I'm trying to setup a workflow with automatic transcoding of ingested files to SD and HD with ffmpeg + audio normalization according to EBU R128 (loudness normalization).
I'm starting this discussion for others to chip in, if you have suggetions for how to best do this or if you are looking to do the same.
Ideally this would end up with something similar to TelVue Connect, where users can submit shows online and they are automatically transcoded and normalized. But in the first step I'm just thinking of setting up a Linux computer with ffmpeg for transcoding and r128gain (or perhaps even better FreeLCS) for loudness normalization.
My main challenge is how to setup ffmpeg to transcode to the best output. We basically use two formats: SD PAL (mpeg-2 576i) and H.264/AVC/mpeg-4 part 10 (720p/50). Does anyone have suggestions for ffmpeg commands to transcode to these broadcast formats?
The workflow would then be to copy a file to a watch folder that checks the video format, transcodes it if necessary and moves the file to the audio normalization watch folder. There it is loudness normalized and finally moved to the output folder, ready for broadcast and in a format that the playout can play.