I'm looking to build an automated render pipeline that will crunch proxies for a large volume of footage at one time. I'd ideally like to run the script on a batch of arbitrary files that could include image sequences, audio, and any video format. The script will create an H264 preview of each clip, as well as extract all available metadata on every file.
It's a tall order, and I'm hoping it's possible with ExtendScript and Media Encoder or Premiere. I've dug as much as I can through the Object Model Viewer, but the lack of proper documentation is making this very difficult. I'm hoping to get a discussion going here that might provide help to future Googlers with similar ideas.
My workflow will be this:
1. Load a file (need a way to "stitch" image sequences and multi-file clips e.g. R3D into a single clip)
2. Extract metadata from file and save to external document (ideally including available source camera settings like ISO and Kelvin)
3. Set output name and destination for rendered file.
4. Render to H264, MP3, or JPEG depending on input source.
My impression is that it'd be best to load files into Media Encoder, extract, render, and restart as opposed to loading them all at once. I need to be able to handle thousands of files in a batch, and my experience is that the GUI doesn't handle this kind of volume well.
Looking forward to anyone's ideas. And maybe some documentation soon!