Recently, I began remastering a collection of instrumental pieces which Sev, Brian and I recorded a while back. We'd originally recorded them on a four-track tapes. Unfortunately, there were three of us, and the pieces were written for five parts, so we recorded three parts (bass, guitar and recorder) on a stereo pair of tracks, and I went back later and put the remaining parts on the two empty tracks, one part at a time.

It worked okay, but sadly, there was some popping when I transferred it from the tapes to my computer so that the songs could be burned on a CD. Recently, I found the master tapes, and began the process of retransferring and remastering them using Sonar (which I hadn't had back when we originally did the songs).

I got everything sounding great. Everything was nice and balanced, the effects plugins sounded nice and crisp (better than the outboard processors I'd used for the original CD). So I burned them onto CD and played them in my car on the way to work. I could barely hear the bass. My computer speakers include a subwoofer, and I'd tailored the sound to that system. Back to the old drawing board, as they say on Mars.

Unfortunately, when I turn up the bass, there are occasional notes with a little more punch, which get too loud. Normally, I'd just slap a compressor on the track, which reduces the volume when it gets too high — but since we've got three instruments sharing a stereo pair of tracks, the recorder and guitar are also reduced whenever the bass gets loud enough to activate the compression.

Last night, I was reading a review of some of the effects plugins that will be bundled with the new version of Sonar (drool!), and one of them is a multiband compressor; you can divide each track into frequency bands, like a parametric equalizer does, and apply different amounts of compression to each band. So it would be possible to isolate the bass' frequencies and compress it separately while leaving the guitar and recorder unaffected.

Unfortunately, the new Sonar won't be out until October, but this morning I got to thinking: if I apply a parametric EQ to the tracks, I could drop out the recorder and guitar entirely, and bounce just the bass to a separate track (and then reverse the EQ so the guitar and recorder are on their own). Then I could apply a regular compressor to the bass track and solve the problem.

Now, why didn't I think of that earlier? It's so obvious.