I'm doing some experimenting with master bus compression during mastering, and I'm trying to figure out how much gain reduction is typically desired for this. Right now I have a ratio of 2:1 and approximately 2dB GR, and I'm noticing volume fluctuations between louder and quiet parts of a song (e.g. bridge vs. chorus). Any thoughts?

Hi, Vaultwit. Your probably gonna get so many different opinions without hearing the material your working on, it's gonna be hard to make a decision. If it's hard rock or metal the tendency is to ramp that ratio UP as far as you can without distortion. If it's an acoustic ballad you want to maintain as much dynamics, soft to loud as you can. Lower settings of the ratio will gradually change the soft and loud spots where higher ratios will squash the dynamic range so there is no difference what so ever. The real thing to look out for is pumping or breathing. This is the sound of the compressor raising and lowering so fast that you hear background noise being raised between whatever the instrument or vocal is hitting the comp.
There is no rule of thumb. Just try to set it so there are no audible noise problems with the comp itself. I wish you luck and I hope to hear your song posted soon.

There are sooo many variables involved, that I can't imagine anyone can answer that question without analysing your material first. Sidroe gave some valid general lines, though.

For starters: We don't know what kind of compression you have possibly used on tracks or instrument buses and so on. The dynamic range of the material in question is the key in starting to define the compression. Having no idea whatsoever of what you are compressing, it's hard to say how to compress it.

Sidroe and Kalle are right; there is no "automatic gainreduction" for all kind of music. I am looking for that plug in, but haven't found it yet ;-). But: much more important to music is a thing called "dynamic"; and i mean the difference between loud and quiet parts. This got lost over the years, some people call it "loudness war", compressing as much as possible. The result is a lack of dynamic, because the difference between loud and quiet is almost completely squeezed out of that music. But one rule is left: you can't get over 0 dB, cause then all the bits are filled with "1" and that's it ;-) HTH Michael

As noted above, it depends on the source material. I have songs with 9db gain reduction that sound the same, dynamics-wise, as other songs with 1db reduction or less. The former is just starting with a wider dynamic range before being limited, while the latter has a smaller dynamic range from the pre-limited mix.

As a general rule of thumb, though, I like to see steady GR under 3db for most dynamic material, and under 6db for more aggressively compressed music. In order to preserve punch, you have to balance overall fatness with crest factor, which is what you're reducing when you use a limiter.

It just depends on how and what. The more gain reduction, the smaller the dynamics get. There is no right or wrong setting for this. Its just like any other effect, use your ears and adjust to taste.

What? You mean your compressor doesn't have a "MIX GLUE!!1!!" setting? 

I just click the 'make the song rock' button. Works every time.