Hi Ilia, oh-ha, thanks for info.
" ... These calculations gave near the same tables as original Windows drivers sent...." if I look at the two histograms - I could mail them to you - it looks like a 'not too small' difference. "...That could be because the scanner receives two tables (maps), I've called them calibration ones. And in the SANE backend I've used just a simple calculations...." From what does the scanner (you mean the hardware?) receives the tables, form the frontend? or does the backend receives the calibration table form the scanner? I believe, I haven't figured it out in detail. But if you do some increments and decrements of the grey- or color-values and the values are concentrated at one end of the histogram, my desired ;-) details (of the image) could be lost. "...Looks like the maps are for each color channel..." The idea with the color channel could be right, because scanning my drawings in color-mode gives the same result (not to much details on the image). "...and depend on scanning parameters (brightness, thresholds etc)...." within limits, yes. All possible paramters (in xsane) turned to many values, the results are although poor. The strong lines getting more or less intensive but weak details are totaly ignored whatever parameters are given. If those maps are the color-maps, is there an alternative, to get them uncalculated? Or is this, what you meant with "...Another way is to receive the maps directly from a frontend thus offloading the calculations...." As I've understood, the backend does the calculations, right? So, the frontend must get the uncalculated data to give them further, right? Then, should I have a switch for the backend to get the uncalculated data? And do I've a chance to get the uncalculated data? Best Regards Stefan Kempf
