sound compression

I would use the compressor or duynamics processor on the individual tracks first.

You may want to give the levelator a try. It supposed to do what you want.

If you try it could you post back how it worked.

Dual 867 MDD / 12" powerbook  Mac OS X (10.4.8)

Yes, but you use the info window of the individual track, you have to click on the "Details" triangle. And I wouldn't use the Multiband Compressor on a single track, but the Dynamics Processor (which is a compressor and expander).

Hi there. I just finished a two part interview in my podcast with Rick Pepper, an veteran audio engineer on speech processing. He basically gives a master class on processing the voice. He goes through this step by step and I ask a lot of questions along the way. You can find the podcast and screen shots of the effects used at http://digiaudio.blogspot.com.

If your not interested in checking it out in detail here is what he is suggesting these processing steps on the vocal track:
1. Hi pass filter or graphic eq to gradually filter out content below 75- 80hz
2. AU Mulit-band compressor setting the attack rate to zero and mild output gain (5-7db). You really have to play with the rest of the settings to see where they match your voice best.
3. MDA bandisto VST plugin top get some sparkle in the high frequencies. This is a free plugin but it is VST so you need an application that offers VST plugin support - Amadeus II or Amadeus pro on the Mac are the cheapest options on the mac - $30-$40 US. That said, this is really icing on the cake. You can skip this step.
4. AU Dynamics Processor: take advantage of the noise gate and reduce the background noise and boost the output gain a bit more.

Now this requires 3 - 4 insert effects for one voice track. Garage band only offeres 2. So I would do the processing in another tool if you have one available. I do my recording in Audio HiJack, voice processing in Peak, Audacity, or Amadeus. Then I insert the processed track into Garage Band. If nothing else I would run the hi-pass filter (step 1) in Audacity and then multiband and dynamics in GB. The new Audacity beta has AU support.

Its sounds like you used the isight mic and moved around a lot. For best results in the audio, stay close to the mic. The farther away you are the more noise you introduce.

Powerbook G4 867mhz Mac OS X (10.4.8)

Awesome!!! this is the kind of info i was looking for.

I'm not sure. For your problem, all this sounds a bit like "shooting sparrows with a cannon", as we say in Germany. If your main problem is different levels of the tracks, you should set the right levels first, and a compressor is not the first choice tool for that (although it has a volume slider). You should probably normalize your sound files in a simple audio editor (like Audacity or SoundStudio). When they all have the same max volume, you might go ahead and use compression to make them sound more alike.

Yes and no. Doing a peak normalize may not help that much. If your peak level in the entire audio file is -1db, the peak normalize will only raise average levels by 1db.

RMS normalize would be better but to get that on a Mac you have to buy Peak Pro 5 for $599.

The Levelator does a really good job on uneven levels.

I was presenting a road map to getting good sound for speech which is a bigger set of steps than just evening levels.

Ultimately, it sounds like he may not have been close enough to the microphone which is really the best solution

Why not just use the volume automation to raise/lower the tracks when and where it is needed? Then compress to "glue them together" and flatten the dynamics a bit.

X

3 reasons I can think of

1. Raising the volume is limited to 6db
2. raising the volume via automation may cause you to clip (move over 0db) if your not compressing
3. raising the volume automation will likely cause ambient noise to be rasied to so you need a noise gate as well

But, there is more than one way to skin the cat so if this works for you, no reason not to do it