Well, it's not really possible to prevent clipping because a computer program has no way to tell how loud your next note is going to be! All it can do is display the audio signal on a monitor that lets you see when the signal is clipping. Garageband, though, has automatic clipping detection and will halt recording as soon as the signal clips.

The only thing you can do is use a compressor to lower the volume in the louder sections, which will avoid clipping pretty reliably if you set its parameters appropriately. But read up on compression first, it's a tricky beast especially for a noob. Also, if you record with compression you can't "undo" the compression effect later. I'd suggest lowering the input volume until there's no clipping at all, not worrying about the low volume parts, and then fixing things after you finish recording by applying compression to the track. There's a really good intro to compression in [[url=http://[/URL]="http://recording.or…"]this thread[/]="http://recording.or…"]this thread[/].

One final comment: low recording volume isn't a problem in general unless your pre-amp has a low signal-to-noise ratio. Good pre-amps have a very high signal-to-noise ratio, which means they can capture soft sounds very clearly. If you're using a cheap pre-amp, you will run into this problem where loud sections will clip, and when you turn down the input volume the soft sections will be fuzzy-sounding.

monitor the recording to keep it from clipping and going to low in sound?

What you are talking about here (ignoring 'mic technique' which is the free fix) is compression. You should read up on both.

Unfortunately from a purely equipment perspective you are unable to use compression in your recording chain. I've never used a USB mic but afaik despite the claims on the box they are pretty shoddy offerings.

There will (or should be) options for controlling the input level either on the mic, as part of the software provided by AT, or simply in the Windows/Mac Sound Control Panel.

You want to try and get to a point where the loudest bit of your singing is reaching about -6dB in Audacity. This will give you ample headroom and you can then use compression after the event using a free plugin such as the Fish Fillets.

Ideally you would chuck the USB mic, get a decent soundcard, a real mic, and an RNP and RNC. But this would mean spending $1k or so. While there are plenty of cheap bundles that include a wired mic, soundcard, headphones etc (I remember discussing a Lexicon one with somebody? Began with an L anyway) for about $150-250 you should be able to get basic, usable results with what you've got by not overloading the input. You will find everything is pretty quiet or when you turn up the speakers it sounds really hissy so you think you must have it set too low. This is in fact the noise floor of a USB mic and an internal soundcard and you'll have to get used to a certain extent to it being quiet until you've recorded and noisy once compressed - but sticking to that -6dB rule should be your first goal to minimise inadequacies in the equipment being brought to bear on your track.

Both myself and the poster above missed one distinction actually - do you definitely mean the clipping occurs 'when you record' - i.e. are you overloading the inputs so you are recording the clipping to disk, or is it simply that to try and get a good average volume you end up clipping the outputs?

If the former, you need to lower your overall recording volume at the inputs using one of the methods I suggested.

If the latter then yes compression will even things out, I'm still correct in saying that you cannot easily record with compression and a USB mic, nor should you, but compression after the event shouldn't be something that should scare you, just get a simple free plugin, and experiment.