Aaron Ximm wrote:
> I'd be curious if this is an artifact of the sample rate per se, or the
> conversion process -- my money's on the latter, but my mind is open.
I'm of the same bet, it's probably conversion.
> I had two or three ideas about the differences I'm hearing:
> (a) the Sony is doing some un-specified post-processing to handle cases
> where it must generate a lot of 'filler' white noise, to insure that it's
> either masked by higher frequency noise, or as non-periodic as possible;
It needs to generate no filler. I've not seen any sonograms of any ATRAC
recording where deliberate filler was introduced. It does not do this,
it only works to reproduce what it got not invent new stuff.
> (b) EDL is not implementing the ATRAC decode recipe correctly;
Or just differently, note that the difference could also be hardware,
even a clock speed mismatch or such like.
> (c) there are ambiguities in the recipe when the compression is highly
> taxed and I subjectively like Sony's handling better;
The compression is not even getting into heavy breathing on this stuff.
> (d) there is some problem in my SPDIF transfer (eg jitter) that happens
> to produce subjetively appealing distortion in these cases.
Certainly when moving digital through several pieces of
hardware/software the potential does exist for changes. Differences in
the precise timing of the crystal oscillators controlling different
parts, for instance. Differences in exactly what the software really
does, which is usually different from what we think it does, and even
different from the official description of what it does. The trickery in
hardware and software is very complex to get what's basically a simple
adding machine to do complex stuff.
> Any opinions or analysis most welcome -- one thing I *haven't* done (and
> haven't the tools/expertise to do well) is do a spectral analysis of thes=
e
> clips. I can hear differences -- but only characterize them crudely.
Ok, here's a sonogram:
http://wwknapp.home.mindspring.com/images/Flagsono.jpg
This is a composite of all three samples. First part is from the mdte
file, second from the sony, third from the difference. Note how hard it
is to pick out the transition between the mdte and sony part, there is
almost no difference between the two. And what there is appears to be a
slight intensity difference rather than structural. Look at how closely
the difference part mimics the louder parts of the others.
Note:
This is very far from white noise, it's fairly low frequencies only.
It's easy to be fooled into thinking you are listening to something
complex when it's actually far less complex. This is not all that complex.
The differences are frequency dependent, and nearly all below 300hz.
The differences between ATRAC material and CD are almost entirely in the
frequency range above 16khz, way above this stuff. It's always been able
to do quite well with these low frequencies, even if complex. Lots of
samples per cycle makes it easier for the whole process.
> Remember -- both these signals were transfered and decoded entirely in th=
e
> digital domain... :/
The difference I hear is that the sony is slightly more detailed than
the mdte. Why I have no clue. The sony was designed for audio, the other
system was not. It could be anywhere.
Note that all the various proposed errors in compression and what it's
not supposed to get right apply to the encoding. If it "loses" anything,
it does it encoding, the decoder is entirely passive, no loss occurs
there. Once on the disk it's pretty much instructions to a sound
synthesizer for the decode. It is known that slight differences do exist
between the output of the decoders for various versions. Slight
improvements in the decoders have occurred, though the specs for how
they operate have been fixed from the beginning. In general the more
recent the decoder, the better the sound by a extremely slight amount.
It is however, far more important that the encoder be newer.
Walt
________________________________________________________________________
________________________________________________________________________
|
