Martijn van Oosterhout wrote: >On Thu, Feb 20, 2003 at 09:19:26PM +0100, Bertrik Sikken wrote: > >>In the hp3300, the calibration table uses a 16-bit word for each >>calibration value, lower 6 bits are offset and upper 10 bits gain. >> > >Actually, I just remembered. Here is a huge difference. The calibration info >on the hp5470 amounts to 12 bytes per pixel. Three small 16 bit integers >(around 0x05e0) and three large 16 integers (around 0xa000). That why I >think that they're max and min values. At the moment they're hardcoded >because without anything it just produces nothing useful. > >I'll look into all this tonight. > Your idea sounds very plausible, since the hp5400/5470 seems to contain much more intelligence than other scanners with a niash chip. Only experimentation can tell how this works probably.
Regards, Bertrik
