I would guess, based on problems I’ve run into, that Daniel’s statement of “develop in house solutions” is probably about right for whatever software they’re using. Not necessarily scripted solutions, but different tools for different parts of the process.
“Compression”, on the software side, isn’t just a one-tool job if you are die-hard in getting the absolute best quality. There could be better solutions than this, but where I’m at now is: if I need to de-interlace it first, then I use FieldsKit or Magic Bullet Frames. If I need to resize it from rectangular pixels to square, then I use VirtualDub resizing with lanczos3. If my target is FLV then I use FlixPro. If my target is H.264 then I use Procoder (I’ve tested against MainConcept Reference- and for my test footage I did not see much of a difference, but I’m sure the opposite is true sometimes. I’m satisfied enough to use Procoder though). It’s a slow workflow, but it’s using what I find to be the best tool for each stage of the process.
If you’re looking for a “swiss-army-knife” solution from source to H.264 or MPEG-2 on PC, and don’t want to deal with avisynth and virtualdub, then Procoder is your best option in my opinion- much better than Squeeze in my tests. Though that was a while back and against an older version…
We’re moving to Apple though, and can’t budget for Episode at the moment, so I really hope Compressor 3 will do all the jobs pretty well (I’m trying to convince the director to move over to progressive so we won’t have to deal with deinterlacing)!