2002.07.05 00:24 "Re: Photometric interpretation for grey-scale JPEG", by Jon Saxton
** Reply to message from Tom Lane <email@example.com> on Thu, 04 Jul 2002 13:40:38 -0400
Actually, wait a minute. Are those files PlanarConfiguration 2? If so, it'd be legitimate (in fact necessary) for the individual color planes to be represented as grayscale JPEGs, while PI will denote the overall colorspace interpretation to be given to the reassembled data. But as long as PlanarConfiguration=1, PI should match the stored colorspace in the JPEG images.
The images (roughly a million of them) were generated by one of our major financial institutions. They are straight greyscale with PlanarConfiguration = 1.
I suspect that they were captured as colour images in YCbCr and then downsampled by simply discarding the chrominance components. The TIFF writer probably didn't bother changing the PI, regarding YCBCR as an accurate description of the colour model given that the samples per pixel value was 1, implying just the Y channel.
While I recognise that the TIFF specification calls for MINISBLACK on grey-scale images, it occurred to me that maybe in some sense the supplier of these images is correct and that (perhaps in some future publication of the standard) YCBCR should be permissible for grey-scale JPEG images.
Regardless of the answer, I have to deal with these images. I also have to deal with millions of images with other encoding faults and correct them on the fly so these PI=6 images present no problem compared with catering for viewers which ignore the Orientation tag and which won't handle Compression=7.
Jon Saxton <firstname.lastname@example.org> Software developer for UNIX, OS/2 and Windows
U.S. agent for Triton Technologies International Ltd