Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Regarding the lossless codec thread... (Read 6257 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Regarding the lossless codec thread...

Hello all,

This is my first post on this forum and so I want to let you all know before I say anything that I'm not here to offend or preach to anybody.  But I do feel the need to say something, nonetheless.

I found Zen-Infinity's thread (now strangely closed) to be intriguing.  I always enjoy it when someone encourages us to look (or listen) beneath the surface of things normally deemed to be "perfect".

Quick story for those who care:  I work for a mastering studio and regularly create PMCD masters for replication plants.  We have a PC-based machine that tests CD media for quality.  I'm sure some of you have seen these.  Well, when I first saw this thing, I flipped out!  I'd never heard of such a thing!  Here was a machine that could GUARANTEE that I'd created a flawless digital master that could be sent to the plant to create commercial CDs.  Sounds great, right?  Of course it does.  No more would it be truly necessary to LISTEN to the CD I'd just made...the printout from the analyzer told me that everything was cool; why would I question it?

That was, until I had a plant call me on the phone and tell me that a CD-R that I'd sent them had uncorrectable errors on it.  WHAT?  BUT I TESTED IT IN MY CD ANALYZER AND IT TOLD ME IT WAS PERFECT!  I went back and forth with the engineer at the plant about the possibilities here...who's machine was wrong, his or mine?  Either way, I re-ran the PMCD.  My CD analyzer is no longer something I look to as gospel.

What I'm getting at here, is that it's dangerous to just ASSUME anything...to put all of your eggs in one belief-system-basket, so to speak.  I think we tend to turn our back on digital too much and too often.  And that, my friends, is when it tries to kill you.  No, but seriously...you know what I mean.  As soon as we shut our ears off and our minds off and just go by what a computer printout tells us, we are negating progress.  I mean, WE created these machines.  And if WE'RE not perfect, how could THEY be?

For the person that said something like "anyone with half a brain cell wouldn't bother to compare lossless codecs", well you sure don't sound like someone that's about to revolutionize audio technology as we know it.  Think about it.  It sounds like a 17th-century physician saying that there's no way that something could be isolated from mold to create something to kill bacteria.  Or, worse yet, a blind man telling me that there's no such thing as a rainbow.  See what I'm getting at?  If you haven't tried something, you have no real business making a judgement.  At least Zen-Infinity is willing to investigate.  Doesn't sound like a lazy person to me.  Sitting back in your armchair and invalidating someone else's findings with nothing to back it up sounds not only lazy, but closed-minded.

At any (bit) rate I promised not to preach.  But, at the end of the day, I really hope that it is everyone's intent here on this web site to REACH THE POSSIBILITIES OF AUDIO TECHNOLOGY.  That means asking questions.  That means questioning the supposedly infallible.  And that means listening.  It's not about the argument (who's right or who's wrong), it's about the experimentation that leads to the truth.

It's no coincidence that audio quality in general on the mass market is in the toilet.  Only a more inquisitive mindset and closer attention being paid to actually listening will have a positive effect on this current state.

Regarding the lossless codec thread...

Reply #1
I don't quite see what you are getting at - lossless compression preserves all bits perfectly - it is an exact duplicate when decoded. 

I believe you are questioning the methods used to extract redbook cd audio, and error recovery rates involved in that process, not errors involved in lossless compression. 

Apples and Oranges.

Regarding the lossless codec thread...

Reply #2
Quote
We have a PC-based machine that tests CD media for quality.  I'm sure some of you have seen these.  Well, when I first saw this thing, I flipped out!  I'd never heard of such a thing!  Here was a machine that could GUARANTEE that I'd created a flawless digital master that could be sent to the plant to create commercial CDs.


Anyone with half a braincell and some knowledge of optical disc drives will tell you that there is no guarantee at all ... since any disc quality assessment does solely depend on the specific read device's capabilities.

Quote
That was, until I had a plant call me on the phone and tell me that a CD-R that I'd sent them had uncorrectable errors on it.  WHAT?  BUT I TESTED IT IN MY CD ANALYZER AND IT TOLD ME IT WAS PERFECT!  I went back and forth with the engineer at the plant about the possibilities here...who's machine was wrong, his or mine?  Either way, I re-ran the PMCD.  My CD analyzer is no longer something I look to as gospel.


Maybe your analyzer was out of calibration ... or maybe the built-in drive has been damaged ...

Quote
What I'm getting at here, is that it's dangerous to just ASSUME anything...to put all of your eggs in one belief-system-basket, so to speak.


You are right on that point ... that's why HA.org members are e.g. encouraged to perform double blind ABX testing for sound quality assessment.

Quote
For the person that said something like "anyone with half a brain cell wouldn't bother to compare lossless codecs", well you sure don't sound like someone that's about to revolutionize audio technology as we know it.  Think about it.  It sounds like a 17th-century physician saying that there's no way that something could be isolated from mold to create something to kill bacteria.


A lossless codec is lossless for good reason ... a PCM file is (mathematically spoken) nothing more than a huge matrix that can be compressed with a given algorithm ... an algorithm that has proven itself to be lossless multiple thousand times in both listening tests and data integrity verification.

If codec developers and members wouldn't use their ears to listen, we would not have made that progress we face today. Think about it ...
The name was Plex The Ripper, not Jack The Ripper

Regarding the lossless codec thread...

Reply #3
Quote
For the person that said something like "anyone with half a brain cell wouldn't bother to compare lossless codecs", well you sure don't sound like someone that's about to revolutionize audio technology as we know it. Think about it. It sounds like a 17th-century physician saying that there's no way that something could be isolated from mold to create something to kill bacteria.


This sounds like zealot Audiophile speak here.  It doesn't make a difference.  Besides researchers have already revolutionized audio technology in this field. Linear Prediction Coding and Residual Anaylsis is a powerful tool  .  A lot of the more exciting research is Psychoacoustics, etc. 


Quote
You are right on that point ... that's why HA.org members are e.g. encouraged to perform double blind ABX testing for sound quality assessment.


Yes, something that many Audiophiles always moan and complain about. 

Quote
It's no coincidence that audio quality in general on the mass market is in the toilet. Only a more inquisitive mindset and closer attention being paid to actually listening will have a positive effect on this current state.


In what reguard? are talking about mastering of CD's released by the industry? The fact that most recordings are compressed and clipped to death? As was mentioned above it sounds like you are confusing apples with oranges 

Quote
I really hope that it is everyone's intent here on this web site to REACH THE POSSIBILITIES OF AUDIO TECHNOLOGY


That is the whole goal of this site.  It was also started to educate people about audio technology and other forms of interesting Research in the field. Often times people aren't passionate when it comes to audio related Research, because they are often to aesthetic and only care about what's in front of their face, visual related stuff.
budding I.T professional

Regarding the lossless codec thread...

Reply #4
Here we go again...

Quote
Quick story for those who care:  I work for a mastering studio and regularly create PMCD masters for replication plants.


Ah yes.. another "I'm a professional, so I know what I'm talking about"

Quote
What I'm getting at here, is that it's dangerous to just ASSUME anything...to put all of your eggs in one belief-system-basket, so to speak.


Agreed.  The adherence to audiophile quackery in the face of all other formal systems which often trump such superstitious nonsense is disconcerting, to say the least.

Quote
I think we tend to turn our back on digital too much and too often.  And that, my friends, is when it tries to kill you.


I really have no idea what you are trying to say here, and I don't think you do either.

Quote
No, but seriously...you know what I mean.


No, really, we don't.

Quote
As soon as we shut our ears off and our minds off and just go by what a computer printout tells us, we are negating progress.


Translation: we need to "trust our ears" and "keep an open mind to audiophile nonsense."  That's what you meant to say, isn't it?

But in reality, superstitious audiophiles really are guilty of shutting off their minds.  The very idea of critical thinking seems foreign to most of them.

Note:  Believing something because some audiophile magazine seems to think its fashionable, while disregarding all science, math, and just plain common sense in the process, is not using your mind.

Quote
I mean, WE created these machines.  And if WE'RE not perfect, how could THEY be?


Something you need to understand is that there is a huge difference between a mechanical process and a mathematical proof.  You can prove something to be correct in a formal mathematical sense and then it is correct by definition and is irrefutable.  This is something beyond reproach.

Mechnical processes, on the other hand, are different.  You can prove that a mechanical process works correctly within a given threshold of tolerance.  However, you can't predict all aspects of nature.  Because of entropy, mechanical processes are inherently unreliable in the long term.  Your test machine is an example of something susceptable to this.  A mathematical proof of the losslessness of an encoder, is not.

So many audiophiles seem to misunderstand the very nature of digital audio.  They mistake problems at a mechanical level with problems at a mathematical level.  They also misunderstand almost all fundamental aspects of what actually goes into making digital audio work.  Then, they come make foolish posts like yours.

Regarding the lossless codec thread...

Reply #5
Quote
For the person that said something like "anyone with half a brain cell wouldn't bother to compare lossless codecs", well you sure don't sound like someone that's about to revolutionize audio technology as we know it.


First of all, audio technology was "revolutionized" a long time ago.  The audiophiles just were too stuck in their vinyl/tube/analog only mindset to notice.  That, and the lack of desire to really understand digital audio, has led to the spread of all kinds of stupid myths like the thought that "lossless isn't really lossless."

Quote
Think about it.  It sounds like a 17th-century physician saying that there's no way that something could be isolated from mold to create something to kill bacteria.  Or, worse yet, a blind man telling me that there's no such thing as a rainbow.  See what I'm getting at?


Yes, I see what you're getting at, and it's wrong.  The big difference with lossless audio is that the entire concept is based upon irrefutable mathematical proofs.  If you don't understand the significance of this, go learn, and then come back when you are capable of arguing about this topic in a sensible fashion.

Quote
If you haven't tried something, you have no real business making a judgement.


If you can prove the behavior of something, you don't need to try it in the empirical sense.  Admittedly, not very many things can be proven this way.  But data compression, which is all math, operating upon digital audio, which is just a discrete mathematical representation of sound, is fortunately one of them.

Quote
At least Zen-Infinity is willing to investigate.  Doesn't sound like a lazy person to me.  Sitting back in your armchair and invalidating someone else's findings with nothing to back it up sounds not only lazy, but closed-minded.


And ranting about something you have absolutely no real understanding of is foolish.

Quote
At any (bit) rate I promised not to preach.  But, at the end of the day, I really hope that it is everyone's intent here on this web site to REACH THE POSSIBILITIES OF AUDIO TECHNOLOGY.  That means asking questions.  That means questioning the supposedly infallible.  And that means listening.  It's not about the argument (who's right or who's wrong), it's about the experimentation that leads to the truth.

It's no coincidence that audio quality in general on the mass market is in the toilet.  Only a more inquisitive mindset and closer attention being paid to actually listening will have a positive effect on this current state.
[a href="index.php?act=findpost&pid=353440"][{POST_SNAPBACK}][/a]


If anything will change audio for the better, it will be the decline of audiophile quackery and superstition.  When people start getting a clue about the way audio -- digital audio in particular -- works, and when the "hi-end" market with all it's snake oil implodes in on itself, then perhaps we can move forward and start making some real progress again.

Regarding the lossless codec thread...

Reply #6
Quote
For the person that said something like "anyone with half a brain cell wouldn't bother to compare lossless codecs", well you sure don't sound like someone that's about to revolutionize audio technology as we know it.


Actually it sounds like someone who understands what "diff" does.  Its important to pick the right tool for the right job.

Regarding the lossless codec thread...

Reply #7
Ugh. We're basing this discussion on replication plants?

Those guys are bona fide nuts, although they have their reasons. I think they have nothing to do with sound quality. They have more to do with the replication gear they're using (which could be 1-2 decades old) and whatever spec checks they have in place to guarantee against any remotely theoretical compromise in output quality, jitter being among the bigger issues. True story: it's not unheard of for an 8x-burned data CD master to get rejected but a 1x from the same burner to be accepted, even though said 8x master works on virtually every CD-ROM in existence (and the final pressing will be reclocked anyway!).

Do not confuse "rules that a plant enforces on you to press your master" with "rules to create the best sounding CD." Those concepts are entirely independent.

Regarding the lossless codec thread...

Reply #8
Quote
Ugh. We're basing this discussion on replication plants?

Those guys are bona fide nuts, although they have their reasons. I think they have nothing to do with sound quality. They have more to do with the replication gear they're using (which could be 1-2 decades old) and whatever spec checks they have in place to guarantee against any remotely theoretical compromise in output quality, jitter being among the bigger issues. True story: it's not unheard of for an 8x-burned data CD master to get rejected but a 1x from the same burner to be accepted, even though said 8x master works on virtually every CD-ROM in existence (and the final pressing will be reclocked anyway!).


I have an inquiry about cd duplication plants -

wouldn't it make much more sense (from a technical pov) to burn a master cd as a data  track, with the audio cd stored as an image with an MD5 checksum?

Regarding the lossless codec thread...

Reply #9
Quote
I have an inquiry about cd duplication plants -

wouldn't it make much more sense (from a technical pov) to burn a master cd as a data  track, with the audio cd stored as an image with an MD5 checksum?
[a href="index.php?act=findpost&pid=353491"][{POST_SNAPBACK}][/a]

If you're at that, might as well use parity files to allow data correction in case of corruption...

The goal in duplication plants is to simplify the process to the maximum.  Therefore, the master has to be a perfect copy of what the press will make.  Also, your technique poses problems with CD-Extras or CDs with software on em.

Regarding the lossless codec thread...

Reply #10
Hello, Hi-Freak, and everyone.
Zen-Infinity, as well as anyone else, is welcome to perform listenings tests, as long as he performs them in a usueful way. Listening tests performed casually by very experienced audiophiles or proffesional engineers as well as anyone else have been proven not to be reliable.

Here are two good examples :
An experienced audiophile compares his top-of-the-art Bryston amplifier versus a lower end Onkyo. At a given time, the Onkyo is working, but the listener believes that it's the Bryston. Of course, he flames down the amplifier as completely bad sounding, while he is listening to his own Bryston ! http://www.diyaudio.com/forums/showthread....&threadid=12752
An experienced sound engineer explains that MP3 is very bad sounding and that for his level of expertise, it is by no way acceptable. Some random wav files are sent to him, one of them comes from a CD,the other from lossy compressions of that one. He mistakes the files, and find a lossy version to sound better than the original one.
The claims : http://www.hydrogenaudio.org/forums/index....562&#entry47562
The results : http://www.hydrogenaudio.org/forums/index....030&#entry48030

These are examples, if there is need to provide some, that a scientific way to perform listening tests is required to get valuable information about how something sounds to someone's ear. The double blind protocol, followed by a calculus of the probability of guessing is the way used in scientific research, and the compulsury way to go in this board, as stated by the terms of service. Subjective sensations experienced during listening are not accepted as a proof that something sounds like this or that, whatever the background of the listener.

Here is a short explanation of the ABX protocol for blind listening tests : http://wiki.hydrogenaudio.org/index.php?title=ABX
Here is a more detailed explanation of the whole process : http://www.hydrogenaudio.org/forums/index....howtopic=16295&

Regarding the lossless codec thread...

Reply #11
Oh, I forgot : a CD analyzer doesn't tell the number of errors present on the CD, it tells the number of errors that occured when it tryed to play the CD. That's why you can get different results than your plant. The drive of your analyzer was capable of reading your CDR without problem, while the drive of the plant's analyzer could not. Maybe its hardware is getting old.

Regarding the lossless codec thread...

Reply #12
I am closing this thread for now. If there is going to be further discussion about this, let's please do it in a scientific, TOS8 compliant manner.