I'm trying to apply CIFilters to images (using CALayer) and movies (using QTMovieLayer). However, some of these filters return a bigger image than the original content image/movie, such as a custom composite reflection filter I've written.
My current solution is to render the images and movies using a CAOpenGLLayer subclass. This works fine for images, and is actually faster to redraw when filter properties are changed over a plain CALayer. But when trying to draw the frames of a QTMovie the performance is terrible. I'm using -frameImageAtTime:withAttributes:error: to extract a CIImage, but when I profiled it I discovered that it first renders a CIImage. Since this is a known (and measured) expensive operation I attempted to request other types of frame image (QTMovieFrameImageTypeCVPixelBufferRef and QTMovieFrameImageTypeCVOpenGLTextureRef) however these returned NULL, leaving me back where I started. I'm looking for suggestions about how to efficiently render video using Core Image filters and only the 64 bit QTKit. I've found plenty of sample projects, but they all uses functions deprecated/removed in the 64 bit version of the framework. Either that or how to suitably transform the output image of a filter chain so that it lies within the bounds of the CALayer rendering it. A last resort I have considered, would be to write a separate 32 bit process making use of the functions removed from 64 bit QTKit, to render into an IOSurface. Though I'm looking to avoid the complexity. Thanks, Keith _______________________________________________ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
