How to use audio compression?

rbc_tn wrote on 7/9/2002, 11:43 PM

Sorry for cross-posting - I posted this first in the Video forum, then realised this is a much more appropriate place!

I have some video footage of a person talking where the volume of the audio is very low. Using VV3, I'd like to boost the volume for the rendered final video. Obviously I can raise the track volume by up to 12 dB, and I'm assuming this will carry through to the rendered video (haven't tried it yet).

I'm wondering if I should also look at audio compression. There are points on the video where the volume is much higher for a few seconds. I'm afraid I don't understand how to use the controls in the Track Compressor - Input Gain, Output Gain, Threshold, Amount, Attack and Release. The online help is pretty skimpy - no background on how this works. And it appears that audio compression is not covered in the VV3 manual.

Anyone care to write a brief intro to compression for newbies?

Thanks!

- rick

Comments

edna6284 wrote on 7/10/2002, 11:19 AM

Hi Rick,

Compression is basically reducing the amount of dynamic range (loudest parts vs. softest parts) in material. So you can ask the compressor to reduce the volume of certain parts of material that exceed a certain threshold.

The threshold control allows you to determine at what point you would like the compressor to act. The lower the threshold, the lower the level that the audio needs to be before the compressor begins acting on it.

The ratio controls how 'hard' or 'soft' the compressor will act on the signal once the level has exceeded the threshold. By 'hard' and 'soft' I mean how much the signal will be reduced. So a 1:1 ratio will do nothing to the signal (1 dB in = 1 dB out) but a 10:1 ratio will be very obvious (10 dB in = 1 dB out). With 'softer' settings, such as 1.5:1, the compressor will act very gently on the signal, so that its effects are quite subtle and the level will be allowed to exceed your threshold setting in a natural, but slightly attenuated, way. The harder settings, like 10:1, act basically like hard limiters and it is very obvious when the compressor is kicking in because it very aggressively stop any signal level exceeding your threshold setting.

There's usually always an output level control, since along with the dynamic range compressors necessarily reduce the overall volume of a signal. The level control allows you to increase the volume of the signal accordingly.

Attack and Release simply control, respectively how quickly the compressor reacts to signals that exceed the threshold value you've set, and returns the level to normal after the signal has dropped below the threshold again. It's not usually good to have an attack value of zero, since you will usually want some of the sharp beginnings of certain notes to be unaffected by the compressor, such as snare hits, cymbal, picked guitar and bass, etc. These are called 'transients', and they go a long way to making music 'sparkle'. For now it's best to leave the attack and release values close to their default values.

Generally speaking you'll probably want to stick to less intrusive compressor use, unless you want the compressor to be obvious in your recording. The best way to experiment is to record some acoustic guitar strumming (lots of loud and soft material) and work with the compressor to see how it affects your recording.

Good luck
D