Make waveform service more flexible by allowing pre- and post-filters to be configured
The waveform generated by the waveform service could be improved in several ways, e.g. it is weird to have empty areas in case of silence. Also, the representation of the audio level is tricky: with the current default settings, we have seen recordings with actually good audio levels where the video editor indicates a very bad audio level.
The good news is that ffmpeg can deal with all that. The bad news is that this requires ffmpeg filters which cannot be used by the current waveform implementation.
For a slight insight of what I mean, have a look at https://stackoverflow.com/questions/32254818/generating-a-waveform-using-ffmpeg
The goal of this issue is to make it possible to configure filters that are preprended/appended to the filter used for waveform extraction. That way, everybody can decide himself what the best way to generate audio waveforms is (no, it is not obvious).