Regression tests: HP Scanjet 5400C
Regression testing was performed on the HP ScanJet 5400C. The following was discovered:
USABILITY: Default values for options -l and -t are not sane and cause the top of a properly positioned document to be cropped. Default values are "-l 5 -t 52". "-l 0 -t 0" give better results in scanimage.USABILITY: Maximum actual value for -y option is 300. 248 is documented maximum in scanimage help output.- QUALITY: 1200dpi scanning shows noticeable colour banding. Other supported resolutions of 75, 150, 300 and 600 are fine.
- FUNCTIONALITY: Documentation says that the scanner supports 2400dpi optically. Only 1200dpi is supported by the backend currently.
BUG: Specifying an unsupported resolution with --resolution to scanimage results in a stretched and improperly offset image. Example 100_stretched.jpeg attached, which was generated using "scanimage -d hp5400:libusb:012:002 --resolution 100 -l 0 -t 0 -x 215 -y 300" on my machine. Backend should perhaps select an appropriate supported resolution or perform custom scaling.FUNCTIONALITY: Buttons do not seem to be supported. The scanner has an impressive selection of buttons and also has a 2 digit LCD display to support a software "copy" function. In addition, there are buttons for selecting "greyscale" which is not supported by the backend (perhaps the scanner only supports colour internally). This might be worth investigating in the future.FUNCTIONALITY: Add support for LCD copy counter as a sensor.- FUNCTIONALITY: Check Windows driver for native support for grey and lineart modes.
The above tests were performed using the USB interface. Scanner also supports parallel interface, but this was not tested as yet.
Version is 1.0.29 candidate from Rolf's PPA.
Edited by Ralph Little