Compare raw audio data streams (e.g. PCM) by their spectrograms or hashes

[BinToss]

A sub-issue of #1917 to break down that feature request to smaller steps.

Some people may request comparing audio by their waveforms, another graphical representation of audio data. However, waveforms only represent samples by their gain level. They do not represent the gain of each frequency that makes up a sample, unlike spectrograms. A spectrogram can be rendered via ffmpeg's showcqt filter, a custom implementation, or a third-party library. Alternatively, a non-graphical approach would hash the data streams and compare their hashes.

To reiterate, this issue is just for comparing raw audio streams.

---

Only after all of this is implemented (perhaps via a plugin which requires and utilizes the host's ffmpeg?) should audio file format metadata and decoding be implemented, but only for the top 15 or so most popular audio file and stream formats. Later, the audio-related features can be expanded by request.

Full-fledged "support" for audio comparison is a much larger task than users would think it to be.