if theyd just port audio grabber we'd be set.
"The cymbal crashes are the most noticeable difference - they are often a little distorted or have a "phaser" effect."
That is almost always: a) a recording problem or b) an encoding problem.
i've found vqf to be slightly better for classical
music
i think you can get a player and ecoder for free from yamaha (www.yamaha.com)
------- Oh damn.... the Sigfile escaped... -Great OM
Without going into much details, let's say I've had a couple of opportunities to see that this compression algorithm is for far the best I've ever seen yet - and I heard if first about 3 years ago. On a single CD-R, with about 145 megabytes left, I've seen over 210 songs. When listening to the music, I would never be able to distinguish between the compressed version and a CD-quality recording. And the decompressor would probably fit on a floppy if they wanted - but i guess they'll be using the DSP version they made before hacking the Windoze version.
Must... have... one...
C.
I know it's probably not the popular genere on Slashdot, but I have noticed that MP3 mangles classical music. I listen with a nice pair of headphones and there is a perceptable high frequency "hissing" for lack of a better word. It doesn't sound like white noise though and I don't get it when I listen to CD's through the computer. Bumping up to 192kbs doesn't seem to help either. Has anyone else experienced similar problems? If so what is a better format for classical music?
From their web site:
Zero Error Processing (ZEP)TM
The MZ3-5000 features ZEP processing, which is the industry's first system for dealing with 100% of all errors encountered on a physical CD.
Sounds a little bit like IntelliSens e (TM) to me.
With my crappy speakers, it's hard to tell the difference most of the time. But when I put on just about any pair of headphones and really listen, I can usually spot the MP3. The cymbal crashes are the most noticeable difference - they are often a little distorted or have a "phaser" effect. Also, the stereo image sometimes changes in noticeable ways.
That is almost always: a) a recording problem or b) an encoding problem.
I guess it's the encoding then. Xing at 128kbit. The differences are not obvious, but they are there if you listen. I do my A/B "blinded" test by having a friend play the CD or the MP3 through my soundcard and headphones. The MP3's sound great, but in addition to the differences noted above, are just not as "crisp" (I admit this is not a scientific term) as the CD. But definitely good enough for everyday listening.
Anybody out there using a multichannel (three or more stereo outs) sound card with Linux? I would like to do zoned audio at home from a single large HD and server, with client control via IR or home intranet. This way the wife can listen to her Tori Amos in the study while I'm piping Eric Johnson into the living room. Of course, multiple stereo systems would solve this, but that wouldn't be as much fun!
The song attributes and ratings will be kept in a mysql database (with different ratings for each user), and playlist queries will be contructed with a simple intranet form.
The last piece of the puzzle for me is the card. Any multichannel card recommendations out there?
Ken used this compression for his own collection, but that's all. Ken is a computer scientist, not an expert on audio coding! PAC was developed by a team led by Peter Kroon.
Tori Amos CDs and LPs tend to be fairly good recordings. On reasonably good systems it sounds great, but on cheaper systems the music loses all its life. I have yet to listen to her from a compressed source, but I would guess from other experience that it would be sound pretty poor.
You must have amazing hearing because I can't tell the difference between an audio CD and an MP3 (provided the MP3 is at least 128kbps and 44KHz).
-Cerberus
I don't know about you, but my servers run on the power of cotton candy and happy thoughts. -Anonymous Coward
I am sure that it probably wouldn't be too hard to redirect output to the second audio out channel on this card.
In addition to those two analog outputs I just talked about, there are two additional digital outputs (SPDIFs I believe) that can also be used as outputs. Sure, this card isn't cheap, but I like it since I can send up to 16 WAV (is this the right term for this?) streams to the card at the same time, and have them mixed on the fly.
Overall, I think this is a really high performance card and would probably work with what you want if you want multichannel and high quality audio. The mixing ability for multiple WAV inputs would probably be very useful if you want to write your own DJ software to interface with the card.
The highest-quality consumer format for most any music is an 180 gram virgin vinyl LP (preferably manufactured by RTI) with direct metal mastering, played back on a quality turntable, amplifier, and loudspeakers. There is is no consumer digital audio format that can top the dynamic range and frequency response you can acheive from a well-produced LP.
Granted, this is audiophile-level quality; some Joe Schmoe listening to Nine Inch Nails wouldn't necessarily be able to tell the difference.
With regard to Ashwin Philips: Ok, I've known Ashwin for roughly 3 years, beginning when I first hired him as a consultant on a (successful) large-scale client/server development project. In this time, I have never known him to engage in any kind of vapor promotion. I've never had reason to question his personal business ethics. As far as the product is concerned - wait a bit longer, then judge it by its own merits. We've had a few techincal discussions, and - though I have some doubt - I believe it will deliver what's been promised. As for due diligence - 'Evidence' is one thing, slander is another.
Multi Sound cards is a good idea (i've been playign around with it for a while myself). There's just a couple of things to watch for ...
1) Some sound cards don't like running in the same computer with another one of themselves. The old soundblaster cards were notorious for this. These days with PCI soundcards negotiating (is that spelt right?) their resources, it SHOULDN'T be a problem, but I've still had some problems with it. I have two Yamaha PCI DS-XG cards here that refuse to run together (under Winblows, anyway). Both detect and negotiate just fine at startup, but the drivers conflict like nothing else. I haven't yet bothered to go lookign for Linux drivers for this card, to see if they'll work together with a decent set of drivers, tho (as my multiple channel autio output needs are under Winblows at the moment). I ended up getting one of those cards, as well as an ESS Solo-1 PCI card working together quite happily tho.
and
2) Synchronisation between the cards. It doesn't look like this'll be a problem for you, but output over multiple soundcards, can lose synchronisation after a while. (Well, technically, it was always out of synch, but you tend to notice it more the longer it plays). This is only a problem if (like me) you've got one song takign up multiple channels over several sound cards.
Eg, I record stuff for my band to have as backing for when we play. We're trying to make this stuff output over 4 channels so that we have a greater controll over it at the PA. What we've got is 2 mono drum tracks coming out of the stereo output of one sound card, and then two mono keyboard tracks coming out of the other cards output. Problem is that the cards don't quite have perfect synch between their respective sample timings, and so one card can end up outputting the waveform slightly quicker or slower than the other. Now, the difference is less than the space of a sample (1/44000th of a second), but it's still there, and the longer a song goes for (we play 8 minute songs on average, with 12-13 minutes being about our ceiling) the more this slight difference gets compounded. For short songs you may very well not notice it at all. Bear in mind too, that the problem only starts everytime you start playing something, and stops when you stop. It's only for extended playing times that you'll really notice it.
Anyway, if you're planning on having seperate (unrelated) waveforms output from each card it's not an issue for you. If, however, you want the outputs kept in synch for some reason (say, you want the same song comign out all outputs so you can listen to it in 3 different rooms) you may (and I say, only may) start to notice the synch difference.
ANyway, I hope this is helpful to you.
Sounds cool!!
:)
So where do we get it from?