DisplayPixelsProc and 16 Bit image

Hi,



So far I have seen from the SDK's examples (Dissolve.8BF) DisplayPixelsProc accepts 8 bit data only.



This means 16 Bit data must be converted to 8 bit before calling DisplayPixelsProc. DisplayPixelsProc then converts the data to the monitors colorspace and displays it.



This method doesnt work correctly for 16-bit linear RGB data. Information in deep shadows is lost and this results in visible color banding in deep shadows.



Is there a recommended way to handle this problem? Must I convert the images colorspace myself? Must I write my own display-methods or can the colorspace handling of DisplayPixelsProc be disabled?



TIA,



Peter
DisplayPixelsProc and 16 Bit image
Well, yeah - 8 bits/channel is crappy for linear (gamma 1.0) data.



You can use the colorspace conversion and depth conversion callback routines.



Or you can do it yourself.

lips