Victor Mote wrote:
I'm no expert here, but I waded through the TIFF spec a bit, and offer the
following paraphrase, which I ask you to either confirm or deny:
<paraphrase>What TIFF calls sample-per-pixel is the number of channels. So
WhiteIsZero can support bi-level and grayscale images through that one
channel. But RGB photometric interpretation needs at least three channels,
and only has 1 to work with, therefore no color. RGB Palette could use that
one channel for a 256-color palette, but it is unsupported. So the bottom
line is that FOP native can't support color TIFF images at all, RGB or
otherwise.</paraphrase>

RE: Sample per pixel versus number of channels. That's right. AFAIK, the terms "channels", "bands", and "samples per pixel" are all synonymous.


RE: Photometric interpretation. I think I may have exposed a level of detail that makes this more complicated than it needs to be. The code in FOP supports values for the photometric interpretation field of 0 and 2, which translates to WhiteIsZero and RGB. However, since only one band/channel is supported, the fact that the code doesn't bail on a photometric interpretation value of 2 is just a bug. So the final text could probably be reduced to:

"FOP's native support for TIFF's is limited to single channel images (i.e., bi-level and grayscale). Said images must use CCITT T.4, CCITT T.6, or JPEG compression, and must use white-is-zero encoding (TIFF's 'photometric interpretation' field)."

The rest of this I follow, and will make changes to the website accordingly.
Again, thanks for your efforts here.

It's my pleasure to contribute some small amount to a project that has proved immensely valuable to me. :-)


Ben


--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]



Reply via email to