When I upload a video, I typically head straight to sentence mode. I have it set to translate so I can get the english context and see new vocabulary translations right away. Sentences can be quite clunky. 99% of the time its never a complete sentence. The first word of the next sentence is always spoken at the end of the current sentence. As annoying as it is, they are minor.
However, I then move over to the karoke section where its more fluent and watch the video in full context. Heres the problem (for me). I dont remember any of the words I learned. So I have to pause the video to look them up. Google translate camera setting is helpful but still tedious.
Now, I dont like having english translations on when trying to watch a video, its too distracting, but what would be nice if you could hover or tap over the word and a quick 2 sec bubble pops up with the translation