[argyllcms] Re: Strugglin' with scanin
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Thu, 04 Feb 2010 09:57:41 +1100
Hening Bettermann wrote:
./scanin -v9 -dipn -p -a -G1.0 5D2_938.tif ColorCheckerSG.cht
ColorCheckerSG_Graeme.cie
-->
./scanin: Error - CGATS file 'ColorCheckerSG_Graeme.cie' read error : Unable to
open file 'ColorCheckerSG_Graeme.cie' for reading
So you've given it a path to a file that doesn't exist.
i.e. if you do an
ls -l ColorCheckerSG_Graeme.cie
what do you get ?
./scanin110 -v9 -dipn -p -a 5D2_938_curv_exp0,5_rot_crop.tif
ColorCheckerSG_Brian.cht ColorCheckerSG_scaled_1931.cie
-->
./scanin110: Error - 4 chanel input '5D2_938_curv_exp0,5_rot_crop.tif' must be
a CMYK tiff file
It's got 4 channels but it's not CMYK.
I would have thought you actually have an RGB file, so Pascal de Bruijn's
suggestion is a
good one, maybe it has an alpha plane. Note that TIFF has several different
flavors of
alpha channels, some of which Photoshop might recognize, and some it might not
(associated and unassociated alpha). You need look to your workflow, and the
settings of whatever created the TIFF file.
(I'd suggest looking at the tiffinfo and tiffdump output of the file, but you
probably don't have that tool unless you're compiling from source).
All along, it has been very picky with the image files, for no apparent reason.
An
image that was only rotated by eye-sight (to be horizontal) was accepted,
while others
that were carefully rotated in Photoshop using the marker crosses of the chart,
were
rejected ("pattern match wasn't good enough").
Are you leaving enough of the surround ? scanin doesn't expect charts to be
cropped. As usual, looking at the diagnostic raster output often
gives a hint as to what's going on (ie. the diag.tif file resulting from -dipn).
[You can always supply the fiducial locations to fix a chart recognition
problem.]
Graeme Gill.
Other related posts:
