Graeme,
Thanks for the quick reply. I'm glad you were able to reproduce and a fix is
coming.
I realized there is still something wrong with my approach as even though with
other target captures I was able to complete the ICC creation, I'm getting
errors of a very high magnitude (max ~51, avg ~28, RMS ~30). Per your advice I
will try and improve the target image input, but it seems my calibration data
is also off.
Cheers,
Nic Nilov.
On Apr 28, 2022, at 17:29, Graeme Gill <graeme@xxxxxxxxxxxxx> wrote:
Nic Nilov wrote:
Hi,
I was trying to create an ICC camera profile from a CC24 passport and am
observing a weird behavior
where scanin output gets stuck and the resulting .ti3 file keeps growing
endlessly. It reached 12Gb
when I terminated the program. The diag.tiff meanwhile is only 8 bytes in
size. I'm doing this all
under macOS Big Sur.
thanks for the files, so as to be able to try and reproduce.
I couldn't reproduce the file size problem on my MSWin machine,
but can on my OS X machine. So it seems like a slight difference
in floating point handling (log10(inf) result) is to blame for that.
That's something I'll fix in the next release.
Your underlying problem though, which you would see in the diag. image
if scanin would finish, is that chart recognition fails because the
image has too much extraneous around it. If you roughly crop it
(see attached for what I mean), it recognizes correctly, and will finish.
Cheers,
Graeme Gill.
<chart-raw.crop.jpg>