This routine, I think, should be reduced to just this:
// Returns true if the video mesh should be scaled to match the width of the screen // Returns false if the video mesh should be scaled to match the height of the screen private bool ShouldFitWidth() { return (mTextureInfo.imageSize.x >= mTextureInfo.imageSize.y); // if cam image is wider than tall, it should fit width }
The original code looks at screen width and height, but I don't think it matters, and was giving me the wrong answer in landscape orientation.
Ah ok, my misunderstanding. That's good to know, and I think I want it as an option in my project, stretch-to-fill or stretch-to-fit.
It was the Xoom and Xoom 2 that were different. And they were different from each other too. Ex: the Xoom's camera image filled about 1/3 of the screen, where same code running on a Droid was just fine. And I have not seen 'orthographic' work on any device, image is always too small. I suspect there is some assumption of position or something I'm missing.