I'm not sure if it is Chrome's doing or YouTube's, but if I open a couple of youtube links in tabs (a selection from the "related" list to the right of one I'm currently watching for instance) the clips in the new tabs don't start playing until that tab gains focus. This avoids the auto-playing sound overlapping problem and means you don't miss the start of the clip and drop in part way through. Surely this is the solution to that selection of requirements?
I've noticed the Chrome feature as well. I watch a lot of Twitch streams, and it causes a frustrating experience, albeit with possible workarounds by Twitch. It's that Twitch is about live streams and live chat, but now the video autoplay is delayed until I activate the tab, but chat is still live from the moment I clicked on the link. It basically buffers the stream now, but makes me go out of sync with chat. One alleviator could be Twitch adding a "go live" button that shows up when I'm not live.
There are still two classes of problems unaffected by this change.
1) As is popular with gifs, there could very well be multiple of these autoplaying videos on the same page, probably even multiple fitting in the same viewport area. If they had sound, these would interfere with eachother.
2) My understanding is that many listen to music from some source unrelated to the webpage containing these autoplaying videos. So even if there's only one video, its autoplaying sound could easily mix with my tunes.
Doing autoplaying sound is hard, but with gaze tracking and knowledge of system sound usage status, I think a pretty good solution is possible.
