[tyingq]

That seems like it would be tricky, as the video and audio would get out of sync. You would have to remove, then "fill" to keep the timing. Though this product does mention it works with multiple speakers on different tracks...so they are already somewhat in that space.

[autoencoders]

For video is quite tricky. One thing with Video is that you don't want to over edit the audio, since its then very hard to keep the video synced. That said for explainer video it should work ok, but for a Video Podcast it would be horrible. I have an idea how to deal with this, but this is not now available.