2016-06-22, 13:13
(2016-06-22, 07:15)Memphiz Wrote: I think the difference here is that the other rendercapture implementations capture the visible pixels only - mmal capture captures the decoded format - is this how it is atm?
On Pi rendercapture will capture all *decoded* pixels from the video.
Normally that would be something like 1280x544 for a widescreen video, even if the area above and below video appears black on a 16:9 display.
However this video encodes the black pixels. We have no way of knowing which pixels are the black bars without iterating across them and I'm pretty sure
rendercapture won't do that on any platform. If it is done at all it would be more reasonable to do at the boblight level.
