I'll try to explain it. First, highly compressed video sacrifices quallty. It's the only way to get the file size down. More compression, more distortion, lower quality.
When editing from such files, the video is first totally decompressed, your edits applied and it still contains the original distortion. After editing, if you're saving the resulting video as the same sort of compressed file, it has to not only preserve the existing distortion, but also adds more distortion. Regretably, the distortion is at exactly the same places as the first compression's distortion and the compressor is going to have to give up file size to preserve that distortion.
The best way to preserve quality is to decompress to some much larger, less compressed (and less dstorting) format (DV is good, raw RGB is best - and largest file size), add a little bit of blur (to take the edge off the distorted spots), do the editing, and look at the results. If all looks good, recompress to the final desired format using two pass, variable bit rate settings that will give the file size you need. This should keep file size from blooming and save as much quality as possible.
Fav quote - "Experience is whatcha don't get 'till ya don't need it no more."
System - Athlon 1.4GHz, Win98, Hauppauge PVR250 receiver and compressor.
Software -Magix Movie Edit Pro 10, Nero 6 + NeroVision Express, Moho 4.61, PSP 8.1, Bryce, Quicktime 6.52 pro, Goldwave 5, DVD-Lab.
Cameras - Panasonic GS9, Canon ES8400V, Canon EOS D20 and Canon A70