In this case "cleaning up sound" is not the goal. It's making it as
undeststandible as possible. As such, my adivce is not to be concerned
with whatever noise or listening fatigue are brought out in whatever
proccesses are applied. I've found that compression and very narrow band
peaking enable such decipherings. The band peaking's in the 3 - 5K range in
which it's the overtones, rather than fundamentals that come through
When digitizing the originals, if at all possible do the compression BEFORE
the A/D. Otherwise the very low level voices, when brought up will suffer
from very low bit resolution that'll cause unnecessary noise.
If you have access to any good old analog outboard noise reduction gear
(i.e. Dolby or dbx), then their recording outputs will be compressed with
pre-emphasized highs.
That's a good way to pretreat the signal BEFORE digitizing it. An
adjustable compressor (UREI, Orban, Manley would be better suited, but you
can work wonders with the noise reduction encoders.
Best,
Shiffy
|