Tips needed for recording vocals

What should I do about the varied volumes during a verse? Should I record and set the levels so that the loudest parts don't clip but the quieter parts come through very low, then compress the audio? Should I compress while recording so that both the louder and quieter parts are recorded at an acceptable level? I've been working on just the instrumental tracks for a while and would like to begin record vocals to get practice mixing complete songs.

I would first try to get the singer to control the level him/herself by playing the mic. If he could turn away from the mic during loud parts and get a little closer when soft, then the level wouldn't vary so much. That is a good place to start. I would be scared to compress too much while recording, because you can't undo it later. Paul

I always record vocals with compression these days. While it is true that it is difficult to undo compression after recording (though not impossible), there is simply no way to get the sound of an analog tube compressor with a plug-in. Even if there is hardly any compression, just the sound of an analog tube compressor in the signal chain makes it worth using, IMHO. You could always try different thresholds and ratios until you're getting the sound you want. Even if you're not a wiz with compression, it really isn't that complicated. To be sure, I record most things with absolutely no signal processing whatsoever, but vocals are a different animal. They almost always need compression if the singer sings the way he/she wants to because there are going to be dramatic differences in volume levels -- often in the space of a single phrase. You can always add additional compression later and/or automate volume levels, but that analog compression really does make a difference. Live dangerously -- it's worth the risk!

If you don't want to compress to disk, try recording with a limiter. I'm not saying "squash your vocals". Set the limiter to only grab the worst spikes, so that the rest of the track is left unaffected, but you avoid that whole "perfect take with one clip" problem. I do this not only on vocals, but on almost everything I record into my 001. Again, not to alter the sound, or squash it, but just to get the very large spikes.

Another option, should you choose not to compress, is to set up two tracks with different gain structures - one for the louder passages and another for the quieter ones. You can record to them one of a couple ways 1) Set up just one mic and do a pass to the "loud" track first, with your mic pre set for lower gain, then do another "quiet" pass to the "quiet" track with your mic pre set for more gain. Then, comp/edit the two tracks together. or 2) Have two mics next to each other set up to record the singer - each going to their own properly gain-staged micpre and track. Again, One for "louder", another for "quieter". And again, comp the two tracks to one.