Try this: uncomment both "norealcal" and "noprecal", and then send me the
resulting debug log, of a color scan.
(Make sure you get a little sleep first! :)
The two are misnomers (which I intend to change this week), so here is a bit
of explanation:
o By default, the scanner firmware performs some little automatic calibration
at the start of every scan. This takes a few seconds.
o There is an inquiry bit which indicates whether or not this calibration
can be disabled.
o The Scanmaker E6, and maybe others, has an additional calibration feature
which has its own SCSI command. This I call "real calibration", because
it actually involves the backend performing calculations on data returned
by the scanner.
and
o The backend tries to be clever, by only allowing the calibration once per
session (which shaves time off of all your scans).
So, the (purposely flawed) logic is as follows:
o If calibration can be disabled and not (norealcal),
use the 'real calibration' routine.
o Else, if calibration can be disabled and not (noprecal),
use a clever trick which does a microscan once with default calibration,
and then disables it for the rest of the session.
o Else, do not use 'real calibration', and don't disable default calibration.
Actually, in v0.12.1 of the backend, I quickly changed the behavior to
act like "norealcal | noprecal" for everything but the E6, just because
of the volume of complaints. (I originally wanted to see which scanners
used the 'realcal', so I made it the default and waited for the bug reports
to come in. That was a mistake.)
-matt m.
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com
This archive was generated by hypermail 2b29 : Mon May 21 2001 - 07:13:59 PDT