Christian,
On Sun, Jul 4, 2010 at 9:50 AM, Christian Henning
Nathan, it just occurred to me. Why don't you use bit_aligned images for your sample code below? You min and max are the same.
typedef bit_aligned_image1_type< 12, gray_layout_t >::type gray12_image_t; typedef bit_aligned_image1_type< 14, gray_layout_t >::type gray14_image_t;
Thanks for the suggestion. I considered that when I first adapted my code to use GIL. At the time I decided not to use them because the data (as provided by hardware over which I have no control) is byte-aligned -- there are 12 (or 14) bits of data and 4 (or 2) bits of padding. I may be misunderstanding bit-aligned images, however. I will be needing to add bit-aligned pixel types soon, however, and I'm pleased with how simple that should be. Thanks, Nate