2-pass processing of inputs, generating statistics
|Reported by:||burek||Owned by:|
|Blocking:||Reproduced by developer:||no|
|Analyzed by developer:||no|
I would like to suggest a new feature which would first read an entire input file (if it is a finite source, like a file, not a live stream) and generate a lot of statistics, needed for the further processing requested on the command line.
For example, something like this:
ffmpeg -generate_statistic -i input.mp4 -ss -45 -c:v copy output.mp4
would make a 2-pass reading of the input source. First time, it will read entire file, generating various statistics, one of which could be the number of total frames in a video stream. The second time it reads the input file and now it can "understand" that "-45" time point means "45 seconds before the end", because it has got a number of total frames, gathered in the 1st pass.
This would help implement other features too, which can't be easily implemented in a 1-pass reading of the input, which ffmpeg implements so far.