Full Version: Optical Media Error Correction
Text deleted by author
to Bob Speer:
A very interesting read, thank you for explaining a couple of things to me.
A very interesting read, thank you for explaining a couple of things to me.
QUOTE(Bob Speer @ Mar 6 2006, 02:29 AM)
Even in the digital realm, there is no such thing as a perfect copy (Bit for Bit transfer). This is truely a subject where there is an abundance of misinformation.
I think you'll want to clarify or qualify that statement, because right now, it's quite widely open to misinterpretation.
If I read it correctly, you're really saying that it's not really possible to get "perfect" copies of digital media. But this doesn't mean anything about perfect copies of the actual data/information on them. Correct?
Text deleted by author.
QUOTE
Because of the "sampling" nature of digital audio, there cannot be a perfect BIt For Bit transfer of data. There will always be errors (missing information), however small, everytime we sample/resample a signal
If copying from Analogue I agree with that statement, going to the digital domain looses information, but copying digital to digital * is * bit perfect (the CD might have write errors (a rareity in my experiance), or degrade over time), but the statement there will always be errors does not ring true (take a store bought CD, burn it and verify the write, if it verifies then the two are bit for bit perfect (there might be various sub-codes missing, etc, but not audio data).
Anyhow, if you must use CD-R or DVD-R then go for lossless (this allows a normal data file system with more error protection than audio cds).
Text deleted by author
We are talking different things 
my comparisions are of test done with audio cds (www.accuraterip.com), many CD-Rs were burnt and sent out to many different people (in the early stages of accuraterip). They all ran ripping tests and the results pretty much 99% of the time came back as perfect rips, without error (that is a bit for bit identical Audio CD rip using a CD-R), there were the odd errors but that was mainly down to older CD drives not liking the CD-R type (old drives, modern dyes).
If it were impossible, the results would not have had matched (even 1 byte of audio data that is wrong is flagged up as a bad rip).
This is all regardless of how many C1 errors, what was compared was data before it went onto CD-R and data that came off the drive (after error correction), not theory rather practice using many people and many drives.
my comparisions are of test done with audio cds (www.accuraterip.com), many CD-Rs were burnt and sent out to many different people (in the early stages of accuraterip). They all ran ripping tests and the results pretty much 99% of the time came back as perfect rips, without error (that is a bit for bit identical Audio CD rip using a CD-R), there were the odd errors but that was mainly down to older CD drives not liking the CD-R type (old drives, modern dyes).If it were impossible, the results would not have had matched (even 1 byte of audio data that is wrong is flagged up as a bad rip).
This is all regardless of how many C1 errors, what was compared was data before it went onto CD-R and data that came off the drive (after error correction), not theory rather practice using many people and many drives.
Text deleted by author
After the third time the same link was posted I had to read that article
Like others I found this part quite disturbing:
but then I got really scared when you posted this :
in any of your articles is really not a good sign.
Then you go on with :
not "estimate" anything. It is a deterministic error correction algorithm
for digital values. Just like in your introduction you mix up error
correction and sampling, two completely different things.
whether it comes from a physical defect or not.
fight against misinformation. You really should get a technical person
to proof read your articles before you publish them.
Like others I found this part quite disturbing:
QUOTE(Bob Speer)
Because of the "sampling" nature of digital audio, there cannot be a perfect
BIt For Bit transfer of data. There will always be errors (missing information),
however small, everytime we sample/resample a signal
BIt For Bit transfer of data. There will always be errors (missing information),
however small, everytime we sample/resample a signal
but then I got really scared when you posted this :
QUOTE(Bob Speer)
When you make a digital copy or transfer, you're resampling a sample. Since we are only dealing with samples in digital audio (not the entire signal) it's likely that additional errors (missing information) will be created during the transfer or copy process. Because of the "sampling" nature of digital audio, there cannot be a perfect BIt For Bit transfer of data. There will always be errors (missing information), however small, everytime we sample/resample a signal.
This is just plain wrong, and the fact that I cannot find "Nyquist" and "Shannon"in any of your articles is really not a good sign.
Then you go on with :
QUOTE(Bob Speer)
Audio written to a disc contains no error correction of it's own.
Of course it does, it's called C1/C2 parity bytes.QUOTE(Bob Speer)
Missing audio information is replaced by an estimated value by the CD player
during playback using CIRC Logic
No, CIRC has nothing to do with replacing missing information and it doesduring playback using CIRC Logic
not "estimate" anything. It is a deterministic error correction algorithm
for digital values. Just like in your introduction you mix up error
correction and sampling, two completely different things.
QUOTE(Bob Speer)
C1 Errors refer to the block error rate (BLER), which consists of bit errors.
No, C1 errors are errors in bytes, just like C2 errors.QUOTE(Bob Speer)
CU errors are either the result of damage to the disc, or C2 errors
that cannot be corrected.
CU errors are by definition C2 errors which cannot be corrected,that cannot be corrected.
whether it comes from a physical defect or not.
QUOTE(Bob Speer)
As I stated in an earlier post, this is one area where there is much
misinformation. We'll simply have to agree to diagree on this one.
Indeed, and unfortunately you don't seem to be on the right side in this misinformation. We'll simply have to agree to diagree on this one.
fight against misinformation. You really should get a technical person
to proof read your articles before you publish them.
Text removed by author.
@bob :
Some people, like spath, choose to contribute. He is also a moderator on the CD Freaks forum, and has (apparently) been a CD drive designer for quite a few years. I would have a slight tendancy to give credit to him instead of you, esp. since he proved me wrong previously (albeit with character). No need to be sarcastic.
There are quite a few resources on the internet explaining the Redbook standard, and concepts that it implies. Any user interested by this type of info can start off with a "Red book" search on google, or on wikipedia
Peace,
T.
Some people, like spath, choose to contribute. He is also a moderator on the CD Freaks forum, and has (apparently) been a CD drive designer for quite a few years. I would have a slight tendancy to give credit to him instead of you, esp. since he proved me wrong previously (albeit with character). No need to be sarcastic.
There are quite a few resources on the internet explaining the Redbook standard, and concepts that it implies. Any user interested by this type of info can start off with a "Red book" search on google, or on wikipedia
Peace,
T.
Yikes, what a bad article.
Mr. Speer, go buy yourself a copy of Nika Aldrich's book on digital audio for audio engineers, post-haste.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.