Audio Question

[Rob Forsythe]

Yes.

Compressor/Limiter.

But you will be much better-off MANUALLY adjusting for EACH voice as Compressors tend to have noticeable audible “pumping” artifacts.

If you want to try the Compressor:

Add the Compressor/Limiter Audio Filter to a clip and click it into the viewer “Filters” window.

As a starting point, adjust the Settings to my “all 2’s” formula:

Threshold (db) -20
Ratio 2
Attack Time 20
Release Time 200
Preserve Volume CHECKED (very important!)

Now click the viewer to display the actual audio track and set the Level to “2” or “3.”

Now, when you play the track it should be at a “more consistent” level.

You can play with my “all 2’s” formula if you want, but its a good starting point for compressing the levels without a lot of compressor “pumping.”
The “Threshold” setting will affect the clip the most… as you slide left, it brings more of the low levels UP.
The actual volume adjustment now comes from the “Level” slider on the actual audio track (and you can key-frame that if you want.)

[Ray Lane]

Well, I know I can do it manually, it just takes a long time to go back and forth, back and forth. It would be great if there was a simple plg that did all the “rubber banding” of the audio volume, and normalized all the audio to the required level.

The compressor/limiter kind of does this, but like you said, it pumps the gain a bit.

[Rob Forsythe]

[Ray Lane] “It would be great if there was a simple plg that did all the “rubber banding” of the audio volume, and normalized all the audio to the required level.”

Unfortunately, software is not quite smart enough to “just KNOW” what we want it to do.

That’s why compressors PUMP… they try to make ALL the audio the “same level” by (in effect) “rubber-banding” ALL of it to within the specs you set up.

If you want the gain to be adjusted in an “intelligent” way, a real-live PERSON must make the decisions.

An alternate way to make this go faster is to SLICE the CLIP every time person “B” (the “low-level” person) is talking (making each person speaking a distinct clip unto itself.)
Then, adjust the level for that person on the first clip, and “Paste Attributes” for each additional clip of person “B” speaking.

If you need more than +12 db, you can add the Audio Filters > Apple > AUPeakLimiter and adjust the “Pre-gain” to any level you want.

[Nick Meyers]

what you’re after ray, is a “normalize” plugin.
unfortunaltey, FCP doesnt come with one (vegas and premiere do, i beleive)

Sound Track Pro might have that function.

[Rob Forsythe]

“Normalize” can be just as problematic to add “wholesale” to a poorly-recorded clip, in its own, way as is a compressor.

For the most part, “normalize” examines the whole clip and sets the HIGHEST PEAK to a given level, allowing the rest of the track’s levels to remain (differences and all) unadjusted.
This is useful if the entire track is recorded correctly (within itself) but the general peaks don’t match OTHER complete clips in maximum level.

Normalize is effective if you are, say, taking several CD cuts from various albums and making a “mix disc” and you want all the varied cuts to have a “normal” peak level throughout the CD.
You don’t want to disturb the dynamics (relationship of high-to-low-levels) WITHIN each cut, but you don’t want the peak level of any cut to be vastly different than the others.

However, the way I interpret the situation in the case of this thread: TWO different voices are recorded TOGETHER on one track with one being LOW in level and the other much HIGHER.

Any kind of “automatic filter” applied to try to force BOTH of these situations to “match in level” would need to be set so BROADLY that the filter’s automatic operation would reveal itself in audible pumping and/or distinguishable background-level changes.

One possible “fix” would be to create a “stack” of filters customized for the lower-level voice.
The stack might include an EQ or roll-off to lessen the amplified background noise, and a Normalizer (if you like) or even the Compressor (that is actually adjusted to specifically address the unique conditions of the lower voice). Once these are set, they can be added to MANUALLY- determined slices of the audio track. If any of the transition-points are still “rough”, an audio dissolve could be added at the cut-points).

[Patrick Boyle]

As this was posted back in 2005, is this still the case (having to manually raise the gain for each person speaking on a single track) or has the technology moved on and provided one for us? A plugin/filter that is.

[Nick Meyers]

the situation re plugins/ filters *within* FCP is the same.

there is an app made for blogers called levelator:
https://www.conversationsnetwork.org/levelator