Document "programmatically generated audio" options

Currently it's not really clear in the documentation that non-real-time programmatic generation of audio is possible via use of `AudioStreamSample`.

While the upcoming 3.2 [real-time programmatic audio generation functionality](https://github.com/godotengine/godot/commit/e33764744cb2bf72ee77c823c3beeb6dc870d2dc) is required for *real-time* generation, it's not actually required for *non-*real-time generation. (Nor will the new functionality be particularly suited when one wants to loop a single generated buffer because it requires constant filling of the sample buffer with no way of looping.)

In addition, there's some subtle variations on what people mean by "programmatically generated" so it might be good to mention other related solutions, as I outlined in https://github.com/godotengine/godot/issues/18135#issuecomment-481976825.

An additional section in https://github.com/godotengine/godot-docs/blob/master/tutorials/audio/audio_streams.rst might be a good place for this information. (Note: https://github.com/godotengine/godot-docs/pull/2361 is a current pending commit to the same documentation.)