[vdm]

I just read on "The Computer Audiophile" website that he will be writing a regular column in the TONE Audio magazine about this subject. This is where I'd like to raise my concerns. Based on several postings of the ComputerAudiophile it is obvious that he is very uninformed and without a technical/computer background. A case in point is his recent postings on lossless compression which made it clear that he doesn't understand it at all.

Don't get me wrong. I don't have anything against the ComputerAudiophile - I salute his enthusiasm for the subject. It's just that I'd prefer that the informed teach the uninformed about a fairly new subject. In this case I feel it will be like the blind leading the blind... not an ideal situation. If he continues with this endeavour it is important that he does not purport to be a technical expert on the subject and should make that clear in his columns. I do appreciate him keeping us informed with news on the subject - something that he seems to be good at...

[The Computer Audiophile]

Hi vdm - Thanks for your feedback. Can you be more specific about your comments? I actually do have a very technical computer background and I have setup and managed global networks around the world.

Can you let me know what makes you think I don't understand lossless compression at all? I have a very good understanding of it and I'm interested to see what makes you think I don't have an understanding at all.

Is there anything else that makes you think I am uninformed about the subject? The term expert when it comes to technology is quite an ambiguous term although it may seem otherwise. There is always someone who knows more about everything and those who think they know everything actually know nothing. (Certainly not a comment about you.)

I look forward to your response and maybe we can get a good dialog going so we are all on the same page. Also, feel free to contact me personally and I'll be happy to chat.

[vdm]

Hi Chris,

Firstly, I don't mean any disrespect since I have not met you in person and I don't know your background. Just for reference I have a Masters Degree in Electronics Engineering specializing in Digital Signal Processing.

Ok, regarding your question, let's start with this posting where you suggest that people re-rip their whole audio collection because they have used Apple lossless compression:

http://www.audiocircle.com/index.php?topic=49258.msg450685#msg450685

This is exactly what I referred to in my original posting. As a result of your posting these people may just spend many hours, if not days or weeks, re-ripping their whole collection. All that is necessary in order to get back the original uncompressed versions of their ripped libraries are to use one of many freely available tools that can reconvert the lossless files back to the originals (bit for bit accurate) in a matter of minutes, rather than days, weeks or months. So, in my opinion that is irresponsible and very confusing advice to those that are uninformed...

[The Computer Audiophile]

Thanks for the reply and a little background on yourself. You should be a patent attorney with those credentials. Digital signal processing is an area where patent attorneys who go for their EE after law school always have trouble.

My suggestion to re-rip a library because it was originally ripped in Apple Lossless is based on conversations I've had with manufacturers in the high end industry. Would the files be bit for bit identical, using our current methods of measurement, by re-ripping or converting? Of course. The poster asked for my opinion and I stand by that opinion. Some people notice a difference between files that have never been compressed and files that have been compressed and converted to uncompressed. Whether or not this difference is in their head or there really is a difference that takes a $200k system to hear is somewhat irrelevant. If a listener thinks it sounds better, then it sounds better.

Surely this is not the case for everyone and this is where I think I should have suggested additional options to the original poster. It may have been more prudent to suggest converting one album and re-ripping the same album. Then have the poster listen for himself and decide what he wants to do.

"...it is obvious that he is very uninformed and without a technical/computer background."

I do take issue with that statement, but I think we may have more in common than is readily evident and we probably agree about much more than we disagree. I hope you continue to read TONE and give my column a chance. I honestly look forward to your continued feedback when you see fit.

P.S. Thanks for saving both of us from a classic forum pissing match. we'll all get more out of this thread with a civilized conversation.

[TONEPUB]

Gotta back Chris up on this one...

Talked to engineers at Meridian and Wadia about this and they both concurred, if you want uncompressed files, to go back and re-rip for the best results.  It's just like using TIFF lossless files in Photoshop.  Even though they are supposedly "bit correct", there is some image loss.

[JohnR]

I'm not buying it. Either it's lossless, or it isn't. There's no real-time factor to consider here (jitter, real-time error-correction or interpolation etc). It should be easy to test under those conditions whether one file is equal to another.

[Crimson]

A little story, FWIW:

I started with computer audio back in '03 at which time I was using a PC and had everything ripped to wav. I then purchased a Macbook in early '05 and decided to migrate music ripping and playback duties to it. As can be expected, a lot of tag info was lost (roughly half) while porting the files over in spite of my best efforts. After doing some research, I decided to try ALAC and over the course of a month re-ripped a few hundred discs whose tag info had been retained during the migration resulting in duplicate tracks in both wav and ALAC. I spent a few months going back and forth between wav and ALAC (at times not even knowing which format was playing), and decided to rip a few hundred more. The end result was that by early '07 my entire library was available as both wav and ALAC. In October of last year I wiped out all the wav files. Am I hard of hearing? Is my system not up to snuff? Well, at least I don't think so.

[sts9fan]

So that lost data should be able to be defined.  Lets see it.

[dwk]

Gotta back Chris up on this one...

Talked to engineers at Meridian and Wadia about this and they both concurred, if you want uncompressed files, to go back and re-rip for the best results.  It's just like using TIFF lossless files in Photoshop.  Even though they are supposedly "bit correct", there is some image loss.

Sorry, I just don't buy this. Whether you mean to imply this or not, you are basically saying "computers don't work". The idea that there can be some inherent difference in two identical bit patterns based on some notion of past history is just flat out not possible. Suggesting that compressing and expanding a file results in a 'different' file would also imply by necessity that the act of ripping it in the first place also alters it, as would copying it; for that matter, it would seem that even reading it off the disk would potentially effect a comparable change, implying that none of these operations could be performed repeatably and reliably.

If people believe that they hear a difference between a newly ripped wav file and a bit-wise identical version reconstituted from a lossless file, then either
a) they are imagining it
or
b) there are other factors at play (i.e. something is messing with the resulting clock)

Certainly I find it easy to believe that 'a' happens a lot. Audiophiles have proven themselves willing believers time after time, and self-deception is pretty much baked into the audiophile experience; I'm actually beginning to believe that it's a required part.

However, 'b' probably also comes into play. While I believe people tend to overblow things, it's concievable that things like having files placed on different portions of the disk or on different disks etc may induce head-seeking and thus power supply modulation, changes in emi/rfi signatures etc.  I find it difficult to see how things like this can be repeatable and predictable enough to establish a reliable percieved difference, but at least it has some potential mechanisms in play.

It is also of course possible that the 'new' rip results in different bits than the original rip, at which point we're dealing with a completely separate issue, although one that I'm sure gets lumped into the same bucket a lot of the time.

Digital audio really is pretty simple - we have bits, and we have a clock. Any differences in resulting playback have to be due to change/contamination of one of these two elements. Given how easy it is to mess up the clock in a computer audio system, it continues to baffle me why people still cling to the notion that there are avenues that can contaminate the bits when both theory and practice suggest that computers handle the bit problem perfectly well. (Of course it also continues to baffle me why people would continue to use a clock architecture which is so sensitive to corruption in the first place, but that's a different discussion)

[pardales]

I have been wondering about this too. I set-up my computer/hard-drive based system in the Spring of 05 and ripped my hundreds of CD's in ALAC - I chose ALAC because I am a MAC user exclusively and it worked well with keeping the file sizes smaller, is lossless, and because it works seamlessly with iTunes and getting album artwork. Recently I have re-ripped several if my CD's in .WAV to see if I could hear a difference on my system. I consider my system pretty revealing though I do not know if it is ultra revealing. I can certainly hear obvious differences in recording quality across CD's/files. Thus far, I have not been able to reliably distinguish between the ALAC files and .WAV.

Good discussion.

[Tirade]

I guess I'll chime in quickly here.

First off "Hi Chris". We've never met but recent CD ripping inquiries have pointed back towards your site as the reference so as they say in the myspace world "Thanks for the ad!"

Moving on, let me preface this next reply with "I have horrible ears". I don't have any hearing problems, I just can't hear some things that others can hear. I rip hundreds of CD's a week and some people prefer FLAC, some prefer ALAC, some want WAV (against my advice) and even others want MP3. When I rip CD's they are verified against (or submitted to) the accurate rip database. If a rip fails the accuracy check its then compared to the source to confirm they are bit-perfect (the AR database is full of errors, but its the only thing we have).

The end result is a song thats identical to the original. The key word here is "identical". If someone hears a difference between 2 bit-perfect songs then the song isn't the problem. iTunes won't play FLAC files so perhaps the user listened to all of his ALAC stuff in iTunes and then when he converted to FLAC he used Foobar. Perhaps his music player uses a different decoder for FLAC than for ALAC, I don't know, anything is possible, but the one thing that isnt possible are 2 identical songs sounding different. If they do, then they aren't identical.

Edit: Not a dig at Chris, this is more in response to the "Audio Gurus" who say reripping accurate ALAC files is better than converting them from accurate FLAC files.

[saisunil]

This is a very healthy and productive discussion.
I applaud everyone's effort by providing their points of views without getting hurt and hurting others

Cheers

Gotta back Chris up on this one...

Talked to engineers at Meridian and Wadia about this and they both concurred, if you want uncompressed files, to go back and re-rip for the best results.  It's just like using TIFF lossless files in Photoshop.  Even though they are supposedly "bit correct", there is some image loss.

Sorry, I just don't buy this. Whether you mean to imply this or not, you are basically saying "computers don't work". The idea that there can be some inherent difference in two identical bit patterns based on some notion of past history is just flat out not possible. Suggesting that compressing and expanding a file results in a 'different' file would also imply by necessity that the act of ripping it in the first place also alters it, as would copying it; for that matter, it would seem that even reading it off the disk would potentially effect a comparable change, implying that none of these operations could be performed repeatably and reliably.

If people believe that they hear a difference between a newly ripped wav file and a bit-wise identical version reconstituted from a lossless file, then either
a) they are imagining it
or
b) there are other factors at play (i.e. something is messing with the resulting clock)

Certainly I find it easy to believe that 'a' happens a lot. Audiophiles have proven themselves willing believers time after time, and self-deception is pretty much baked into the audiophile experience; I'm actually beginning to believe that it's a required part.

However, 'b' probably also comes into play. While I believe people tend to overblow things, it's concievable that things like having files placed on different portions of the disk or on different disks etc may induce head-seeking and thus power supply modulation, changes in emi/rfi signatures etc.  I find it difficult to see how things like this can be repeatable and predictable enough to establish a reliable percieved difference, but at least it has some potential mechanisms in play.

It is also of course possible that the 'new' rip results in different bits than the original rip, at which point we're dealing with a completely separate issue, although one that I'm sure gets lumped into the same bucket a lot of the time.

Digital audio really is pretty simple - we have bits, and we have a clock. Any differences in resulting playback have to be due to change/contamination of one of these two elements. Given how easy it is to mess up the clock in a computer audio system, it continues to baffle me why people still cling to the notion that there are avenues that can contaminate the bits when both theory and practice suggest that computers handle the bit problem perfectly well. (Of course it also continues to baffle me why people would continue to use a clock architecture which is so sensitive to corruption in the first place, but that's a different discussion)

[The Computer Audiophile]

This is a very healthy and productive discussion.
I applaud everyone's effort by providing their points of views without getting hurt and hurting others

Cheers

I had a real strong feeling that this thread was going downhill fast, but I'm happy with the direction it is taking.

[Jon L]

This is a very healthy and productive discussion.
I applaud everyone's effort by providing their points of views without getting hurt and hurting others

Cheers

I had a real strong feeling that this thread was going downhill fast, but I'm happy with the direction it is taking.

Chris, you said in the other thread:

"I spoke with a very high end manufacturer Friday night about this very topic and he said they have done tests with their gear and notice a difference every time. One problem is all the hardcore fanatics who jump all over those of us who prefer uncompressed music. Why would Joe Sixpack voluntarily put himself through the ringer in one of these forums by stating he notices a difference between uncompressed and compressed music."

With all due respect, this is the kind of statement that can create rumours, mislead newbies, and confound the innocent.  Frankly, you need to name the person at the manufacturer and what kind of "test" they claim they did (likely subjective listening tests).  Otherwise, it's worthless information.  I mean, there are manufacturers who claim all kinds of *interesting* things, aren't there?

In addition, I would like to point out that this particular discussion about lossless->uncompressed vs. uncompressed is a different discussion from lossless and uncompressed *sounding* different when played back via certain software/computer environment.  The latter *may* be true b/c there may be different software implementations, lossless playback software/plugin differences, and computer processing power/resource differences, etc.  It's at least intellectually conceivable.

However, IF one is using the same software and hardware to create lossless FILES (not played-back sound quality) and uncompressed FILES on the hard drive, they are identical.  One can convert one or the other a million times to lossless or uncompressed, but the resulting files are *STILL* be identical.  Jitter/timing info does NOT apply when talking about files like these.  Once again, if some manufacturer is sincerely claiming this is not true, I honestly am very interested in knowing who they are and would like a chance to make some inquiries to them.

[Panelman]

I agree. One of the problems of traditional audiophile belief or dogma is that it is very subjective and manufacturers can make statements that cannot  be refuted because of that subjectivity.  However with digital audio, at least file formats, it is possible to validate statements or at least provide data to support statements. Chris, in the USB cable thread on your site two "high end" manufactures make contradictory statements about the factors that can affect the sound of a properly functioning USB cable. While not taking sides on the argument it was very clear that one of them was  providing information that was not  valid as the data standard he described is no longer used. The fact that a manufacturer makes a claim means nothing without some data to prove it especially in an area where so many people don't have an in-depth knowledge of how digital works. I am one of those people however I know enough to understand that a computer file format can be tested and if something is bit perfect it is bit perfect. I also believe that you will provide a better service to the people who visit your site and your advertisers if you require some level of documentation to support technical claims made by posters about computer audio. Or at least request the data from posters. Otherwise people will stop visiting the site.

[kfr01]

I'm not buying it. Either it's lossless, or it isn't. There's no real-time factor to consider here (jitter, real-time error-correction or interpolation etc). It should be easy to test under those conditions whether one file is equal to another.

It is easy to test and it has been tested.  (as you know, lossless = lossless).

Both TONE and TCA provide a great product; they spread news, information, and views regarding our fine hobby.  The fact that they include subjective opinions regarding personal listening experiences is nothing new (see Stereophile, TAS, 6moons, etc.).

I agree, however, with Panelman when he says a better service will be provided to the readers and the advertisers if references, scientific tests, or at least sound reasoning are used when attempting to educate readers.  This holds true for Stereophile the same as it does for TONE or TCA.

TCA lost some credibility (likely only temporarily, he seems like a very sharp guy) to some population of his readership quite quickly when he made the postings regarding lossless v. lossy.  This is not an insult, it is simply true. 

For better or worse, once an individual is in the business of providing some level of technical information, the bar for that individual is raised. 

Me, I'm hoping that TCA and TONE rise to the challenge and become the best source(s) for Computer-as-source information in the audio scene.  Here's to healthy technical discussions, logical reasoning, scientific testing, and fact-based "publications". 

Cheers,
Karl

[The Computer Audiophile]

Hi Karl - Thanks for the feedback / suggestions / opinions. Thanks for the kind words about my site as well.

I look forward to providing the best computer-as-source content available anywhere. I know I can't please everyone all the time, but I hope to please some of you some of the time. I will continue to write about subjects I feel passionate about and I'm sure we'll all disagree here and there. Fortunately we all seem to have a similar passion for music, enjoying high quality reproduction of that music, and furthering this wonderful hobby of ours.

Thanks again for your civilized constructive criticism.

[pardales]

 Jon L. said:
 
"In addition, I would like to point out that this particular discussion about lossless->uncompressed vs. uncompressed is a different discussion from lossless and uncompressed *sounding* different when played back via certain software/computer environment.  The latter *may* be true b/c there may be different software implementations, lossless playback software/plugin differences, and computer processing power/resource differences, etc.  It's at least intellectually conceivable."



This seems to me, a key point in this discussion.

[The Computer Audiophile]

In addition, I would like to point out that this particular discussion about lossless->uncompressed vs. uncompressed is a different discussion from lossless and uncompressed *sounding* different when played back via certain software/computer environment.  The latter *may* be true b/c there may be different software implementations, lossless playback software/plugin differences, and computer processing power/resource differences, etc.  It's at least intellectually conceivable.

This seems to me, a key point in this discussion.

I agree pardales.

[mfsoa]

OK, so who's going to be the one to convert a disc from WAV to FLAC back to WAV back to FLAC etc a hundred (or more) times or so and do some bit-for-bit comparisons and listening sessions?

Seems like the more tech-savvy of you could whip up an automated system to do the conversions.

It'd be good ammo for one side the argument to show that the 1000th conversion is bit perfect and not audibly different.

-Mike