===================================================
= NB: Over 50% of this message is QUOTED, please =
= be more selective when quoting text =
===================================================
Why is that to soft? The soundquality of CD's which are carefully recroded
and from which the levels never touch or exceed 0dB is usually much greater
then from CD's which are recroded to and beyond 0dB, thus introducing a fair
amount of clipping.
When I record to MD from analog sources I never record past -4dB and I try
to achieve a average recording level of about -12dB.
Remco
----- Oorspronkelijk bericht -----
Van: David W. Tamkin <[EMAIL PROTECTED]>
Verzonden: zaterdag 7 augustus 1999 05:49
Onderwerp: MD: calculating digital gain for too-soft CDs
> You know, I posted this to alt.audio.minidisc and to the MiniDiscussion
> boards, but not yet here.
>
> A lot of tracks on my CDs were mastered way too softly and peak in the -6
dB
> to -3 dB range. I've an application that unfortunately involves copying
to
> audiocassette, so I need to set the levels on the MD to be as strong as
pos-
> sible without clipping.
>
> My CD changer offers peak search, but the function is not of much use.
> Someone explained to me that it is intended for analog copying, so it
tries
> to locate the passage that averages out the loudest over several seconds,
not
> necessarily the single loudest sample. Multiple runs of peak search over
the
> same track invariably result in its selecting different passages.
>
> Further complicating the matter is that the VU-meters on the MD recorders,
> as Grega Simenc has explained to me, show only about 100 or so samples per
> second, not all 44,100, so a given input at a given gain might like the
> "OVER" indicator on some plays but not on others, depending on whether the
> clipped sample or samples were shown on the meter or not.
>
> I've been resorting to the following procedure, which is so wearisome that
I
> haven't recorded anything from CD to MD in weeks:
>
> 1. I set the MD to monitor the digital signal from the CD changer.
>
> 2. If at any time in the following steps I lower the gain to +0.9 dB or
> lower, I just set it to 0.0 and stop testing. After all, the correct
> setting can't be negative in this situation.
>
> 3. I take a peak search and set the gain high enough for the passage to
clip.
> While the CD changer repeats it, I lower the gain until it can play
three
> times without lighting "OVER".
>
> 4. I take another peak search and, if at the gain level determined in step
3
> it lights "OVER" on one try out of three, I lower the gain until it
> doesn't.
>
> 5. I take a third peak search and do it again.
>
> 6. I play the track, watching for lighting of "OVER". Whenever it lights,
> I pause the CD, lower the gain, rewind the CD player before the passage
> that would have clipped, and resume play.
>
> 7. When I get all the way through the track, if I haven't broken out
because
> of step 2, I lower the gain another 0.2 dB just in case there were
samples
> that didn't make the VU-meter but which would have clipped.
>
> It is very wearisome!
>
> Now, I understand that if I had a working soundcard there are recording
util-
> ities that can scan a .wav file or such, spot the highest amplitude, and
re-
> port its value. But I don't have a working soundcard nor any other reason
> to shell out for one, and my audio equipment is in another part of the
house
> from my computer, so computer-dependent solutions are not practicable for
me.
>
> When I asked on alt.audio.minidisc, Chris Knoche said something that made
me
> think of using the Peak Hold feature of my JE520 instead of the CD's peak
> search facility to get the initial setting for step 6. But its meter is
so
> coarse that I don't know if it would help much.
>
> When I asked on the MiniDiscussion boards, T. Huu and I started a
correspon-
> dence where T. (don't know which pronoun to use) suggested a voltage meter
> with digital display and peak hold of its own. But I thought, first, I'd
> need to get separate readings for the left and right channels, and they'd
> need to be taken from the analog output, and second, I thought it would
take
> quite a few decimal places to get enough resolution to figure the gain
within
> a couple tenths of a decibel as I'd like, but the way the math works out
when
> I try it it takes about a 2.3% drop in voltage to correspond to a 0.1 dB
drop
> in amplitude, and that left me wondering whether I was calculating it
right.
>
> So, is there any resolution to this problem? Is there anything that
improves
> on my current method? A voltmeter with digital display and peak hold
would
> be rather expensive (not to mention dependent on the CD changer's DAC),
> wouldn't it?
>
-----------------------------------------------------------------
To stop getting this list send a message containing just the word
"unsubscribe" to [EMAIL PROTECTED]
