CD-R data integrity test

[jack70]

I've given my opinion about this issue in the past, but some here refused to answer to reason, so listen up. (PS:the world is not flat either) It's a recent article about something I've always wanted to do myself.... but never had the time. Here's the first part:

I don't know if it's because I grew up in the analog era where copies of copies were always of poor quality, but I'm squeamish when I make copies of copies. I remember the good ol' days when I would hook up 2 VCR's together and make copies of home movies or tv shows. The copies were always really fuzzy and washed out, even when using premium quality tapes and good machines. Today, with digital formats, I still get nervous about making copies of copies. For example, when I get files from friends on CD-R's, I always ask if it's a 1st generation CD or if it's a copy of a copy. (My friends have since learned to tell me that it's always a 1st generation copy even when it's not!)

The premise of this article is simple: To burn 100 generations of a CD and then compare the 100th generation copy with the original CD to see if the data is the same or if there are any differences. Now, I don't mean making 100 copies of the same CD; that would be pointless. I'm talking about making a copy of a copy of a copy, 100 times. In other words, take a CD (original) then copy it making "gen 1". Then copy "gen 1" which gives you "gen 2". Then copy "gen 2" which gives you "gen 3". Do this until you get to "gen 100". So you see, "gen 100" is a copy of "gen 99" which is a copy of "gen 98"...... all the way back to the original CD. You get the point.

I know, I know, a copy of a CD with data on it should be exactly the same as the original. But I can't help but feel nervous about the integrity of the data. That's where I got the idea for this article. Actually, I did this for my own peace of mind but thought that others might have the same concerns and would be interested in the results of some CD-R testing.

Click here (http://www.dslwebserver.com/main/fr_index.html?/main/100-gen-cdr-test.html) for the whole article (2 more short pages).

http://www.dslwebserver.com/images/100-gen-cdr-test/total.JPG

[Jim Clark]

I don't know if anything new has been broached here, I was under the impression that this was all common knowledge by now. There are always going to be a few holdouts I suppose and I doubt this will do anything to change their opinion. There are people who swear they can hear an audible difference between an original and a CDR though I've never seen it demonstrated nor have I come across any sort of plausible (or even inplausible) explanation for their beliefs. Of course many of these are the same people who swear that Black CDRs have inherently better sound. I've even seen internet forum postings that claim black CDRs sound better than the original. All poppycock of course, except for the tales I believe in.

jc

[ocrazor]

Good test! But-- it doesn't really apply to audio cds. Data cds are broken up into small blocks, each of which has its own section of error correcting information. Audio cds are just one big stream of data with no markers or error correction along the way. If you were to try this with an audio cd, even the very first copy would be different because the laser would start writing in a slightly different spot. (A little more or less silence at the start). Also errors might appear throughout the cd more often, but good cd drives usually cover them up.

[Mash]

no blank CD or DVD has a physically perfect recording surface. There will be error variation in the physical recording surface of any CD or DVD disc that will cause dropouts. Surface variation can be measured as so many sigmas i.e. a 3.5 sigma process, but it is far more difficult tp predict specifically where a variation will occur on a given disc. This is why CD / DVD players have error correction circuits to fill in missed data bits. Different players will, however, demonstrate different error correction competencies.

So- are you repeating each copy step with the same equipment? An unlikely real-world scenario.

The 100th copy may very well sound different because it contains excessive recorded error corrections made by the upstream equipment, with even greater variation in recorded error corrections if a different recorder/player combination is used each and every time. But who in the real world is going to be using such a 100th copy?

[jack70]

I don't know if anything new has been broached here, I was under the impression that this was all common knowledge by now. There are always going to be a few holdouts I suppose.
More than a few Jim. I see it here, and I see it on other boards ALL the time. Just like the writer who did that test says... we’re coming from an “analog” mindset, where duping tapes (cassette or video) leave a way of thinking that's now wrong in the digital landscape. And since our audio systems (& the world in general) are still 95% analog, it's not an easy thing to get past. Reminds me of the Popes who refused to believe Kepler's & Copernicus's new vision of the universe...

We have similar debates about how burning at slower speeds is “better”... and it’s not... in fact it’s often worse. A little different reasoning, but it comes from the same “analog” mindset.



...and I doubt this will do anything to change their opinion
Gee, if something like this don't get those holdouts thinking differently, what will? LOL. Anyway, that's why I linked the article. Yeah, I know, there actually is a "flat earth society" too. (I prefer cursing to hitting my head against the wall myself).



So- are you repeating each copy step with the same equipment? An unlikely real-world scenario.
The 100th copy may very well sound different because it contains excessive recorded error corrections made by the upstream equipment, with even greater variation in recorded error corrections if a different recorder/player combination is used each and every time. But who in the real world is going to be using such a 100th copy?<br>
....but it is far more difficult tp predict specifically where a variation will occur on a given disc

You're missing the point of his test I think. It shows that digital copying is essentially different than analog. It proves data integrity exists where it doesn't in analog systems. And it shows that disc surface defects (rare because of automated QC) didn't make a difference in this test. If it was a common problem it would have showed up.

Your point about different playback systems has some merit, but that's a different issue. It's one that exists today even with first generation (perfect) playbacks.

[audiobill]

It's funny, but this concern of integrity was nibbling away at the back of my mind, too.

Just curious....did you let your ears do a final listening test, between the music on the original and the 100th disc.??

Thanks-a-ton,
Bill aka audiobill

[ocrazor]

So there IS some error correction in audio cds, just not as much as in cd-roms. Still, the the thing that I don't like is that while a cd-rom drive will not "allow" errors to be read that would alter the data, when playing an audio cd it will simply cover these errors up with silence or interpolation and keep on playing. Audio information could be altered or lost due to scratches and the cd would never skip and you would never even know about it.

[codebutcher]

Effects of interpolation or other means of error correction in audio CDs are nothing comparing to loss of quality due to repeated dubbing of audio/video tapes. And such imperfections that need correction are seldom even in low quality CDs, recorded on mediocre equipment. And CD player has a major role. IMHO, there's nothing to worry about. Actually, this is pretty close to slight paranoia. Just kidding, don't mean offend anyone :)

[jack70]

It's funny, but this concern of integrity was nibbling away at the back of my mind, too.
Just curious....did you let your ears do a final listening test, between the music on the original and the 100th disc.??
Thanks-a-ton
Well, that wasn't me that did that test... I came across it on a techy news site... thought it would be interesting reading to many here because this (digital, technical) subject has periodically come up on this board.

As some others have mentioned above... this test does NOT test, or go into the variables of digital to analog conversion (DAC's, error correction, purity of the original analog wav signal, etc). It's a look into how digital data, when copied, has a great integrity factor. What do you think a 100'th generation video-tape, or cassette tape would sound like... the change in sig/noise, distortion, freq resp, dynamics, etc?

It should be noted here... these copies were NOT decoded BACK into wavs and then digitalized AGAIN and saved back to a new CDR, etc. We're talking simply about the process of RE-copying digital data. It's a small point, but one many don't appreciate (from many things I've seen people say over the years).