Banding on prints using custom profiles

Robin Simmons <[email protected]> Fri, 19 Aug 2022 17:45:57 +0100
Newsgroups gmane.linux.printing.gimp-print.devel
Organization Lenken International
Message-ID <[email protected]>
Good afternoon,

I am getting banding in areas of gradual luminance gradient when printing with Gutenprint, as described below.

OpenSUSE Tumbleweed KDE, plasma version 5.25.4
Kernel 5.19.1-1-default (64 bit)
Gutenprint version 5.2.15
CUPS version 2.4.2
Epson Surecolor P600 printer

In order to obtain accurate colours with the printer, I have custom ICC profiles created by the suppliers of the photo papers I use. The suppliers' templates (patch files) are printed from Darktable v4.0 in the normal way by turning off all colour management in Gutenprint and Darktable. The resultant ICC profiles are very good for colour accuracy. However, where a print has an area of gradually changing luminance, visible banding occurs. This is most obvious in desaturated regions; it also happens in more saturated areas but is much less marked.

Darktable's soft proofing displays exactly the same banding. I have created two simple monochrome images with linear luminance gradients and taken screenshots of the results in Darktable with soft proofing on and off respectively. The results are attached.

I have so far had 5 custom profiles created from three different suppliers. All of them show the banding to a greater or lesser extent. The same image files printed on Windows computers and the same model of printer, using the Epson driver and custom profiles for that combination do not show any banding. The conclusion must be that the custom profiles created via Darktable and Gutenprint/CUPS are somehow faulty, i.e. the patch files have not been created correctly.

I also occasionally use Qimage Ultimate v2023.100, a Windows program running under WINE. The ICC profiles created via Darktable produce prints (and soft proofs) in Qimage which are virtually indistinguishable from those in Darktable, i.e. the ICC profiles are fully interchangeable between Darktable and Qimage. Furthermore, if I print the colour patches from Qimage instead of from Darktable, the results are exactly the same. This leads me to believe that the problem cannot lie with either Darktable or Qimage, since the colour patches are printed independently and give identical results.

I cannot compare with images printed directly from GIMP as the ICC profiles created with Darktable or with Qimage are not interchangeable with GIMP - the colours are no longer accurate. I do not know why.

Perhaps I should also mention that prior to using Gutenprint I worked with TurboPrint and results from this show no banding at all.

The only common factor is therefore Gutenprint/CUPS, hence I can only conclude that this is very likely to be a Gutenprint bug.

Kind regards,

Robin Simmons

_______________________________________________
Gimp-print-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gimp-print-devel
Screenshot_soft proof OFF.png (image/png, 144.5 KB) - not displayed
Screenshot_soft proof ON.png (image/png, 257.2 KB) - not displayed