Hi David,
Good starting point settings for a compressor for speech are:
Set the threshold to 0. This tells the compressor to leave uncompressed any signal that stays below 0, and to compress any signal that exceeds 0.
Set the ratio to 3:1. This is how much compression will occur when the threshold is exceeded. For every 3 dB over the threshold an incoming signal goes, the compressor will output only 1 dB.
I’m not familiar with Soundtrack Pro, so I don’t know what other parameters you can adjust, but other common ones are attack, which tells the compressor how fast to react to a signal;
release, which means how quickly to “let go” of a signal; and output or makeup gain.
In your case, if the dialogue is not peaking but the applause is, adjust the threshold so that the compressor only “works” during the peaking sections. Adjust the ratio to get the applause to stay out of the red.
In general, if you are getting more than around 6 dB of comression on speech, then you need to back something off, either the ratio or the threshold, because you will begin to hear what is called “pumping” as the compressor alternately compresses and releases the signal.
Peter