What's the difference between compositing an image and just plain drawing it?
Another question: I have a custom NSView, and the graphics in it scroll up, but it continues to draw onto the window beyond the bounds of the rect. How do I avoid that?
If you limit your drawing to inside your -drawRect: method, your drawing should be clipped so you can't draw outside of it. If you want to draw into your view from anywhere else, remember to call -lockFocus before you start and -unlockFocus after you're done. -- Bo
I encountered a similar problem and have reported my solution here: CustomViewDrawsGarbageOutsideRect.
"Compositing" refers to operations combining pixel images (bitmaps/pixel maps). The window server does a lot of compositing of the bitmaps of the windows, for instance. Generally the compositing operations take an NSCompositingOperation, which defines how the source and destination pixels are combined. ++MarkDalrymple
You can use composite..., draw..., and/or dissolve... methods when drawing an NSImage.
From the Apple docs on Drawing and Images:
Images displayed using the various composite or dissolve methods, such as compositeToPoint:operation: or dissolveToPoint:fraction: , have the horizontal orientation of the base coordinate system; they can't be rotated or horizontally flipped. <some text deleted>
Images displayed using the draw methods, such as drawAtPoint:fromRect:operation:fraction: , have the orientation of the current coordinate system; they are rotated and scaled as defined by the view into which the image is drawn.
I ran across this distinction when trying to scale an NSImage proportionally to fit into a custom view. When I used composite... the NSImage was always full size, but it scaled and shifted as expected when I used draw...
Which is faster? Compositing or Drawing? I have a preview box in my current project, which lays image previews from QTMovie files on top of each other (the NSImage is grabbed via [movieObj posterImage], which works so-so). However, the background color is drawn, and these images are all drawn using drawInRect:, when the NSView's drawRect: method is called, and the performance is crap. The background color is picked with a color well that is attached to the view using bindings. As the user drags in the color picker, the color changes, but because the drawRect: method is called every time that color changes, it's pretty darn slow. Which drawing method is faster/more efficient, and is there a way that I can get the images into a smaller footprint or reduce the performance hit when drawRect is called? The end user may end up having 20 or more of the images being drawn in the view.
Cache your multiple images. Draw them all into a single blank image using [un]lockFocus, then draw that one image into your view when needed. You avoid scaling and blending your 20 images every single time something changes.