Folks --
I am still in QC 2.1. Upgrade to leopard in the long future.
My problem is in a core image kernel. In the process of debugging a simple kernel ( I sample a pixel, then return it ), I was changing the values of the RGBa components, with the other two ( GB ) turned to Zero. I had a billboard plugged in.
The incoming signal is a "data Image". Pixels vary from 9970 to 0.0034 in value. ( values courtesy of mat lab and value finder kernel ). The core image kernel involved in a viewer. incoming pixels were to be converted to values in the 0 to 1 range. I was not getting good results. So I hard set the GB to 0.0,
and typed in values for red from 0 to 0.9. I got green and blue pixels which I thought I could not get.
Are the three pixel values hooked together in some fashion? I wanna retry this today, as I am not entirely sure I was working in a "pure spot" in the code.
Any help with this would be appreciated.
Special Thanks to Sabetelli @ Apple and Sam Kass. I finally have a convolution kernel of size 19 pixels for a bi lateral filter with 2 gaussians ( distance and value ) with variable mode padding being applied. Thanks for all your help folks!
--dan
