use av library instead baby!

[BBC-Esq]

Use the "av" library instead, which "just makes sense" since faster-library already uses it anyways. This removes the pyaudio dependency as well as that of scipy and numpy. It also obviates users having to install ffmpeg since av is basically a wrapper for ffmpeg.

Here's what the faster-whisper repo says about it:

I didn't have the time to analyze other scripts for their use of pyaudio, ffmpeg via command line, scipy, etc., but I image that there are places you can use av instead to make things faster, easier for the user, and to simplify the code and reduce the amount of requirements.

It also uses PathLib instead of string manipulation. I'm a big fan of pathlib because of its syntax and it's more platform independent.

As always, let's discuss before implementing!

[makaveli10]

@BBC-Esq Looks good to me, thanks for putting this together.
Although, i see this warning in the resampling process:

Timestamps are unset in a packet for stream 0. This is deprecated and will stop working in the future. Fix your code to set the timestamps properly
Encoder did not produce proper pts, making some up.

[BBC-Esq]

I get the same error message but it doesn't seem to affect the quality or speed of the transcriptions. It might have something to do with timestamps though..Thoughts?

https://superuser.com/questions/1226305/ffmpeg-warning-timestamps-are-unset-in-a-packet-when-converting-h264-to-mp4

https://stackoverflow.com/questions/48927762/ffmpeg-timestamps-are-unset-in-a-packet-for-stream-0-non-monotonous-dts-in-outp