%PDF- %PDF-
Direktori : /usr/share/doc/libsane-common/gt68xx/ |
Current File : //usr/share/doc/libsane-common/gt68xx/gt68xx.TODO |
gt68xx.TODO -*-text-*- General: -------- - some sort of fixed lookup-table may be necessary to compensate for the non-linear CCD/CIS behaviour. A gamma value is not good enough. Or maybe line calibration can be used for that? - Some Plustek scanners are advertised with e.g. 1200x2400 but the maximum vertical resolution seems to be limited to the horizontal one. Check if that's really the case. - Keep a log of all gain/offset combinations already tested and break if one is repeated to avoid endless loops. - Print commands that are send to the scanner in clear text. - Try to implement non-blocking USB bulk reads, check libusb CVS. CCD: ---- Mustek BearPaw 2400 TA Plus: - some lines at the beginning of each image have wrong colors (linedistance?) This seems to happen with other scanners, too. - 600 dpi no backtrack, 1200, 2400 dpi backtrack (gray), check with windows for color - 600 dpi 12 bit seems to work in pixel mode only (but long lines are too big!) Mustek BearPaw 1200 TA: - similar color problem as 2400 TA Plus, doesn't always happen? Mustek BearPaw 2448 TA Plus: - all modes are slow (massive backtracking). If connected to a USB2 port, the backtracking ich much less despite this scanner can't do USB2. Strange. - Sometimes there is a USB protocol error when cancelling the scan? - top borders move with resolution (happened in the early days with 2400 TA!) - gray scans have different brightness - scan motor is bumpy and freezes sometimes, looks like acceleration is too big (firmware problem?) - vertical stripes - TA settings are wrong - no backtrack >= 600 dpi (also gray!) - don't move before and after scan - check jpeg output, maybe this can be used for faster scans Plustek 1248U: - Linedistance seems to have problems in the first few and the last lines. - Gain and offset values are quite high. Optimum values are not reached. Problem with coarse calib? Better algorithm needed. - Images in color mode sometimes look too redish (result of above problem). - quite slow when full width is selected. Lots of backtracking. - Check if pixel mode is possible. - Check again 600x1200 dpi. Plustek U16B: - 600 dpi 16 bit is limited to about half the size of the scan area because only pixel mode works and 16k is the maximum buffer size. - After scanning, a run of sane-find-scanner leads to a protocol error and scanning does no longer work. Genius ColorPage Vivid 1200XE - 16 bit per color is advertized but this mode doesn't seem to work (only 12 bit)? - 1200x1200 and 1200x2400 dpi doesn't seem to work despite being advertized - Y-positioning varies slightly with resolution (some mm) CIS: ---- - It may be necessary to do an additional exposure calibration before the offset/gain calibration to compensate for different intensities of the colored leds. Especially for the Mustek 2400 CU Plus. - Changing exposure doesn't seem to have any effect during calibration sometimes while it has an effect on the final image? Mustek BearPaw 1200 CU - ok Mustek BearPaw 1200 CU Plus (0x021b) - ok, but similar exposure problems as 2400 CU Plus Mustek BearPaw 2400 CU Plus - ok, but exposure setting may need to be done automatically if the intensity of the colors varies for different scanners. - calibration/exposure has sometimes sharp jumps in intensity. Reason is unknown. Mustek ScanExpress A3 USB - still slight vert. stripes - why is line distance setup necessary? - line distance varies with resolution (0,5,5 at 300 dpi, 0, 6, 7 with 150 dpi) - exposure doesn't seem to have any effect Mustek ScanExpress 1248 UB ok Plustek OpticSlim 2400 - only up to 1200 dpi vertical works - only 8 and 12 bit work, 16 doesn't (check Windows) - exposure is extremely high and seems to have similar problems as with other CIS scanners (reaches limit also) Mustek ScanExpress 2400 USB - make it work (nobody seems to have this device) Visioneer OneTouch 7300 - 16 bit doesn't seem to work despite being advertized