[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ARSCLIST] Lossy compression losing quality (was Re: [ARSCLIST] Pristine Audio and the Milllennials . . .)
Howard Friedman wrote:
Mike, you wrote,
3. Copying a digital file either loses nothing or fails; the exceptions
to that are exceedingly rare. That is, if the copy is not perfect then
the transfer will fail, usually due to the dreaded CRC error. There is a
tiny chance of dual failure which would permit copying a file with an
error, but that is negligible in practice.
Does that mean that ripping CDA format files from a CD to MP3 format on my hard drive loses nothing?
No - because ripping audio is not copying. Because a CD-DA is recorded
with one less layer of error correction, the probability of a bit error
is substantially higher than when copying a file. For better or for
worse, playback takes advantage of the ear's insensitivity to very brief
transients and just masks out the error - error concealment is the
technical term.
Now, the right equipment and good media used near optimum speed will
give very reliable extraction, especially with Exact Audio Copy to do
the ripping. But the 13% of extra audio playback comes at the cost of
far higher risk of accepting a bad bit here and there.
A case in point arose during some experiments in steganography - hiding
encrypted data within some masking information. I ran a test in which I
buried text in a nearly full 74-minute T-Y disc in CD-DA. Using a 12x
Plexwriter at 8x, I recovered the full text without error. (The source
material was a live Qawwali performance; there was no way to tell that
anything but music was on the track.)
The probability of error during extraction in the best case can be made
very small, but still far higher than the probability with the same
audio as a file - WAV, AIFF, APE, FLAC or other lossless format - using
the same setup.
Mike
--
mrichter@xxxxxxx
http://www.mrichter.com/