Feature request: Ability to have much less samples

[Joshfindit]

I have files that are 60-500 minutes, and all I need are at most 900 samples.

An example file: 141482 frames, (61:35.856) 3695.856
Even --pixels-per-second 1 gives 22K of JSON data.
We can pre-calculate the desired 'pixels-per-second' value (in this example case, it would be roughly --pixels-per-second 0.2165. 500 minutes would be roughly .0267), but audiowaveform seems to only take an integer.

Blue-sky, it would be great to have a '--desired-samples` that uses the first frame, the last frame, and calculates the spacing needed to reach the right number of frames, but that feels like a complex ask. :)

[swordsreversed]

+1

[chrisn]

audiowaveform computes the data you need in order to output PNG images, but I'd need to give some thought to the command-line interface for it to output JSON or binary data instead: --start, --end, and --width get us most of the way there...

[timobehrens]

+1