[Image-SIG] 16-bit unsigned short: PIL tile descriptor?

[K.Arun]

Hello,

        I'm trying to write a PIL plug-in for a in-house image
format. It's a very simple format with a 1024-byte header followed by
unsigned shorts in big-endian byte order. I tried using both the 'raw'
and 'bit' decoders without much success - while 'L' works with 'raw',
in that I don't get any errors, my images look weird when
processed. Using "F;16B" in the parameters tuple for 'raw' mode (what
should self.mode be in this case ?), results in a 'ValueError :
unrecognized mode' exception being thrown. Could someone throw light
on what self.mode needs to be set to and the correct tile descriptor
values to use ? Thanks,

                                                  -arun