Hmmm...I guess the least lossy way to do this is to construct a series of AviSynth scripts, one for each job and pump them through TMPGEnc. I seem to recall some issue with frameserving so I'll need to research that.
FWIW, I've found storage codecs are causing a problem. HuffYUV or the more recent more-compression lossless codecs are nice but require a LOT of space.
There's an idea for a data-based script writer starting to gel in my mind. A text file would be constructed with one-line entries consisting of comma-delimited variables that are used to write a series of scripts. The variables include source file name, and other variables particular to the source. The output is proper AviSynth scripts, one for each clip and, hopefully, a properly constructed set of queue data for TMPGEnc.
The reason is TV captures. Natural breakpoints are at commercial points and the easiest/safest way to enforce these is to construct a series of files, one for each segment. Filtering isn't so tough, a script modification could limit the frame numbers.
Oh, as I'm obviously thinking with my fingers right now, maybe the missing part is import of variables and proper conditional processing in AviSynth....(shuffles off looking for sh0dan...)
|