The correction in a CD is NOT in the DAC. The correction is in the encoding of the CD data - IF it plays back it is lossless! This is Reed-Solomon encoding and the way it works is that the data is NOT serially encoded, it is interleaved with a type of checksum that makes it possible to recreate a certain amount of missing data. Beyond that amount, you get an audible glitch.
So, the "correction" takes place in the uproccessor chip that *reads* the data from the CD itself.
No DACs that I am aware of do ANY error correction at all. It is possible that the Digital filters (they do upsampling) *may* have some sort of provison for making sure the bits that get in, get out, but they don't do error correction afaik.
A digital to digital copy, done with EAC or a similar program, should be identical to the original.Pro gear or not, if the data is identical, it is identical.
The question of how it sounds will depend upon some details that include jitter on *playback* - and that will in turn depend upon the details inside the particular unit chosen for the playback chore.
It is possible that on the mastering end that some details that are related to the clocks used in the transfer *can* have an effect on the final product. This was a big problem early on in the CD duplication industry, but presumably has been identified and eliminated.
Having said all that - in practice all bets are off. Ymmv.
- |
I agree with this point of view. So the copy may sound better if it is played on a machine or DAC which has a worse error correction capability than the machine on which the copy was made. For example playback on a portable CD, copy made on HHB burner.
But the copy can never sound better than the original when both are played on the machine which makes the copies ... you can't doubly correct the errors : the copy will not benefit from the error correction when it is played on the burner since its errors are already corrected.
So I think you'd do better to buy the best possible CD playback machine, rather than a cheaper playback and a copy machine.
And the high error rate on the original is caused by poor mastering, not by jitter inherent in the disk. Jitter is a function of the transport, not the disk itself. |
Okay, I spoke to a recording engineer that I know. He's forgotten more about this subject than I'll probably ever know, but I will relay what he told me. In the recording process errors occur in producing the glass master from which the final CD is made. The reason we know this is where the errors occur is if you take 3 CDs of the same title and production you will get the same error rates (or extremely close). If you take a mass produced CD and a high quality re-master you are very likely to get a very low error rate on the re-master as opposed to the mass marketed one. He says this is normal. It is caused by clocking errors in producing the glass master. Some production houses are better than others. All DACs make an attempt to correct errors, whether they are caused by jitter or by error in the recording process. Some DACs do a better job than others. Some buffer the information and read the bits on both sides of the error and average them to correct the error. Some simply take the adjacent bit and fill in the error. The process of re-recording the CD can improve the error rate. It will fill the error bits with whatever protocol it uses. The result will be a disc that has a lower error rate. This still leaves a question of is the original better than the copy. The copy now has a lower error rate, but it is just processed the errors and "pre-corrected" them prior to burning the copy. If the protocol for correcting the error in the computer prior to burning the CD is a better protocol than the DAC being used, the result should be a better sounding copy than the original. For example if the computer corrects by bi-interpolation and the DAC uses only adjacent bit correction, then the copy will be better in THAT DAC. In another DAC that uses a superior error correction scheme this might not be the case. Well, I think that answers it. I hope I was clear in translating a recording's engineer's wisdom on the subject. |
Your right that analog tape won't have jitter issues, but I'm not sure I understand how that relates to this thread. I really would like to understand if these error rates can not be corrected through reclocking and re-recording why do I (and apparently a few others) have the experience of copies actually measuring better from an error (and/or jitter) rate. I am not a recording engineer, but will pose this question to the recording engineers I know. In the meantime, if anyone knows the answer to this (and it could be that my method of measuring is completely invalid for some reason--and you really can't correct the problems on the original disc) I would like to know the answer. |
Rives, I've got an intriguing question here:
What if you simply transfer your CD into simple analogue cassette than you won't realy have any jitter issues will you? |
Unsound: That's the same experience I've had. I can make copies of some pretty poorly mass produced CDs that sound much better once I've made the copy. But take a JVC XRCD--I can't make it any better it only gets worse. I used a Genesis digital lens to read the error rate (at least I think that's what it measures). I would play the original, then make a copy and replay it. For those poor quality original CDs the copy had a lower error rate as measured by the digital lens. Now, the ones that were really good to begin with never improved--usually got worse. I have heard that people have used the Genesis digital lens in front of a CD recorder and gotten similar results. I have not done this as my only CD recorder is the one attached to my computer. Perhaps someone else has tried this experiment. So any ideas to what's going on here? |
Way back when Stereophile ran one of the their first shows at the Waldorf in N.Y. someone by the name of Thomas W. Shea with someone from Audio Alchemy were offering free Cd copies of any Cd you wanted. These copies were always better than the original. It was a case of diminishing returns. The worse the original the better the copy. The better the original the less improvement in the copy. |
Rives .. reclocking is certainly helpful on playback if the transport does not have a very good clock to begin with. I think jitter is the death of CD sound quality, but I also suspect that manufacturers of cheap CD players use jitter as a kind of "dither" to mask shortcomings elsewhere in their signal path .. I'm certain Marantz does this in their low end players. Jitter adds a warmth and mush which helps out a cheap analog output stage.
As for the error rate there is no way to improve on the error rate you started with since if the bit on the original disk is in error then the information which that bit carried is lost forever.
If this bit error is correctable by the error correction coding then it would be just as correctable on the original disk as on the new disk.
Perhaps you can produce a copied disk which will play better on a marginal transport due to having better reflectivity, but in order to produce this better copy you would have had to play it on a better transport in order to read it.
I maintain that a digital copy can only be the same or worse ... it cannot be better, because it cannot retrieve information lost on the original disk.
I'm not trying to pick an argument, and I am really interested if someone can explain why I'm wrong ... because I have been wrong about audio many times in the past (e.g. digital cables can't sound different ... now I realise they can, and that their are sound explainations as to why). |
Jacks: That's helpful and clears up some of the confusion. I had confused jitter and error rate. I had been told by someone in the industry of CD mastering and manufacturing that there are clocking errors on the record end that can be reduced by extracting the data and reclocking it. I equated this error to jitter as jitter is a clocking error, but as you have pointed out it is an error rate and not really jitter at all. At any rate, it does seem the case that the error rate can be reduced by making a copy that re-clocks the data, and does so in a superior method to the original. At least this is what I have been told by those in the CD industry. |
I've had good results, re-recording CD's on my Alesis Masterlink. There is a significant reduction in error rate. I've heard that there is a new CD burner out that burns more precise bits in the CD, which further reduces the error rate. I believe that Yamaha builds it.
Coming from a "pro" background, I believe that many confuse jitter with error rate. Some audiophile CD manufacturers specify a certain reject percentage for error rate. This said, some CD's that are manufactured end up in the incinerator because they don't meet the specifications. FIM Music is one such companies that take this approach, though most of their catalog (musically) suffers ... |
Rives, It's rather playback system has level of jitter X but not the original CD for christ sake.
The recorded CD will than have 1.1X level of jitter played in the same system.
Yes the future high end CD players might have a memory buffer that will store info first and than play. That will take much longer waiting time for the first track to play but Yes the jitter level will be significantly reduced let's say upto 0.05X.
Mara |
Are you saying that uneven spacing of the pits on the CD contributes to jitter during playback ?
I don't think I can believe this because data is retrieved from the CD at a different rate from the rate at which it is fed to the DAC (this must be the case since the data on the disc also contains error correction redundancies). The clock rate for data retrieval is not the same as the DAC clock rate.
Someone who knows how CD playback really works please straighten this out ... does the inherent jitter in the pit spacing on the CD (which I guess you'd see as a closing eye pattern on the opto detector) translate into greater jitter during playback ? I just can't believe it, but I concede I could be wrong. |
In digital, unlike analog, re-recording can actually make it better (not just sound better, but really technically better). Jitter occurs at both the record end and the playback end. If you can eliminate or reduce the jitter at the recording phase, then the result is a better disc that has less jitter on the playback. Now, how does this work? First assume the original disc has jitter of X. Your playback system has jitter reduction to some degree and will reduce the jitter to only 50%. Now you load the system onto a hard disc and then into RAM where it is heavily buffered, the process involves reclocking the data stream as it is written to disc. Let's say you can reduce the jitter by 90%. So now you only have 0.1X as your jitter being recorded. Now you playback and because the jitter is low you only reduce the remaining jitter by 20%. The result is playback that was original 50% of the original jitter vs the copy that is now only 8% of that original jitter. These are, of course, hypothetical numbers, but the principle is sound (no pun intended). This is one of the few areas that copies can actually be (not just sound) better than originals. |
Could anyone give a technical explanation of how the sound is supposed to improve ... after all its bits, 1s and 0s on the disk, and the timing (e.g. jitter) is not derived from the disk itself, but from a reference oscillator inside the CD player.
I'm not being a pain-in-the-butt skeptic type .. just really interested how the sound could possibly improve. I can only see room for degradation. |
If you are going from a digital source to digital source and use any recorder that has jitter reduction you can often improve the sound. Professional equipment usually has very good clocking and you can often improve the sound. PCs can too improve the sound, but as Rec points out, if you are using an analog source you need to have a good sound card or your copy will sound considerably worse than the original. |
you should admit that re-recording will not be better than original in any case.
you can make it sound better if you will re-master from original tape but there you should face the copyright issues. |
I have made recordings from a Pioneer cd recorder, Fostex pro cd recorder and from a pc. The Pioneer and the Fostex copies sound the same as the source, the pc recordings to me sound less dynamic and more compressed. This may be due to my cd recorder and computer soundcard. I notice that if I copy at four times speed on the pc the sound is is more compressed and less musical. This may be system dependent as I have much more money in the Fostex and the Pioneer recorders. I also use a Fostex DAT and it makes excellent recordings. Another positive with pro equipment there is no copyright issues and you can make as many digital copies as you please and you can use computer blank cdr(less expensive than cdr for a home audio recorder) |
i'm not sure about professional cd recordings but if you get dat and make a recording from vinyl you'll be indeed more successfull and have a recording much better than on regular red-book cd. |
