Room, room and room. You will not begin to get "professional" results until you have a good acoustic space to record in.

The FastTrack USB is noisy and has little gain. To compensate, you will need to use a microphone position that is closer to the speaker than would be optimal. Yes, you should apply some compression and maybe a little EQ when processing the recorded signal. Excessive compression will have the effect of pumping the noise (hiss) from the interface up and down, which could make the result sound worse than not compressing.

What audio software do you have?

I use Cubase 5. The rooom has acoustic panels all over it. I try to keep the speaker close to the mic and use ae a pop filter. When I have all the tracks down, then I apply noise gate, remove all the breaths, normalize all the tracks. All tracks have compression and a limiter on the inserts. I'm thinking about getting rid of the compressor, since as far as I know it works very similar as the limiter. I just want to get the highest possible level at the end. Will that do?

A limiter is a particular compressor which has no attack time and infinite ratio. Removing all attacks (limiter) on a voice simply destroys intelligibility. A regular compressor tames audio levels while preserving intelligibility when att the attack time isn't set too short. I don't understand why you need to normalize, if your limiter is correctly set. A voice must be loud, but it has to be clear also. The limiter must only prevent overloads on A/D converters.

While I agree with Boswell that a great room makes a great voiceover, many (most?) professional voiceovers are cheap commercials made in rooms that are simply dead. That's what you describe and it will probably be OK for this project. (I am assuming the project is instructional and your goal is maximum intelligibility - you aren't trying to make the speaker sound "good.") I am also assuming that you are using plugin compressors and limiters. You want to play very carefully with the threshold on these with each sample. There is a tradeoff between boosting the softest sounds and making them easier to hear and trimming the attacks and making them less distinct.

Laurend, post: 349268 wrote: A limiter is a particular compressor which has no attack time and infinite ratio. Removing all attacks (limiter) on a voice simply destroys intelligibility. A regular compressor tames audio levels while preserving intelligibility when att the attack time isn't set too short. I don't understand why you need to normalize, if your limiter is correctly set. A voice must be load, but it has to be clear also. The limiter must only prevent overloads on A/D converters.

I actually apply a compressor and a limiter and rather than having the intelligibility destroyed, I find the results crystal clear....I find the compression adds punch and the limiter actually adds clarity lol....odd.....I use a "vocal presence" pre set on the limiter (T-racks)

I agree on using both a compressor and a limiter. Compress first, and set it to apply gain reduction only in the loudest half of the sounds, and at a fairly low ratio, like 2:1 or 3:1. Set up a limiter (a look-ahead, brickwall limiter is best for this purpose) after that, and set it to brickwall the signal at just under 0dB (-1dB or less if you can), then turn up the make-up gain at the output of your compressor until the limiter starts to catch the plosives and peaks. The less often the limiter kicks in, the more natural your signal is likely to sound, but by adjusting the threshold and output gain on your compressor, you can achieve your favorite comprimise between loud and natural.