Jeremy Martin wrote:
>
> At 03:06 AM 9/16/2002, you wrote:
> >One issue in the post below is the statement that the recordings will
> >be subject to various kinds of processsing (paragraph 3). None of
> >the subsequent posts supposedly on this topic have considered this.
>
> Walt did mention this a little - since I will be using noise reduction
> software I'm definitely going with an older Sony model where the recording
> level stays fixed the whole time you're recording. I will have to set my
> levels up better before hand but that will help wonders with removing any
> steady background noise (ambient or from the mic/pre).
Um, I think you have got it slightly off. All Sony MD's, even the new
ones will record with a fixed gain. If left entirely to defaults they
come up with the auto level set, you have to set the recorder to manual,
on my old MZ-R30's that was as simple as holding the pause and record
buttons for a few seconds and it would indicate manual in the display
and you could set the level you wanted. You could also get at it through
some menu choices. On all Sony's you would then press the pause button
to start it recording once you had the gain where you wanted, they do
generally remember the gain setting, at least mine did. On the newest
Sony's, you can operate the gain controls while it's recording as well
as when it's in record pause. But they do everything else too.
As I was saying, riding gain during recording does have problems when
you get into processing. That's not just a MD matter, but true of any recorder.
> >A number of commentators elsewhere have suggested that while the
> >immediate user of a MiniDisc recording might be entirely satified
> >with it, once one goes beyond that immediate use problems may arise.
> >To test this contention, we have a lot of work to do. One thing
> >ought to be clear though - ATRAC was not designed for anything other
> >than immediate use of a recording (or a clone). If you want to
> >process signals, it may be wise to do so with non-compressed source
> >recordings. Or, as Jeremy is possibly implying, maybe one should be
> >happy to tolerate any subsequent non-linearities.
You should be aware that this poster is speculating. Once we finally got
the detractors to mostly admit they could hear no difference they raised
this issue. It's a non-issue, at least as far as ATRAC goes. The issues
with additional processing problems are virtually all sync problems, and
are in common with any digital record method.
In many cases all types of compression are thrown into the pot as if
they were identical in what they did. Some types of compression can
interact with processing (which is often done by tossing out data, ie
the processing may be compressing too). ATRAC most closely resembles
MIDI in how it works in this regard. Both use the stored data to drive a
synthesizer which recreates the sound. We don't see too much complaint
about MIDI, which is severe compression.
I've asked for actual examples of these problems with current versions
of ATRAC itself. So far I've not seen any put up anywhere.
MD recordings are being used to mix and master CD's now. So it must be
capable of being done. The one that surprised me this year was how many
mic reviewers were using a HHb Portadisc to record from the mic they
were reviewing.
While mp3 is decidedly designed for immediate use and deadend, I don't
believe Sony views MD the same way, and they are the ones that designed
it. It is true you don't want to use it over and over on a piece, and
that's a common practice with DAT in mastering. You want to use it as a
record medium, then transfer it and handle it uncompressed from there.
> As I was planning to use my recordings in fairly abstract ways a lot of the
> time, the compression wouldn't bother me that much anyway. In fact I may
> often distort things more on purpose. It's odd, but for some reason low-fi
> sounds with lots of distortion actually sound good in some things,
> especially percussion. With today's software samplers, plugins, not to
> mention Reaktor/MAX MSP/etc, anything's possible, and I really doubt ATRAC
> is going to make it sound all that worse.
You may be disappointed if you are counting on some damage from ATRAC to
make your sound. Expect to deal with recordings that, as far as the
ATRAC part, are better than CD. Of course it's easily possible to mess
the recording up somewhere else.
> But for those more concerned about the differences (for scientific reasons,
> or whatever) here's a test you can do. Record something in over DAT. Make
> sure that same sound is transfered digitally to a PC, then digitally to a
> MD (so the only A/D conversion will be on the DAT's first recording - the
> MD will just be encoding the DAT's recording so the MD's converters won't
> come into play, right?). Now digitally send the MD's encoded version back
> to the PC. (So a high end MD recorder like the one Walt uses would be
> necessary).
>
> Note: someone who has a DAT recorder, could record the first sample, then
> send it to Walt or whoever. Then Walt can do his digital MD thang and send
> the file back for comparison. No one person has to have both a DAT recorder
> and a high-end MD recorder, in other words.
>
> Ok, then the comparison: open up something like Emagic's Logic software.
> Something where you can play two .wavs (or .aiffs or whatever) at the same
> time. Phase invert one of the two waves. Make sure they are lined up
> *exactly* of course and are the exact same length. If the two files were
> exactly the same, once one was phase inverted there would be total silence.
> Since the MD did encode it using ATRAC, what will remain is exactly
> whatever effects ATRAC has on the sound. It might not sound like much but
> we could zoom in down to the sample level and see exactly how big of a
> difference it makes.
>
> If you guys get the 2 files together I can do the phase inverting etc on
> this end, unless you already have software that will do that. Cool Edit
> probably does it even.
Way ahead of you, nearly two years ago now I did this in a slightly
different form.
Let's take as a given that CD will be the same as DAT, even though DAT
recorders have their own problems. This thread is not about DAT. Now
let's write some tracks to a audio CD, from CD sources (probably DAT
originals, but it does not really matter). Be sure and put sync points
in the beginning of each track. Now lets put this CD into my Sony MXD-D3
deck. Stick a blank MD into the MD side and press syncro record. This
will make a "clone" of the CD onto the MD, all internal to the deck, all
digital transfer, but the MD is now in ATRAC code. Now transfer those
tracks back to the computer from the MD. As this is a analog transfer
from this deck, we will also transfer the tracks from the CD back to the
computer via the same path.
We can then line up the sync marks, inverting the MD and subtracting the
MD from the CD track. You then have a track that's what ATRAC removed,
the only difference between the two as the deck uses the same D/A for
both disks. Or at least as close as I've managed to get so far. Note
this deck is using ATRAC 4.0, the same version as is in the older Sony
MZ-R30 to R55's.
Trot off to:
http://www.naturesongs.com/walt/difference.html
There you will find all the binaries, soundfiles and sonograms. You will
also find the blind listening test we did, though the key is there, so
it's no longer blind. And you will find a set of comparison samples with
the HHb Portadisc, with the transfers all digital.
There is no difference test for the all digital with the Portadisc
because you run into a clock difference between my G4 mac and the
Portadisc that leaves you looking at the beat frequency between the two
clocks on the difference test. 2.6 samples per second difference, but
it's enough to mess it up. I'm pretty sure the G4 is the one with the
slightly slow clock, it loses time at about that rate. There may be a
way around this, but I've not spent the time to sort it out. The
difference test is far too sensitive for any editing of the tracks in
the G4, what I need is a way to send the sound out through the loop at
the same time I record it back, without going through ATRAC. The
Portadisc can do it, but the computer does not seem to be able to do that.
The Portadisc is ATRAC 4.5, btw.
Walt
________________________________________________________________________
________________________________________________________________________
|