RFC: Direct Grayscale Type Constants

[Steve Sisak]

The next issue is "direct" (no CLUT) grayscale.

The 8-bit grayscale supported by the Apple 'raw ' codec is inverted 
(black=255, white=0) from what most image processing applications 
expect (and what is produced by most cameras.

Apple's 16-bit grayscale codec

   k16GrayCodecType              = 'b16g',

Has the normal orientation (black=0, white=65535)

We are currently using:

enum
{
     // Direct gray (black = 0, depth = bitsPerPixel)

     kDirectGrayPixelFormat		= 'YYYY',
};

For consistency, I would like to change this to:

    k8GrayCodecType		= 'b08g',

and add

    k12GrayCodecType		= 'b12g'

for 12-bit gray, packed in the standard Macintosh format, making the 
algorithm for big-endian packed grayscale 'bNNg' where NN is the 
depth.

Note that these constants fall within the Apple reserved range, so I 
would much have an OK before shipping with these constants.

(I know the issue has come up before on the list)

Cheers,

-Steve