"Mike Rivers" wrote in message
news:znr1057584387k@trad
In article
writes:

Thanks for the numbers. So, it's not a common problem. That is,
it's much more common for an analog device to be capable of putting
out a clean signal that will clip a digital recorder's input, than
the opposite (i.e. that its maximum undistorted output will -not-
reach 0 dBFS, and thus it's unable to use all the bits (which isn't
really a "problem" when tracking, at least where I sit)).

Actually it IS a common problem, I'm just not sure how common.

OK, so its questionable exactly how common it is because we don't know how
common it is. BTW, is that glass half empy or half full? ;-)

Arny has set up far more different sound cards than I have. Personally I
wouldn't calibrate a digital recorder at -12 dBFS for the nominal
input level. That's not enough "digital headroom" for me.

Agreed.

However it
probably works pretty well for modern music where you're recording
samples that are already compressed, or you put fairly heavy
compression between the live source and the recorder.

Is that really all that common?

I'd much rather
have 20 dB of headroom built in and still stay on the conservative
side.

Agreed.

People have problems in both directions. There's the preamp with a
nominal output level of +4 dBu and maximum output of +28 dBu "easily
overloading" a sound card input that's set up for a nominal input of
-10 dBV and 12 dB of headroom.

An obvious mismatch that requires mixing equipment rated at -10 with
equipment rated at +4. Many sound cards have switchable input sensitivity so
this is a misunderstanding, not a permanent problem.

And then there's the preamp with a
nominal -10 dBV output going into a sound card that has a nominal
+4 dBu for -20 dBFS which makes recordings that aren't "hot enough."

Frankly, I know of no sound cards with 20 dB worth of headroom. 12 dB seems
to be a VERY typical number.

Then it's further complicated when connecting an electronically
balanced output to an unbalanced input in a way that only gives half
the nominal output voltage. This drops the record level by 6 dB, which
on most waveform displays, is half the waveform area, which worries
people a whole lot. It also means that there may be more gain required
on playback, but I've rarely found this to be a technical problem,
only one of perception.

I agree with you that this is a problem of perception. You've scoped it out
perfectly - many people are too sensitive to appearances without considering
what those appearances actually mean.

People are afraid of boosting gain for fear of
increasing noise. On some equipment it's a problem but that should
tell you to get better equipment, not to agonize over recording
levels.

IME far more recordings have been damaged by audible clipping than audible
noise floors. These days even relatively cheap consumer cards have noise
floors that are 80 dB or more down.