The overlay mixer provides the best performance (in the machines I've tested, at least) as AFAIK it uses hardware conversion and scaling. It is a bad thing to need to switch to another renderer to see subtitles (these others just don't render as smoothly in my case).
Why is this necessary? Can't the text be printed on the decoded YUV frames before displaying? Just modifiying the Y plane could be enough, maybe. Is it too hard to implement or is it just not possible?
Log in to post a comment.