From: Michael R. <mr...@us...> - 2004-03-29 16:44:42
|
Hi James, > > I thought you were going to introduce VO_CAP_CROP? What if XShm does not > > have this capability? > > It would indeed work that way, but I'd prefer a solution that could be > easily applied to most (if not all) of the video drivers, particularly as > the point is to improve performance which is often an issue when using > xshm. I understand. So what you need is a way to propagate the cropping info to video out already at get_frame(), right? I think we need a flexible way to extend get_frame() requests anyway: If we want to do automatic conversion of frame formats and automatic insertion of deinterlacer post plugins and similar things, get_frame() has to become more powerful, so that the engine can determine before the first frame is processed, what post plugins need to be automatically inserted. I think get_frame() should look like this: vo_frame_t *get_frame(xine_video_port_t *, frame_request_t *) where frame_request_t can be extended to contain more details on the frames. Michael -- "#define QUESTION ((bb) || !(bb))" -Shakespeare |