TAS 194: The State of USB Audio
September 6th, 2009 -- by default
Source: The Absolute Sound
Source: The Absolute Sound
The high end’s trademark is sonically meaningful evolution; revolutions are rare. USB, however, gives every indication of being the latter. Though its roots are pedestrian—the interface was designed to facilitate connectivity between computers and standard peripherals—USB now augurs a potential sea-change in the way audiophiles acquire, store, and play music.
In the USB world, PCs serve as the primary audio source. They store entire music libraries, yet offer the ability to access desired material—based on any number of criteria—without the need to scour a single dusty shelf. Silver discs are history, all of them having been “ripped” to the PC. New music need not ever reside on such physical media, for it can be downloaded with a click. Playback emanates from hard drives, which sound better than optical disc readers. And, perhaps most importantly from a sonic standpoint, CD-quality sample- and bit-rates are no longer a barrier; PCs and USB both support much higher resolution levels.
Dedicated music servers can do all this, too. But they tend to be expensive and are still in limited distribution. In contrast, USB empowers anyone who owns a PC to reap the benefits of purpose-built music servers.
Of course, a stock PC does not a music server make. To assume that mantle, PCs require a support ecosystem that includes hard drives with the capacity to store a music library, Web sites offering a varied selection of music in formats from MP3 to HD, and user-friendly music-server software. As it happens, all of these essentials are already in plentiful supply, many of them for free or close to it. The final necessary element in this ecosystem is a means for the PC to connect to the rest of the playback chain. This is where USB comes in.
True to its first initial—the other two stand for Serial Bus—USB is indeed “universal” among PCs, making it the natural choice for such a connection. But there must also be something within the audio system that supports the same interface. For high-end installations, that something is a USB DAC—and manufacturers are unleashing a flood of them. Indeed, USB has become a checklist feature; new DACs exclude it at their competitive peril. Meanwhile, for those with “legacy” DACs, a new component class has emerged: the USB-to-S/PDIF converter. With these offerings, the audiophile industry has delivered the last pieces needed to realize a sonic revolution.
Assessment
But let’s step back a minute. Before clasping USB to our collective bosom, shouldn’t the high end first ask some basic questions about this interface, such as how does it actually sound? TASthinks so. Which is why yours truly spent roughly six months listening to and generally playing around with USB as delivered through a score of PCs, software programs, cables, DACs, and converter boxes. I was seeking credible, reliable answers to fundamental USB-related questions. And because my experiences with the interface proved remarkably consistent across time, products, and source material, I believe I found them.
Before delving into those questions and answers, though, a few notes about the test bed. The primary source device through this journey of exploration was my own current-generation HP Pavilion laptop PC running the latest version of Windows Vista. (As anyone who follows the various computer-audio Web sites knows by now, Vista outstrips XP as an audio platform because of the latter’s sonically destructive, difficult-to-bypass kernel-mixer module.) Other sources in my test bed included an Apple iMac G4 running OSX, as well as a custom-built desktop PC sporting Windows 2000 and an upgraded internal sound card. These disparate machines and operating systems enabled me to assess not only the nature of USB audio, but any platform-based variations.
Naturally, all three computers had USB ports, but the HP and Apple also offered a FireWire interface, while the custom machine’s enhanced sound card could deliver high-resolution coax S/PDIF. All this flexibility allowed me to methodically compare USB to alternative interfaces.
All content, design, and layout are Copyright © 1999 - 2011 NextScreen. All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
Comments
Obviously Alan Taffel is on Crack! Tthis is the most rediculous article I have read to date.
Pretty ridiculous spelling on your part.
Obviously Alan Taffel is on Crack! Tthis is the most rediculous article I have read to date.
You wrote: "However, there is as yet no such thing as a FireWire DAC."
There certainly is, for quite some time already!
See:
http://www.weiss-highend.ch/minerva/index.html
Daniel Weiss
....and the fine line of RME DAC's, such as the one I own, like the Fireface 400. Were you deliberately excluding pro firewire DAC's? It seems like a pretty big oversight for an otherwise thoughtful piece.
You wrote: "No doubt audiophiles would also welcome a FireWire-to-S/PDIF converter built to audiophile standards."
Yep, see:
http://www.weiss-highend.ch/vesta/index.html
Daniel Weiss
There are many faults to the methodologies and reasoning used Mr. Taffel's negative commentary on the state of USB audio that have been thoroughly exposed. I currently have a USB DAC connected to my PC and am absolutely thrilled at the fidelity achieved playing back high-resolution files I've downloaded. I am also thoroughly addicted to the ease of access to my music through a well-structured database and intuitive user interface. Achieving this level of playback and ease of access wasn't easy. I now know much more than I intended about the USB protocol, the Windows audio stack, and ID3 tags. But computer audio is in its infancy, and I'm confident that vendors will eventually be introducing high-performance products that are much more plug-and-play. Downloads of high-resolution audio content is the best possible method (outside of the DSD/DXD studio masters) to deliver software that can provide the ultimate in sonic experiences to audiophiles.
I also feel that the current focus on USB as a transport method will pass. I chose the USB path only as a stop gap until reasonably priced network players with sophisticated user interfaces become available. Linn and Naim currently offer some very high performance networked players, but their pricing is outside my acceptable range. It was really encouraging to see Onkyo include the ability to stream 24/96 audio files in their new networked receivers. I really don't want a PC with a keyboard and mouse as a part of my entertainment system.
So what change, exactly, did the USB cables make to the digital data going through them, which causes the resulting change in sound? Either the bits made it through the line successfully, or they didn't. You cannot have degradation of a serial data path without data loss, so please, help my poor little brain understand how anything other than "sex appeal" of the cable caused any audible difference. You can get away with such tripe with analog signals, because the subjectivity is such that nobody can argue with you about whether or not you can hear the diff between 14 ga lamp cord and "oxygen-free copper." You can't get away with this with digital.
Wait... I should say, "You SHOULDN'T be able to get away with this with digital." Sadly, you can gold plate volume knobs and convince some people to buy them based on "improved soundstage."
This is the equivalent of a typewriter user reviewing the latest version of Microsoft Word...he is just completely out of his element and confused by everything he is encountering, and generally not very happy to be using something other than the typewriter he has grown to love over the years. What a complete waste of time.
>>>Either the bits made it through the line successfully, or they didn't.
You are never going to convince the subjectivists of this. They don't understand digital audio and they never will. They think audio is like a car - that it can always be tweaked to sound better. They simply do not understand that what's recorded on a CD are lands and pits representing (via transition) 1s and 0s. They refuse to accept that two 16-bit binary words with the same values will always sound the same when played back on the same equipment.
I got into a bunch of arguments over digital audio on the Beatles Remasters forums on Amazon and finally gave up. There was a guy swearing that pressed CDs from the same master, but pressed on different presses, sounded different. I and others argued "not if the bits are the same". He wasn't having it. The don't believe in blind A-B testing either. They believe that there are things that can't be measured (even though that has nothing to do with blind or double-blind A-B testing.)
I'm okay with them claiming that sampling rate or word-length on a CD isn't enough for proper reproduction, even though numerous tests have been run that demonstrate that the vast majority of people cannot hear the difference as compared to a high-resolution digital format - at least those differences can be measured. I'm okay with them claiming that quantizing errors destroy digital audio. But when they make claims that a copper interface sending digital audio makes differences to the sound of the audio, there's no arguing with them.
This is no different than arguments over religion and they frequently turn violent. There are those who don't trust science and believe what they believe on the basis of faith and in the opposite corner are those who want measurable proof. Try to convince someone who believes the earth was created 5000 years ago with carbon dating results or the fossil record. It can't be done. They simply cannot accept evidence that isn't consistent with what they already believe. But...on the up side...it does sell a lot of audio equipment and software too. On the Beatles Box forum, there are apparently numerous people who spend substantial portions of their income seeking out every single version of Beatles music ever released, constantly looking for the holy grail: the one with what they perceive as the best sound, which usually winds up being "the purple and pink vinyl pressed in Estonia in 1974, but only from press #31 and only the second pressing".
All of you people who always answer "bits are bits" are ignoramuses.
It simply is irrelevant. The bits on your CD are not delivered as precisely through your CD player as they are from a hard drive, for example. Ever notice that CD players come with X times over sampling, etc? That's because the medium is an inherently error prone one - not in the storage, but in the playback.
Ever heard of jitter? It's the digital equivalent of distortion. Lots of ways to introduce jitter into the digital signal path, and some setups have measurably and audibly less.
I could go on for pages of examples that prove that bits aren't bits - the ones you start out with aren't necessarily the ones delivered to your DAC, and by extension, your amplifier.
Same thing with USB. Some devices transfer it better than others. Period. If you don't think this is true, do a little research and find out how wrong you are.
All of you people who always answer "bits are bits" are ignoramuses.
It simply is irrelevant. The bits on your CD are not delivered as precisely through your CD player as they are from a hard drive, for example. Ever notice that CD players come with X times over sampling, etc? That's because the medium is an inherently error prone one - not in the storage, but in the playback.
Ever heard of jitter? It's the digital equivalent of distortion. Lots of ways to introduce jitter into the digital signal path, and some setups have measurably and audibly less.
I could go on for pages of examples that prove that bits aren't bits - the ones you start out with aren't necessarily the ones delivered to your DAC, and by extension, your amplifier.
Same thing with USB. Some devices transfer it better than others. Period. If you don't think this is true, do a little research and find out how wrong you are.
When I read text of my hard disk drive via USB words are not magically changed, HOT is not changed to HIT. The USB cable and protocol guarantee 100% error free data or no data at all. The DAC sees the EXACTLY the same set of digits whether USB, Firewire or S/PDIF or indeed any and all digital interfaces that uses error correction eg CRC checks that make sure that if the data is not 100% accurate it is re-sent until it is.
After reading this USB article and his review of a USB->S/PDIF converter I am now firmly of the view that mister Taffel has NO idea what the difference is between digital and analog.
Is Alan Taffel a spoof writer inserted to test the gullibilty of your readers?
To all,
I understand your skepticism over differences in USB cables. However, sometimes there will be a difference due to jitter. Please allow me to explain...
In digital audio, one of the most devestating variables is jitter. There is plenty of information regarding jitter on Wikipedia, AES journals, etc, but let me state it simply. Digital converters are synchronous devices. They convert in a periodic manner, one sample at a time. The accuracy of this periodicity is crucial to the quality of conversion. Jitter disrupts the accuracy of this periodicity.
Benchmark converters utilize a clever clocking mechanism (UltraLock) to buffer and re-clock the data using an on-board clock. By buffering, the timing errors from the cable, source, etc are moot because the data is no longer periodic. Its just sitting in a buffer. The on-board clock takes over the duty of synchronizing the data and the D/A converter. This way, all cables and sources WILL sound the same, as long as the 1's and 0's are the same, when using Benchmark converters.
All the best,
Elias Gwinn
Applications Engineer
Benchmark Media Systems, Inc
Data is data... or is it? If I play the exact same data, with the exact same DAC, and the exact same amplifier, with two different DAC clock sources, one clock source having very high jitter, and one clock source matching the exact sample timing of the original data, I will hear two very different sounds.
Look at the difference in clock generation for USB and for SPDIF.
The stuff about USB cables seems crazy though. I can not imagine that a basic USB cable has such a high BER (Bit Error Rate) that one can not reliably move the data in sufficient time.
CDs use a 44.1kHz 16-bit format. If this is ripped to the computer hard disk in a lossless format one might presume that the same 44.1 kHz 16-bit data will emerge from the computer upon playback. Unfortunately this is often NOT the case.
Media players and operating systems both have the ability to apply digital signal processing (DSP) to the audio before it is delivered to the output port (USB, Firewire, or S/PDIF). Some DSP operations such as digital volume controls, pan controls, and L/R balance controls, are relatively benign when used in moderation. Other DSP operations are very destructive. Computer operating systems have the ability to apply DSP operations such as sample rate conversion (SRC). SRC processes can be executed with great precision if one is willing to expend large portions of the computer’s available CPU time. In contrast, very poor quality sample rate conversion can be accomplished with very little CPU time. Unfortunately, most operating systems attempt to conserve CPU time by utilizing very poor SRC algorithms.
With the proper outboard USB hardware it is possible to avoid damaging DSP processing inside the computer.
The first requirement is that the sample rate transmitted over the USB interface must match the sample rate of the audio being played. If these sample rates do not match exactly, the operating system will immediately begin applying low-quality sample rate conversion to make the audio play. After all, playing something is better than playing nothing. Unfortunately many USB devices are not frequency agile. The outboard USB DAC must be able to respond instantly to requests to change sample rate. Asynchronous USB interfaces fail the frequency agility test.
The second requirement is that the interface must support 24-bit data even when playing 16-bit files! The reason for this is that the operating system and the media player both have digital volume controls, and the operating system also has a digital mixer. If the gain controls are all set to exactly unity gain, and all other input channels are muted, and all pan and balance controls are exactly set at center positions, then the data out will match the data in. Unfortunately the graphical interfaces used to set these controls make it nearly impossible to get all of these setting exactly right. The result is that the 16-bit audio gets multiplied by a number that is very close to 1 but not exactly 1. This extends the word length of the audio to 24-bits when multiplied by an 8-bit volume control setting. If the USB output device can support 24-bits, then no damage is done (other than a very slight gain change). In contrast, if the USB output device can only support 16-bits, the operating system will simply truncate the data from 24-bits to 16-bits. This 16-bit truncation causes a grainy distortion that is very noticeable in quiet portions of a recording. USB interfaces must support 24-bits.
The third requirement is clock stability at the D/A converter sample and hold circuit. Low jitter clock performance can be achieved using a variety of techniques. These techniques include: multi-stag analog PLLs, digital PLLs, storage and retrival systems, and ASRC reclocking systems that employ digital PLLs. USB DACs must have the ability to reject (filter out) the jitter on the USB interface in such a way that it has no significant influence on the D/A converters sample and hold circuit.
The Benchmark DAC1 USB interface meets these three criteria, and our test show that performance can equal that of S/PDIF interfaces even when the computer gain controls are not set to exactly unity gain.
John Siau
VP, Director of Engineering
Benchmark Media Systems, Inc.
Can't wait to see what Gordon Rankin has to say.
Obviously none of the units tested were Asynchronous USB.
You should include a Wavelength DAC and Ayre USB DAC in your comparison before firmly stating your conclusion
that USB can't sound as good as SPIDF.
Mr. Taffel,
There are essential settings that must be minded when playing music from a computer. Benchmark has done extensive research on this, and the results are posted on our Audio Wiki.
If you don't properly configure the settings in the computer, the errors can kill the quality of the audio. Sample rate conversion, word-length truncation, and other DSP settings are insidious.
We’ve done plenty of listening tests AND bench tests that show that the USB input is indistinguishable from the other inputs. But that is only when the settings are correct!! Otherwise, all bets are off.
All the best,
Elias Gwinn
Applications Engineer
Benchmark Media Systems, Inc
Mr. Taffel,
There are essential settings that must be minded when playing music from a computer. Benchmark has done extensive research on this, and the results are posted on our Audio Wiki.
If you don't properly configure the settings in the computer, the errors can kill the quality of the audio. Sample rate conversion, word-length truncation, and other DSP settings are insidious.
We’ve done plenty of listening tests AND bench tests that show that the USB input is indistinguishable from the other inputs. But that is only when the settings are correct!! Otherwise, all bets are off.
All the best,
Elias Gwinn
Applications Engineer
Benchmark Media Systems, Inc
Mr. Taffel,
There are essential settings that must be minded when playing music from a computer. Benchmark has done extensive research on this, and the results are posted on our Audio Wiki.
If you don't properly configure the settings in the computer, the errors can kill the quality of the audio. Sample rate conversion, word-length truncation, and other DSP settings are insidious.
We’ve done plenty of listening tests AND bench tests that show that the USB input is indistinguishable from the other inputs. But that is only when the settings are correct!! Otherwise, all bets are off.
All the best,
Elias Gwinn
Applications Engineer
Benchmark Media Systems, Inc
Clock jitter is measurable, and it is digital.
Good to see people are responding so eloquently and making such convincing, thoughtfully-reasoned rebuttals here....
I tend to get a little twitchy when people talk of audio components as "having poor timing," suggesting that a drum solo played over a $13,000 pair of Magnepan MG20's rather than a $300 pair of Monitor Audio BR1's will not merely sound better but have the drummer be more on beat. Either the fellow's hitting the cymbal at the right time or he isn't.
Having said that, jitter is a real electrical phenomenon, not something that audiophiles have made up to get you to buy volcanic rocks to put under your CD players, folks. And I certainly can't rule out that Mr. Taffel is picking up a real, concrete, well-documented difference between Firewire and USB. See, USB is a "host-based" protocol, requiring a computer; the CPU is actually doing all the data pushing. Firewire is peer-to-peer, and doesn't actually require a computer at all. In effect, So with Firewire, the CPU can just hand off data "chunks" to the dedicated Firewire hardware; with USB, the CPU is actually doing the work, streaming bit-by-bit to the USB port. Which of these is more subject to software-induced jitter? You guessed it: USB.
It doesn't seem to me that this is insurmountable, of course -- a USB DAC could have anti-jitter circuitry in it in much the same way the internal DAC for a CD player does. (Or an outboard DAC for those of you who do such things; as much as I enjoy reading TAS articles, I do sometimes wonder how in touch with real world incomes the writers are when they describe a $3000 amp as "great for a budget system." But I digress.) But having a better source stream in the first place is going to be preferable to trying to retime the signal later in the audio stream.
Martin (assuming Martin is your first name),
Thanks for an excellent comment. Most 'digital source data is just 1's and 0's and the associated digital transmission electronics can't affect an audio system's sound' proponents just do not understand enough about the engineering details to realize the impact that a small electrical phenomenon such as jitter or delays in the processing caused by a CPU's involvement in the transmission process can have on the final audio output.
In reference to your 'twitchy' comment about timing: I am the first to admit that I do not know enough about the electronics and physics involved in reproducing recorded sound via loudspeakers (and the human body factors involved in the hearing of that sound) to understand the 'timing' phenomenon; however, I have heard it quite frequently with speakers, amplifiers, and source components such as turntables and cd players. For some reason unknown to me (and I assume to many audio component designers) some components just do not seem to reproduce the pace, rhythm and/or 'timing' of a recording properly. For example, I once read review comments of a Magnepan model (don't remember which) which stated that "these speakers just don't rock". When listening to a normally lively pop rock recording over a different model of Magnepan speakers, I would have wholeheartedly agreed with the reviewer. They simply sounded rather slow and lifeless as far as the 'timing' of the music is concerned. It has nothing to do with the original timing of the musicians or even what has been recorded on the source medium. Somewhere within the reproduction component chain the 'perceived' timing is affected. I say 'perceived' since it may not be heard by every person listening and may have something to do with how people hear and perceive the timing of music. Possibly frequence response affects the manner in which we perceive music's timing and/or an individual's hearing may have a significant affect on whether or not that individual perceives the apparent 'timing' problem. For me it is a significant issue when choosing audio components.
Regards,
Bob
Actually, Watts is the first name, but a lot of people make that mistake. :)
I'm never willing to go so far as to say that because I personally can't hear a difference between audio components or formats or such that no difference exists. I demonstrated to my own satisfaction a while ago that while I can tell the difference between SACD and CD, and between a CD original and an 160kbps MP3 file, I can't tell the difference between the CD original and a 192kbps AAC file, even with a pretty good set of headphones and headphone amp (and a "very good for a 16-bit" USB DAC), and so far I've had no reason to feel ripping my CDs to 256kbps AAC was the wrong choice. (This seems to be about the equivalent of a 320kbps MP3 file.) I don't know whether many people can tell the difference between a 256kbps AAC file and the CD original reliably even among those who are reasonably sure they can -- but of course there are differences, so given good enough equipment and good enough hearing...
I suspect that "timing" in speakers may have to do with how accurately the speaker itself can convey the attack-delay-sustain-release components of music. Signal jitter might play a part there, but I'd bet physical characteristics make the most dramatic difference. Speakers are, after all, the point where all the signals get converted into actual sound! Magnepans make some really unusual choices with respect to physical music reproduction, and that's certainly going to make a difference.
Alan,
Surely you jest?
Chris Connaker
Founder
Computer Audiophile
The lack of scientific nature of this entire article is astounding.
The completely absurd conclusions follow.
I have to disagree with Mr. Taffel. I recently acquired a $99 USB DAC and it's performance blew me away. I have had very high end equipment for a long time and can definitely tell the difference in sound when I hear it. Some of the details Mr. Taffel mentions did not come accross to me. I have wondered how higher end USB DAC's like the one he used would sound in comparison to the cheapy piece that I used...
If anyone is interested in facts, and how any of this stuff actually works, read on....
Firstly, bits are bits. Full stop. There is no magic about "bits being convenient representations of an analog signal, so things can be close to being inbetween".
Secondly, USB ( and firewire ) both use Isochronous mode to transport data (audio) from one end to the other. This is good ( Isochronous means that the two boxes are guaranteed to be able to transfer the data between them without ever running out of bandwidth ), but bad because this gurarentee means no retries. If the bits are wrong, they are discarded, so if your USB link is poor, you don't get subtly degraded signal, you get missing audio ( crackles and pops ).
OK, so most people have probably skipped now, saying "this idiot is just saying all USB sounds the same".
Well, what hasn't been touched on here is: All the USB products reviewed use adaptive mode USB. In this mode, the audio is transferred at a rate defined by the USB controller inside the PC, and the DAC MUST either: Track the audio clock - typically by using an adaptive PLL, that syncs itself to the ( jittery USB SOF signal ), or manipulate the data such that the DAC has the right amount of data to work with ( by again, measuring the SOF ).
So, the DAC either compromises it's timing integrity ( tracking the frame ), or the data integrity ( by changing the data to suit ) - the right sample at the wrong time, or the wrong sample at the right time.
This is why people are willing to spend more money on one of the (few) asynchronous mode USB devices. In this case, the DAC controls the rate of flow from the PC, such that it doesn't have to manipulate the data, and the USB is just doing what it should - transferring the data, with the important clock ( in the DAC ) being in control. NB - this can also be the case in Firewire and network DACs
I do think it's instructive that Alan compares the problems he found with USB to those of Toslink ( where the main problem is jitter ), but I find it disappointing the lack of any asynch USB products is not even touched on in the review.
your friendly neighbourhood idiot
Just in case you think I'm not a picky audiophile here's my system: JM Labs Alto Utopia Speakers with Shunyata Gemini Speaker Cable powered by a McIntosh MC402 Amp connected by Nirvana Cables to a Pass Labs Aleph P Preamp and a Meridian 508-20 CD Player. I listen exclusively to classical music and expect it to sound as it would in the concert hall. My most important critereon before going to USB (to more easily access my 3000 CD music collection) was: NO LOSS OF FIDEILTY from what I already have playing CDs directly. It is true that it took a while to find a DAC that would do this and, by the way, I tried the Weiss Minerva (which uses Firewire) and was not satisfied. I also tried a modified Benchmark, Audio Reserach DAC7, MusicStream and BelCanto. None worked as well as my CD player. Then I tried the Chord QBD76 and used the Poeimia USB cable manufactured by Ridge Street Audio. I also was using two 1.5 TbSeagate Freeagent E-Sata drives and set my screen to go dark in one minute. This set-up provides complete and total equivalence to playing the CD directly provided I use Foobar and resample to 96K with the USB output set for 48K running under Vista so no Kernal processing occurs. It is obvious to me that you conclusions were drawn without ever hearing the new Chord DAC or the Poeimia USB cables and may not have taken the trouble I did to use extrnal disk drives that are not USB, or to compare players with on-the-fly upsampling. If you really are in search of the truth, and not just writing an article with the limited information you have, I urge you to try the set up I have or, if you happen to be in Northern Caliifornia, you can come and listen for yourself. But I do not think you should let the article stand without an update.
....should I bother to read past page 1? Me knows none of this stuff, I just put a cd in and go, and generally like what I hear. BUT, I would like to go this route in the future, some day, and have started putting cd's on an external drive in FLAC format. Where to go from here would be great, I know many of you know, could you share some links, etc?
Thanks, Richard
... digital communication media so the moment the author started talking about how "USB Sounds", they proved they had NO freaking clue what they were talking about and lost me. In fact, the only real aspect of the USB as it pertains to a sound system is the amount of bandwidth it can provide. Beyond that, it's completely up to the host/guest to determine how audio will sound. You could develop a device (such as a DAC) that had proprietary drivers for the PC that served audio in a streaming format over a USB connection, or you could use it to transfer files to another device, but ultimately, the USB connection has no control over the sound quality unless you found it was creating a bandwidth bottleneck.
Someone needs to check the author and perhaps replace him with someone more competent.
Please you guys that are a lot more into audiophile issues than me, give me feedback on what I am about to say. I have a dell dual core pc with a 150gig hard drive, OS is Microsoft XP Pro service pack 3 I use Zune software and my cd's were ripped using wma 240-255kps per second. I use an extrenal usb LG drive to rip. I listen to my collection using a peachtree audio DECCO connected to JPL L880s.
I have 12220 songs and 468 artists. I think my music sounds pretty good. I have a non pc based system using Sunfire amp, Nad transport and Magnepan IIIa's. I just don't think the Micrsoft OS kernel affects my music. I am not a Mcorosoft fanboy but working in the business world Microsoft server and client software rule. What is wrong with music servers? I tried MediaMonkey and I tunes but Zune to me has lots of nice features and is free. I think PC based music will only increase and its easy to back up.
Who is the guy who wrote this aricle and is he an authority?
Thx
It is obviously clear that Mr. Alan Taffel is unable to adapt to the changes that are taking place in the way music is stored, played and listened.
I am completely sorry he is total denial and I hope that one day he will accept the reality !
Until recently I did tech support for a Manufacturer's Rep. One of our lines was a high quality music server. As an audiophile I was fascinated by the hype over USB and asked the chief designer of our server line when they would add a USB output for use with newer DAC's. To my surprise, he told me that their research - and ears - told them that their existing S/PDIF interface sounded better and that there were "problems" with USB that hadn't been overcome. Shortly after this converstation I bought a new high-end DAC with one USB and two S/PDIF inputs. The DAC builder believes he has solved some of the USB issues and the forums are full of praise for the unit used with a variety of computers. I hooked up with Toslink for a first listen and quickly confirmed that this was, indeed, the best DAC I've owned to date. Then I listened to my music server via the S/PDIF coax and was rewarded with the best sound I've ever heard in my home. I couldn't wait to hook up my MAC Mini to try the USB input. When I did, I was astonished to hear the same thing Alan described in his article - a pleasant sound (in my system, a step above though similar to the Toslink) but slightly cardboard-like and two dimensional. I experimented with settings, etc. for a while then concluded it must be the computer. It certainly wasn't close to the S/PDIF interface. So, all the technical articles and rave-ups aside, I for one can confirm that even with a top USB contender, good old S/PDIF still sounds better.
I'm really not trying to be an advocate of one thing over another, but your post leaves me with a few questions....
You play back music on your server via SPDIF and it sounds great. You then connect a Mac via USB and it doesn't sound as good. So:
Same content? (i.e. same file?)
The Mac has a rate converter built in, so same sample rate? iTunes in particular makes it easy to switch in one or sometimes two sets of unnecessary rate conversion
Was the DAC from one of the 3 manufacturers to do USB properly so far?
I have to ask, because the Bryston( for instance ) gets good write-ups as a DAC, but uses a poor USB implementation. This thread is about an article that dismisses USB full stop, so getting the facts straight is important....
your friendly neighbourhood idiot
Hi Savant (since I don't know you, I refuse to call you Idiot)...
Didn't mean to imply that my experience was definitive - just surprising to me. All files were the same, all ripped from my original CD's. All music I have listened to for years. Apple was set on lossless per several recommendations. My server also uses lossless compression but it is proprietary. That could be the difference. Or not. I didn't experiment with USB cables, just used the one supplied by the DAC maker. I don't know if his is "one of the three" to do it right since I don't follow the forums all that religiously. I do know that my DAC has received rave reviews (it is not a Bryston). I'm not comfortable about publishing the brand since most others have reported fantastic results with USB and I don't want to start another flame war or in any way detract from his sales. In fact, I would not have responded to this article at all (I'm thrilled with my DAC/S/PDIF combo with my music server as source) if the world hadn't seemed to be insisting that the author (whom I don't know and whom I had barely heard of before) is either deaf or insane or both to dare take exception to the USB received wisdom. The fact that what he heard coincided with what I heard (and was backed up by the designer of my server before I bought my new DAC) prompted me to comment. If you insist that an increased effort toward proper set up of my computer/USB interface will yield results more in line with what other users (including other owners of my DAC) have experienced, I will take your word for it. But, frankly, I don't intend to bother. I've got too much music to listen to and too little time to do it.
Hi Bill, don't worry - I get called an idiot every day :)
I'm not at all trying to question what you heard - at the end of the day, all of this is about how your music is presented to your ears, and this seems as good a place as any to try and help... I perfectly understand how you don't want to waste time messing about with settings etc., but in case you do...
So, on a Mac, you must use AudioMidi setup to set the output sample rate to be the same as your audio ( so normally 44.1k ), and it's important to do this before you start iTunes ( if you've already started iTunes, you've got to QUIT iTunes from the menu - pressing the red dot thing doesn't make it quit ).
I don't think it's any secret that there are three manufacturers doing USB "properly" in the high-end market:
Wavelength ( first to the market, single-ended triode DAC )
Ayre ( license code from Wavelength, zero-feedback, solid state, asymmetric filter )
dCS ( very expensive, done their own way apparently )
The above three have had equipment reviewed/measured by Stereophile, where their jitter claims seem to hold water
ALL the other high end DAC manufacturers either use a really horrible USB->I2S converter ( PCM270x ), which you can identify by the fact they only do 16/32,16/44.1 and 16/48, or a better implementation provided by centrance ( does up to 24/96, but will be more jittery than async, and more jittery than most SPDIF implementations )
your friendly neighbourhood idiot
And don't forget Empirical Audio: usb converters and DAC with usb
Alan,
Get a hold of an Emu 0404 24/192 USB DAC. You can find 'em on ebay for well under $150.
Then download the J River Media Jukebox; the download is free.
Hook up the DAC to your laptop via USB but use ASIO to move the data.
Get a hold of the free 24/96 downloads from HD Tracks or any other hi res flac or wav files you may have on hand.
I guarantee you will be smiling at what you hear, and I daresay you may gain a new found appreciation for the concept of bang for buck.
...and I read Alan's article. I do not feel qualified to dispute him, but I have a suggestion - something that he might try if he feels that USB isn't ready for prime-time audiophile usage.
Aside: I'm relatively convinced that most USB implementations are not bitperfect outputs. In my office, my HP Z400 runs MediaMonkey 3, and sends USB output to a Pro-ject USB box. The analogue outputs of the Pro-ject go into the line inputs of my Monitor Audio iPod dock. It's totally usable, sounds generally great, and quite often - when generating WAV output from a FLAC file of a good recording (e.g. Sara K's "Closer Than They Appear") - it raises goosebumps. Let's just say that, whilst working, this system is enough for me.
My main floor system is a bit upscale from the office box. It has two source inputs: A Musical Fidelity A308 CD player (somewhat dated, but still very smooth) and an older IBM Pentium 4 running MediaMonkey... and here's the difference: the IBM sports a C-Media 8738/8768-based soundcard (the very inexpensive SIIG Soundwave PCI 7.1). The analogue outputs of the card are not used - and therein lies the difference, IMHO.
I have disabled the motherboard sound hardware, and added a set of Open Source sound drivers for the SIIG card from http://code.google.com/p/cmediadrivers/, built by a gentleman who goes by the moniker Dogbert. The story behind the driver development is outlined on his site, and also on the Hydrogenaudio forums. I have corresponded with him, and he is quite clear that he spent the time building this bitperfect driver because of his dismay with the stock C-Media chipset drivers.
So, in my system, the S/PDIF output of the SIIG card feeds a vintage GW Labs DSP reclocker/resampler. Digital output from the DSP goes out balanced to my MSB Nelson Link DAC III - another vintage device, but one that has weathered time well. Analogue output from the Link DAC goes out via some extremely overpriced cables to a Musical Fidelity X10v3 Tube Buffer. Another set of overpriced RCA cables connect the outputs of the X10v3 to the inputs of another vintage (but completely overhauled) piece, a McIntosh MA6100 integrated amp (reworked by Audio Classics last spring). Speakers are Paradigm Studio/20 v.4s. Speaker wires are solid OFC copper with an ultrathin laquer insulation. Not state of the art, but clearly not consumer crap, either. :-)
The result? To my ears - the IBM with the bitperfect S/PDIF output - playing FLAC files ripped with an AccurateRip(tm) supported interface - sounds as good as - or better than - the source CD in the A308.
Thoughts? I suggest that Mr. Taffel tries this approach - bitperfect S/PDIF out from a PC instead of USB. Total cost should be about $31 - that's the cost of the SIIG soundcard from Amazon. And of course, Dogbert's driver is free.
My $0.02,
-Don Strimbu
Jeez... when will people stop claiming themselves to be computer audio experts and generate articles like these?
I'm running M-Audio gear. PCI cards, USB and Firewire. All are AD and DA. Using XP Pro, Cubase, Foobar, etc.. The USB Transits sound damn good.
You see, the trick is to use the ASIO drivers. Done that way, trust me, you wouldn't be hearing the difference that you claim.
My reference recordings are my LPs recorded at 24/96 WAV into a specially built PC running XP Pro, M-Audio Delta 24/96, my Linn LP12, Ittok, Grado Reference Sonata, Grado preamp and CJ preamp. The PC is configured as a "digital tape deck"
Anyhow, we got like six audio systems at home, all computer based nowadays, plus my Transit/Grado amp/Grado SR325 portable set up.
USB sounds fine if you know how to configure your PC. Maybe next time you should spend some time on the head end, huh? It's like using an 8-track as head end to my ARC amps and Maggies. Caca in, caca out, dig?
Regards
TonyE
TonyE Has the right Idea. And so do many others.
Just use a pro firewire deck and be done with it. There are many to choose from. USB IS GREAT! For convenice. Yes it can! DO IT ALL. But if you know more about the inards of it. IEE 1394 A/B just makes more sense for most people.
Problem is always configuration. Connection Second, Hardware is third. . And the OS is more of a culprit then any thing on this list. I wish I could write an artilce and have it published in TAS about this. It seems like people just take connectors for granted.
A good example of this is when your computer runs CRON scripts and backround stuff. It can take your audio experice from Nirvana to Burnt Biscuit in no time. Having a MAC makes no diffrence. LINUX can also be troublesome as well.
If you plan to make a PC be your media center, THere is allot of work involved in making it trouble free. People In IT will understand. (And those in the Prouduction side of things. Others will just be happy that it plays and works great.
I never listren to music from cd's anymore unless I am in the car. Best bet is to rip it to the HDD, with Exact Audio Copy, and then use a very good standalone player to an EXTERNAL FIREWIRE DAC.
And in terms of power supplys, Stay away from li-Ion packs and stick with lead acid, with a very good 14.4 volt chargoing system. (Not no cheap wall wart) and disconect it when not needed. (most firewire stuff can be bus powered but pro stuff always has power on board anyway. )
I hope this makes it easy on a few of you. REMEMBER, the reason why HIFI stuff cost so much is becase someone put there time and effort to make sure that the stuff that you buy works like you want and elminates compromise.
Here is another thing I just hate. USB cables that cost more then $100 bucks. Why? Chances are the conectors on them are more of the problem then the cables. Be sure to use the Caigs! on all contacts. AND BE CARFULL WITH APLICATION! The stuff is conductive and can do more harm then good causing a short circit or grounding of contacts.
Viddy this me dorogoy malenky brothers! Can I believe me glazzies?
Big malchik Alan T govoreeted that spdif (AKA shitty pisspoor digital interface) is the the best and only way to play me Beethoven symphonies!
Your humble narrator disagrees, he prefers a piece of carbon rock to connect him to his music residing on pop-disks to any of the others.
As the malchik who has no nozh to grind (but maybe a britva) and no filly in this race (but many horrorshow pop-disks) I must declare:
This supposed article is a bolshy crock of chepooka cal and it smells of the varsay and non may abide it (counting the responses)! And it makes me a bit bezoomny.
If a starry veck reviewer who only does seedee had globbed his lapas on several horrorshow record players and dumped them on spare chairs.
If he somehow gotten sound out of them using a grotty starry cheap preamp from the bottom of the basement (only one to still have a phono-stage), should he be praised of his efforts?
If he then ittied down to the local 2nd hand record boutique and kupetted a bolshy stack of the 50 pence rejects from the bargain bin and then proclaimed based on his slooshing to all this “that at this stage in its development, LP does not rise to the level of an audiophile-quality”!
Should he be applauded by his readers or would he invite a swift great big noga into the yarbles from such strong molodoy malchiks such as Alex the Droog?
Just like any veshch in horroshow stuff, maintenance and dobby setup is needed. For horroshow computer audio as much as for the horrorshow speakers, dobby valve amplifers and record players that interessovat Alex the Droog so much.
Any old veck may do as he pleases.
A reviewer malchik on the other rooker, on whose govoreeting a golly lot of malchiks work out how to spend their hard crasted cutter and pretty polly can not.
He must take care to smot how to get horrorshow results from what he reviews.
Hey, Alan T, you big Malchik, can you really govoreet you did so horroshow on that?
Thought not.
And Mr. Editor, you let that cal get into print, what is your appy polly loggy?
Time to meet my droogs over at the Korkova, I’ll ask for extra knifes in the Moloko this notchy to sharpen me up, after reading this.
Yours, Alex the Droog
What a joke! I've read Alan's article in The Absolute Sound magazine (TAS). The guy is completely out of his element and is barely "computer literate". Most anyone here and on the various online computer audio forums are way ahead of Alan. He has no credibility in computer audio. Perhaps he should stick to reviewing $500 AC power cords.
What a joke! I've read Alan's article in The Absolute Sound magazine (TAS). The guy is completely out of his element and is barely "computer literate". Most anyone here and on the various online computer audio forums are way ahead of Alan. He has no credibility in computer audio. Perhaps he should stick to reviewing $500 AC power cords.
What a joke! I've read Alan's article in The Absolute Sound magazine (TAS). The guy is completely out of his element and is barely "computer literate". Most anyone here and on the various online computer audio forums are way ahead of Alan. He has no credibility in computer audio. Perhaps he should stick to reviewing $500 AC power cords.
1. Via an Analogue to Digital conversion process an Artist records an analogue sound to a digital medium that then contains a digital file - eg CD, computer file that can be transported and duplicated millions of times
2. The digital medium is placed in a vehicle (PC, Mac, USB, SPDIF, Internet, CD transport) that transfers the digital information (at a time of my choosing) to a device that converts the digital information back to an analogue signal - the DAC
3. Unless the vehicle is broken, there can be no variability in the information (repeatable delivery of identical information), only the speed/pace at which the information is delivered
4. If (a) the DAC "buffers" the information received and controls the pace at which it converts it to back to an analogue signal and (b) the delivery vehicle never allows the buffer to empty (jitter), then no matter which vehicles are involved in the delivery, the outcome must be the same.
So the trick is how to deliver a digital file from the original recording (1) to the DAC (2) reliably, repeatably and without corruption. All the talk about computer settings, music software, cable types etc suggests a premature digital to analogue conversion process or alteration (corruption) of the digital information before it reaches the DAC. For example, if you can change the volume of the music delivered by your vehicle, then you have already corrupted the original digital information before it reached the DAC. If you can hear a difference by changing a USB cable, then you are not listening to a digital file faithfully delivered to your DAC.
Unlike a cheaper CD transport, a computer can easily read and transfer a digital file in exactly the same way, millions of times without missing/changing a single bit of information in the process. If it couldn't, every Word and Excel document in the world would be unusable.
Over to the DAC makers. How do I get the digital file from the medium (CD, internet etc) to you via a PC/Mac for you to "do your thing"?
Sonic Luddite-
No real argument with you, but your comments are overly simplified:
1. Most PC setups resample the music before converting it. Not all resamplers do an errorless job or a good job, and the act of resampling changes the bits - either adding or subtracting. It's easy to do an a/b comparison of the before/after and hear the difference if you have good equiment. Music heard through Windows Kmixer and through ASIO drivers don't sound the same. Some DACs upsample, and apparently some do it better than others "on the fly". All of these things, and other similar factors, mean that the "the bits" you start out with aren't necessarily "the bits" you end up with.
2. As you hinted, most of the other problems are jitter due to timing (clock errors). This is inherent in the way most people transfer music data and turn it into actual music. Even the common S/PDIF connectors add jitter to the signal (if you don't believe me, read up on I2S interfaces), and there are lots of other jitter sources in most PC to stereo or CD to amp setups. These can be reduced (not totally eliminated) but that involves a cost in either time, money, or both.
Exactly firedog. If you think about a CD player as having 2 functional components - (1) the transport to read the digital file and (2) the DAC - the way forward for PC based high end audio is to purely replace the CD transport function using a device (PC) better suited to reading, storing, cataloging and presenting digital information.
If the PC has to interpret the data (sampling) before presenting to the DAC then IMHO we've failed.
A major source of confusion in this debate stems from the fact that PCs/Macs contain an audio subsystem (software drivers, DACs, amps etc) to enable the PC to directly drive analog PC speakers, headphones sockets etc. Hence the talk of midi settings, drivers, cables etc.
For true high end audio we have to be able to bypass all of the PC's audio hardware and software and get back to the pure task of a surrogate CD transport managing digital data.
I don't qualify as an audiophile yet but I use iTunes to rip with Apple Lossless, an AppleTV to stream to a Pioneer VSX-LX60 for DAC, pre-amp and feed a Whatmough Impulse sub, and a Yamaha AX750 to drive a pair of custom built Dynaudios. To my ear the sound quality is superior to playing the source CD through a Panasonic DMP-BD10A.
Once I better understand the problems of managing digital music I plan to invest in a Benchmark or similar external DAC.