Post by KeithL on Aug 15, 2019 12:01:08 GMT -5
Much as I hate to open an old can of worms that is well past its use-by date......
This article is both quite informative, quite accurate, and extremely misleading.
(The article applies completely to the vast majority of hardware or "appliance" audio CD players.)
As the article describes, CDs have very powerful error correction, which can correct most minor digital errors perfectly.
And most CD players also include a final level of interpolation - which does its best to fill in gaps that are simply too extreme to be corrected perfectly.
And, as they mentioned, most CD players make no effort whatsoever to inform you when this happens... so you really don't know whether you're getting all the right bits or not.
HOWEVER, almost always, the exact opposite is true for computer-based playback and ripping of CDs.
First off, computers treat audio CDs like data, which means that they DO NOT interpolate.
So, if there's a spot on the CD where the data is too badly damaged to be repaired by the error correction, you WILL actually get an error... and the disc will stop playing.
Note that this process operates at the physical level - error correction operates at the level of ensuring that what you read from the disc is what was originally recorded there.
However, this sort of error correction won't know if, for example, there were errors in the data before it was written to the disc.
So, if I were to write a CD-R disc, from a damaged file on my hard disc, the error correction in the player will ensure that file, including the damage, is reproduced perfectly - but will not report or correct it.)
HOWEVER, most modern disc ripping software takes this a big step further....
Programs like dBPowerAmp, after they rip each track, calculate a checksum, and compare that checksum to one they find in an online database.
This confirms that the track you ripped is IDENTICAL to copies of that same track that other folks ripped from their copies of that same CD.
Again, no interpolation, no guessing, and no errors allowed....
And it DOES tell you if they are not.
So, if you rip that same CD on two different computers, both of which use this technology...
Then, yes, they will be absolutely identical, and you will know it, and there is no possibility that something will have been interpolated.
And, as a sort of side effect, doing this gives you a good idea about how many CDs are out there that actually are damaged and have irreparable errors on them.
(And this will depend to a major degree on how carefully you handle your CDs, how clean they are, and the quality of the disc transport in the computer.)
I personally am somewhat obsessive about how I handle my CDs.
I don't touch the playing surface, and I don't touch them with dirty hands, and I most certainly don't leave them laying around without their sleeves or jewel cases, or use them as coasters.
Of the last 500 or so CDs I've ripped.... the software detected a total of THREE uncorrected errors (we're talking about a single error on a single track - equivalent to a single tick on a record).
Of those....
- one was fixed by carefully cleaning the disc
- one was a scratch on the disc which was simply too bad to be fixed (that disc had been tossed into an envelope without even a paper sleeve - not by me)
- and one turned out to be an error on the pressing master disc itself (I bought another copy of that same disc and it had the same error)
I should also point out, in case it isn't obvious, that the CD transports found in most modern computers are similar in quality to the transports used in most low cost audio CD players.
So, the idea that these sort of uncorrectable errors, that end up being imperfectly interpolated, and causing an audible loss of sound quality, are common... is a HUGE exaggeration.
(You might as well worry about the optional volcano and shark bite insurance.)
As someone else noted - there is one area in which a transport that is otherwise functioning properly may vary - which is jitter.
Jitter is a collective term for timing errors that occur in the clock that is used to synchronize the audio data stream.
It is also true that some transports may output a data stream with more or less jitter, or different sorts of jitter, than others.
But whether this makes a significant difference or none at all will depend largely on how your DAC or other down-stream components handle jitter.
(Some older DACs, and some oddly designed modern ones, are quite sensitive to jitter - but most modern equipment is totally or almost totally immune to it.)
