Compressing final mix

[chuchosay]

I know most software compression (and some hardware compressors) have the option of compressing based on RMS average value or peak value.

I know what the difference is, but which one should you use on compressing the final mix? I would think the peak detection method would be best, but I would like some other ideas based on more experience than what I have.

Thanks.

 

[mixsit]

Depends. What effect do you want on the mix?
Very fast attack and release to get small peaks but remain more or less 'transparent' a-la Waves L1. As you slow it down (and perhaps use more) it can leave more of the short peaks that keep it lively, but maybe pump a bit with the rhythm. With plenty of options in between.
Try some fast settings and watch what it does to your reverb and ambiences. A of a sudden you may find your hearing them much more and might want to change that aspect as a result.
Wayne

 

[Igormeister]

Use RMS detection compressor with attack time 25-20 ms, and release time 0.5-20 ms, ratio 2:1 - this allows transient peaks to pass, but compress average RMS level and increase loudness without pumping - overall definition and clarity of mix will increase
Then chain fast brickwall limiter, like L1 Ultramaximizer, to cut transient peaks. This is LOUD mastering
You can use multiband compressor to have total control over freq response- experiment.