I have read a couple of good posts here and there about the theory behind converting pictures to an MPEG movie, but I had a few specific questions about the process for NTSC.

1/ SCANNING: What exact dimension should I scan my pictures in? 720x480 or 640x480?

2/ PHOTOSHOP: This is where I expand the width to 110%, right? Should this be done regardless if I use 720x480 or 640x480?

3/ PREMIERE: What exact dimension should I use for my Premiere project? 720x480 or 640x480?

4/ TMPEGenc: What exact dimension should I use for the MPEG-1 XVCD encoding? 720x480 or 640x480? (I know that standard MPEG1 files are smaller, but these higher resolutions work on my Pioneer 343 and give clearer pictures.)

5/ TMPEGenc: I experimented a bit and found a different between Maintain Aspect Ratio and Fill to Frame. Which is preferable to use when encoding the MPEG1?

Is there something I have left out here?