Combining audio files from different sources

Thanks, Bob and Steve.

tomcalwriter  wrote I have been doing it separately but have found a lot of variability in dynamics and difficulty in matching volume once they are combined.

Just to add a bit to what Bob's said here: Before you do anything else, normalize all of the tracks to the same peak level - in this case 0dB. If you don't do this, then the results of any compression or limiting will be variable - possibly making the situation worse, not better! The reason for this is that all compressor and limiter settings are referenced to 0dB, so unless the files start with the same peak levels, the effect will be different on each one.

In order to preserve most of the dynamics, but not have massive level differences, limiting the tracks above a certain level (and this rather depends on the peak values you find) is a good starting point. So for instance, after you've normalized the tracks, try running a peak limiter at, say, -9dB on them, and then applying a few dB of makeup gain. This will make it much easier to sort out the final levels for each one. The snag with compressing the track is that compression almost inevitably raises the noise floor, but limiting on its own won't do that - the only noise floor increase will be caused by the gain makeup. Don't take -9dB as gospel; the level you need may be different, although that's not a bad starting point. The idea though is to keep the peaks down and the 'body' of the sound at a more consistent level. When you listen to this, you'll hardly notice the difference - except that everything will appear to be louder!

In any case, what suggestions might anyone out there have?  The reverb particularly is a problem and I have installed the plugin DeVerberate that is kind of hit and miss.

My experience with any form of reverb removal is that the side effects are more distracting than the reverb in the original. But, if you're going to do this then almost certainly it will be better applied to a file that's been limited and level corrected first, although without actually trying it, it's hard to be sure about that. I'm only mentioning it really because if it doesn't work one way, it may well be worth trying the other.

Since noise reduction is a destructive process, I'd do that part of the process separately on each portion of the recording (make a copy for destructive editing so you don't lose your original).  Once that's done, I'd also normalise each portion to whatever arbitrary level works for you--say -3dB(FS).

Once that's done, assemble all the tracks in order and apply the compression of your choice to the whole track--properly set up this should even out the dynamics.   After the compression, you may find that you need to normalise everything agin.