I am in the process of editing a two-camera video of a semi-acoustic band
using Adobe Premiere 1.5, with audio separately recorded onto DAT.

Mindful that television audio needs to be on some sort of constant plane of
volume, my problem is that the DAT recording contains some very major highs
and lows. If I put the audio on "as is" all the low-level parts (mainly
where the singer is speaking to the audience) would be pretty much
inaudible.

What I would ideally like to do is boost the low level signal without even
touching the higher volume stuff. I've tried compression but I'm finding it
very difficult to find a point where I can clearly hear the spoken parts but
which doesn't distort the louder more dynamic music (bass and drums
particularly) very badly. To hear the speech at an acceptable level I would
need to compress it (I'm using Sound Forge Wav Hammer) with a threshold
somewhere between -30 and -35 dB and ratio at 3:1 to 4:1.

What I want, I guess, is the opposite of compression... where instead of
suppressing the high level I'd be bringing the lower level stuff up to meet
it. Or maybe it's the same thing and I'm missing something! I'm seriously
considering leaving the music as it is and applying compression to the
spoken parts only, but I'm worried about sudden leaps in gain at these
points.

Can anyone assist me, please?

"P" wrote in message ...
I am in the process of editing a two-camera video of a semi-acoustic band
using Adobe Premiere 1.5, with audio separately recorded onto DAT.

Mindful that television audio needs to be on some sort of constant plane
of
volume, my problem is that the DAT recording contains some very major
highs
and lows. If I put the audio on "as is" all the low-level parts (mainly
where the singer is speaking to the audience) would be pretty much
inaudible.

What I would ideally like to do is boost the low level signal without even
touching the higher volume stuff. I've tried compression but I'm finding
it
very difficult to find a point where I can clearly hear the spoken parts
but
which doesn't distort the louder more dynamic music (bass and drums
particularly) very badly. To hear the speech at an acceptable level I
would
need to compress it (I'm using Sound Forge Wav Hammer) with a threshold
somewhere between -30 and -35 dB and ratio at 3:1 to 4:1.

What I want, I guess, is the opposite of compression... where instead of
suppressing the high level I'd be bringing the lower level stuff up to
meet
it. Or maybe it's the same thing and I'm missing something! I'm
seriously
considering leaving the music as it is and applying compression to the
spoken parts only, but I'm worried about sudden leaps in gain at these
points.

Can anyone assist me, please?


feeding the audio through an Aphex Compellor or some device with similar ACG
function may be your best bet.

the Behringer Ultramizer may be a good alternative.

if the end use is for broadcast you may not need to do anything at all. the
processors at the transmitter will squish the audio just fine

feeding the audio through an Aphex Compellor or some device with similar
ACG
function may be your best bet.

the Behringer Ultramizer may be a good alternative.

Thanks for the advice, however it's a not-for-sale venture and I won't be
forking out any more on it than I already have, other than my time.

if the end use is for broadcast you may not need to do anything at all.
the
processors at the transmitter will squish the audio just fine

It's definitely not for broadcast - permission granted only to video and
distribute to people who couldn't be there, or those who were and want to
remember. The effect you're talking about would be ideal - a dynamic
compression that adjusts according to input. I can't see anything within
Sound Forge that would do that.

"P" wrote in message


I am in the process of editing a two-camera video of a
semi-acoustic band using Adobe Premiere 1.5, with audio
separately recorded onto DAT.
Mindful that television audio needs to be on some sort of
constant plane of volume, my problem is that the DAT
recording contains some very major highs and lows. If I
put the audio on "as is" all the low-level parts (mainly
where the singer is speaking to the audience) would be
pretty much inaudible.

What I would ideally like to do is boost the low level
signal without even touching the higher volume stuff.

The best way to do this is to manually identify, select and adjust the
volume of the various logical segments of the recording.

What may not be obvious is that the subjective loudness of audio is not
directly related to visible amplitude on the screen. You'll have to work out
the necessary adustments by ear.

This is easy enough to do in Audition, but over the span of a long recording
it can involve some work. I never promised you a rose garden! ;-)

More specifically, my tools of choice are Effects, Amplification and
Effects, Envelope.

One helpful tool is a collection of envelopes that start and end with unity
(1.0) gain, and smoothly but rapdidly transition to the desired amount of
gain or reduction, and then return.

"P" wrote in message ...
feeding the audio through an Aphex Compellor or some device with similar
ACG
function may be your best bet.

the Behringer Ultramizer may be a good alternative.

Thanks for the advice, however it's a not-for-sale venture and I won't be
forking out any more on it than I already have, other than my time.

if the end use is for broadcast you may not need to do anything at all.
the
processors at the transmitter will squish the audio just fine

It's definitely not for broadcast - permission granted only to video and
distribute to people who couldn't be there, or those who were and want to
remember. The effect you're talking about would be ideal - a dynamic
compression that adjusts according to input. I can't see anything within
Sound Forge that would do that.


im not familiar with the workings of SF but you can set just about any type
of compression/expansion in cool/edit/Audition or save the time and squish
it in real time with an Ultrimizer (about $50 used) or an Ultradyne (also
low cost). a Composer or many other types of comp/limiter should be usable
for you task, just set the attack and release times to slow.

Thanks Arny, I will check out Audition.

In the meantime, I think I've finally got my head around Wave Hammer in
Sound Forge. There are two parts to it - Compression and Volume Maximizer.
The idea seems to be to lop off the level from the high peaks using the
Volume Maximizer at a point with minimal distortion, then compress the
result so that anything above a certain threshold is boosted according to a
selected ratio. In preview mode the results have been OK but not perfect
yet, and both the processes in Wave Hammer can be done simultaneously. The
result has been a much fuller sound from the band, and audible speech in
between songs. The trick is to find the point that minimises distortion and
"compression fatigue" for the listener. But I think I'm on the right track
now.

just set the attack and release times to slow.

Thanks, I've been doing this with the release time and the results have been
much more distortion-free. Don't fully understand it! Why the attack time
also?

"P" wrote in message

Thanks Arny, I will check out Audition.

In the meantime, I think I've finally got my head around
Wave Hammer in Sound Forge. There are two parts to it -
Compression and Volume Maximizer. The idea seems to be to
lop off the level from the high peaks using the Volume
Maximizer at a point with minimal distortion, then
compress the result so that anything above a certain
threshold is boosted according to a selected ratio. In
preview mode the results have been OK but not perfect
yet, and both the processes in Wave Hammer can be done
simultaneously. The result has been a much fuller sound
from the band, and audible speech in between songs. The
trick is to find the point that minimises distortion and
"compression fatigue" for the listener. But I think I'm
on the right track now.

AFAIK Sound Forge has similar gain adjustment facilities as Audtion.

Dynamic range adjustement by automatic means can't equal that which can be
done by an intelligent operator.

Dynamic range adjustement by automatic means can't equal that which can be
done by an intelligent operator.

I just hope I have the intelligence (and the ears) for the job. I think I
do.

"P" wrote in message ...
just set the attack and release times to slow.

Thanks, I've been doing this with the release time and the results have
been
much more distortion-free. Don't fully understand it! Why the attack
time
also?



a typical compressor can be used in multiple fashions: fast attack, fast
release, high ratio, high threshold and it's a limiter. medium attack medium
release, moderate ratio and it's a compressor. slow attack, slow release,
medium to high ratio and low threshold and it's an AGC.

in addition some units have a 'gate' setting that 'freezes' the compression
level in place when audio falls below its threshold. this is useful to
prevent 'pumping'. in other words having the AGC/compression audibly
noticeable.

from what i can gather you want an end result of audio where the
listener/viewer can set a comfortable level and not be forced to make
constant adjustments up and down. this calls for gated AGC action and some
peak limiting (to tame clipping)

personally i like the all in one solutions offered by the dbx quantum or
Behringer Ultradyne. these do the AGC split the audio in separate bands and
compress, recombine and limit. the process is sometimes called 'finalizing'
the dbx DDP has a preset called 'finisher' that does a similar process (but
not multiband)