rec.arts.movies.production.sound

wrote ...
> Richard, here is a link to a 15 second clip of the scene.
>
> I'd be very interested to see how you approached it.

/media/

I tried two different types of processing..

1st phrase: I used the "Effects / Noise Reduction / Noise Reduction"
function. I took a sample of the noise from the end of your example
(but not including the "clink" sound!). And then I used it at 90%
strenth to "reduce" the noise. The result was a bit "metalic sounding"
(which is to be expected). But IMHO, acceptable as the background
with a translation over the top.

2nd phrase: I used the "Effects / Amplitude / Dynamics Processing"
and tuned the threshold to eliminate most of the noise except
during the speech. I used soft splines to avoid a really harsh
switching profile.

1st & 2nd phrase: I added some pink noise to mask the hard
silence between phrases. This makes it less obvious that some
hard processing was done.

Remainder of sample cut: No processing (for comparison).

I just did a couple of quick-n-dirty processing tests to your
sample. The same kind of attention I would give it if it were
a small part of a larger job. Note that if this were the primary
track (rather than the background under a translator), this kind
of processing might not be good enough depending on the
expectations of the production.

I used Adobe Audition which is essentially identical to
Syntrillium Cool Edit Pro. The same features are available in
more recent versions of Adobe Audition (and in Adobe
Premiere also, at least most of them).

For my taste I would use the first ("Noise Reduction") method
as it tends to also eliminate some of the music even during
the speech (but at the cost of a more "metalic" sound). This is
not what I originally suggested because each case is different
and you never know how well that "active noise reduction" will
work without actually trying it. Sometimes it just makes things
worse.