Question to audio compressor

[hellboy80]

What setting could you recommend me for voiceovers? When my narrator speaks - some words are very loud and some of them are very quiet (I told her about it, but it seem she cant do anything about it ). I could correct it manually, but its very time consuming. I heard audio compressor may help, but I have no experience with it.
What settings could you recommend me? Important notice: according to my client, my video has follow some technical parameters: the audio has to be at -12db audio level (and can reach -9db only occasionally).

 

[ajcourtney]

according to my client, my video has follow some technical parameters: the audio has to be at -12db audio level (and can reach -9db only occasionally).

This makes no sense because you don't know what the reference level is. IOW, what absolute value does 0db equal? Unless, the person is referring to 0 dBfs, we have no starting point.

Anyway, the name of the game today is not just to compress your audio for leveling purposes, but to smash it within inches of its life. So this kind of request is definitely goofy IMO. Since it sounds like you're also serving as the mastering engineer, then you ordinarily should maximize your levels using longer word lengths (e.g., 24-bit depth) before dithering down to the traditional 16 bit finalized audio. Anyone further down the chain could always attenuate your final audio while retaining the highest possible fidelity you achieved during mastering.

If you don't have a lot of maximizer plugs available to you like Waves L1/L3, TC Powercore MD3/6K/Finalizer stuff, etc., then you will probably want to start playing with Sony's Wave Hammer which uses a compressor on the front end to feed the limiter/maximizer. There are some decent presets to start you off with. But your "average" vocal track level is going to determine where you want to begin to set your threshold level. From there, you can follow two routes with the limiter side - (1) feed it a fairly smashed level from the compressor and let it simply limit your peaks or (2) feed it a lightly compressed (read "fairly level") signal and smash the track with the maximizer. I haven't played around with the Sony plug enough to know which approach is best with that particular code since I primarily maximize with the Waves L1 and the TC stuff.

Traditionally, you could push the maximizer stuff harder than a simple compressor and avoid a lot of sonic artifacts since the maximizers have historically used look ahead algorithms and were just more robust in smashing audio.