On 8 Mar 2014, at 2:38 pm, Trygve Inda <cocoa...@xericdesign.com> wrote:
> Profiling this shows lots of time (90%) is spent in drawInRect and it takes > about 2 seconds to run. Is there a better way to do this that is faster? > > Maybe keep everything in CG and only convert to NSImage at the end? Sounds like you answered your own question... > Not sure > how to do this. Instead of creating canvas as a NSImage, create a NSBitmapImageRep and then a graphics context from it, using +[NSGraphicsContext graphicsContextWithBitmapImageRep]. Then you can use CGContextDrawImage with your CGImageRef data, targetting that context. Once you're done, just add the bitmap image rep to an NSImage. Essentially, you want to avoid copying any image data that you don't need to, though in fact your original code *may* already be avoiding that and just wrapping things. Spending 90% of your time in drawInRect is probably a good sign - it means that only the copy that really needs to happen is taking time. It might be that you can't do better (though 2 seconds sounds high, but you haven't said what size these images are). The other thing to ensure is that everything uses the same colour space throughout - if any conversion is needed it will hit performance hard. Lesser performance hits can come from mismatched image buffer formats, so if all your CGImageRefs are the same, use the format information associated with them to create your NSBitMapImageRep to match. If format and colour space are the same, the actual bitblit needed should be very fast as it's just a straight copy, no conversion. --Graham _______________________________________________ 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: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
