The difference matte works by taking two images and subtracting one from the other. A-B=C. Whater pixels are the same in layer A and B, are made transparent. THe tolerance control will change how close of a match the pixels have to match to be considered the same. So basically youre going to have to have a frame of your scene empty, and then the footage with the actor in it. It is important that your footage is shot with a tripod, to keep the background the same. In layer 1 (the top layer), have the footage with the actor in it. Then in layer 2 the still image or footage of the empty backround, making it the same length as layer 1. Then you apply the difference matte to layer 1, select the source (I think that's what it's called, I'll check at home computer), as layer 2, then play around with the tolerance untill you get the results you want. Finally you can precompose this layer so it is one easy layer to use. However be careful with using the difference matte, as it isn't very reliable. Noise and other artifacts on footage from consumer-level cameras can cause problems with the pixels being the same. Hope this helps.
Peace
 | 
Linear Mode
