RSS/Atom feed Twitter
Site is read-only, email is disabled

c2g in Gimp 2.10.12 not performing as in 2.8

This discussion is connected to the gimp-user-list.gnome.org mailing list which is provided by the GIMP developers and not related to gimpusers.com.

This is a read-only list on gimpusers.com so this discussion thread is read-only, too.

8 of 8 messages available
Toggle history

Please log in to manage your subscriptions.

c2g in Gimp 2.10.12 not performing as in 2.8 pistikem 07 Aug 01:40
  c2g in Gimp 2.10.12 not performing as in 2.8 rich404 08 Aug 08:12
   c2g in Gimp 2.10.12 not performing as in 2.8 pistikem 09 Aug 16:08
    c2g in Gimp 2.10.12 not performing as in 2.8 rich404 09 Aug 17:36
     c2g in Gimp 2.10.12 not performing as in 2.8 pistikem 10 Aug 00:52
      c2g in Gimp 2.10.12 not performing as in 2.8 rich404 10 Aug 07:36
       c2g in Gimp 2.10.12 not performing as in 2.8 pistikem 10 Aug 14:31
        c2g in Gimp 2.10.12 not performing as in 2.8 pistikem 29 Nov 01:38
2019-08-07 01:40:17 UTC (over 4 years ago)
postings
5

c2g in Gimp 2.10.12 not performing as in 2.8

GEGL's c2g function in Gimp 2.10.12 performs very differently as compared to v2.8. the outcome is much brighter, and the deep blacks are gone that were so helpful to simulate a film-based appearance. i played with the settings for radius, samples and iterations but nothing comes close to the former appearance. clicking the 'shadow' button didn't improve the result. so, the current stage of c2g is pretty useless for me - unless i do something substantially wrong.

can somebody help me how to get c2g behave in Gimp 2.10 as it has in 2.8 ?

pistikem (via www.gimpusers.com/forums)
rich404
2019-08-08 08:12:32 UTC (over 4 years ago)

c2g in Gimp 2.10.12 not performing as in 2.8

GEGL's c2g function in Gimp 2.10.12 performs very differently as compared to
v2.8. the outcome is much brighter, and the deep blacks are gone that were so
helpful to simulate a film-based appearance. i played with the settings for
radius, samples and iterations but nothing comes close to the former appearance.
clicking the 'shadow' button didn't improve the result. so, the current stage
of c2g is pretty useless for me - unless i do something substantially wrong.

can somebody help me how to get c2g behave in Gimp 2.10 as it has in 2.8 ?

I do not think it is possible. I have looked at this previously so not just a passing thought.

The change log for GEGL is here: http://gegl.org/NEWS.html Down in GEGL-0.3.30 this came in

quote - c2g and stress have gained the ability to toggle the influence of the shadow neighborhood or not, the new default is to not improve shadow detail; yielding a bit more natural renditions - unquote

Gimp 2.10 has moved on since then to GEGL 0.4 Using GEGL 0.4 command line with c2g produces the same effect as Gimp 2.10 so no solution there.

I also have a 'buntu 16.04 with GEGL 0.3.25 Does that work? No. The command line produces a slightly different version but still not the same as Gimp 2.8, still too light.

Best bet. Depends on OS but run a portable Gimp 2.8 as well as Gimp 2.10, otherwise find a combination of other filters that produce that grainy effect.

rich404 (via www.gimpusers.com/forums)
2019-08-09 16:08:01 UTC (over 4 years ago)
postings
5

c2g in Gimp 2.10.12 not performing as in 2.8

you are absolutely right, rich404, i observed the same changes in c2g's rendering pattern with the version changes in GEGL and GIMP, via GUI or CLI. now, this is very unfortunate. the c2g algorithm is to my understanding truly unique and very unusual and special, as it adjust pixel contrasts in an 'organic or neurological or physiological' way, like the original author stated that he was influenced by neuro-physiological evidence. i feel this precious piece of code would deserve to be properly maintained within GIMP and GEGL, or be made publicly available, perhaps ported to Python3. i am not aware of any other (photo) image processing algorithm that 'intelligent' to combine data-driven(!) contrast and color information from a pixel neighbourhood across three channels for the purpose of reducing RGB to one layer. i am truly desperate for getting a new c2g version to work with the appropriate set of adjustable run parameters. what about a python3 port ? i still have a few linux boxes with gimp 2.8 running but i don't see that as a good way out of this problem, perhaps as a temporary solution only...

pistikem (via www.gimpusers.com/forums)
rich404
2019-08-09 17:36:21 UTC (over 4 years ago)

c2g in Gimp 2.10.12 not performing as in 2.8

you are absolutely right, rich404, i observed the same changes in c2g's
rendering pattern with the version changes in GEGL and GIMP, via GUI or CLI.
now, this is very unfortunate. the c2g algorithm is to my understanding truly
unique and very unusual and special, as it adjust pixel contrasts in an 'organic
or neurological or physiological' way, like the original author stated that he
was influenced by neuro-physiological evidence. i feel this precious piece of code would deserve to be properly maintained
within GIMP and GEGL, or be made publicly available, perhaps ported to Python3.
i am not aware of any other (photo) image processing algorithm that 'intelligent' to combine data-driven(!) contrast and color information from a
pixel neighbourhood across three channels for the purpose of reducing RGB to one
layer. i am truly desperate for getting a new c2g version to work with the
appropriate set of adjustable run parameters. what about a python3 port ?
i still have a few linux boxes with gimp 2.8 running but i don't see that as a
good way out of this problem, perhaps as a temporary solution only...

FWIW - You would have to ask the developers (Maybe you already have?)

The closest I can get is the gmic_gimp plugin and the black-and-white filter

but....

It is not consistent. A darker image requires slightly different setting for contrast and gamma than a lighter image and if you do not know what a Gimp 2.8 c2g looks like, then it is a matter of judgement.

screenshot examples have the Gimp 2.8 c2g as an inset.

The only advantage with gmic_gimp is

1. Your can save settings to recall them at a later date.

2. It is scriptable.

best of luck with this one.

Attachments: * https://www.gimpusers.com/system/attachments/1238/original/01-c2g.jpg * https://www.gimpusers.com/system/attachments/1239/original/02-c2g.jpg

rich404 (via www.gimpusers.com/forums)
2019-08-10 00:52:38 UTC (over 4 years ago)
postings
5

c2g in Gimp 2.10.12 not performing as in 2.8

thank you, rich404, for this information ! i didn't know about this gmic_gimp thing - good to know !

now, short question : the large examples you show are processed within gmic with some manual desaturating techniques, correct ? you are not using any c2g algorithm in gmic as we used in gimp 2.8, right ? one definitely sees a wealth of more information in the small inserts or insets as compared to the big pictures !

i am new in this forum (and also in gimp-forum and gimpforums). how can i contact the developers as you suggested ?

pistikem (via www.gimpusers.com/forums)
rich404
2019-08-10 07:36:57 UTC (over 4 years ago)

c2g in Gimp 2.10.12 not performing as in 2.8

thank you, rich404, for this information ! i didn't know about this gmic_gimp
thing - good to know !

now, short question : the large examples you show are processed within gmic with
some manual desaturating techniques, correct ? you are not using any c2g
algorithm in gmic as we used in gimp 2.8, right ? one definitely sees a wealth
of more information in the small inserts or insets as compared to the big
pictures !

i am new in this forum (and also in gimp-forum and gimpforums). how can i
contact the developers as you suggested ?

The gmic_gimp plugin, www.gmic.eu does not use c2g (or GEGL) in anyway. The larger images are the gmic dialogue preview that gives a live (not always exact) view of the parameter changes, and there are a lot of parameters in the Black & White filter. All you can do is try it. Maybe not as good as c2g but....

For better information ask on the pixls.us forum. https://discuss.pixls.us/ where the gmic developer resides. It is largely a photographic forum and you will get better advice than I can give.

The developers use https://gitlab.gnome.org/GNOME/gimp/issues/ as a center for bugs reports and enhancement requests.

rich404 (via www.gimpusers.com/forums)
2019-08-10 14:31:17 UTC (over 4 years ago)
postings
5

c2g in Gimp 2.10.12 not performing as in 2.8

The gmic_gimp plugin, www.gmic.eu does not use c2g (or GEGL) in anyway. The
larger images are the gmic dialogue preview that gives a live (not always exact)
view of the parameter changes, and there are a lot of parameters in the Black &
White filter. All you can do is try it. Maybe not as good as c2g but....

For better information ask on the pixls.us forum. https://discuss.pixls.us/
where the gmic developer resides. It is largely a photographic forum and you
will get better advice than I can give.

The developers use https://gitlab.gnome.org/GNOME/gimp/issues/ as a center for
bugs reports and enhancement requests.

thank you for the links - i will contact the GIMP/GEGL developers.

regarding the non-manual 'automatic' aspect about c2g : i love the fact of not having to influence my picture manually (beside the run usual parameters but they don't influence local contrast nuances that the algorithm auto-magicalty works out). it's like with the previously used b&w filters (green, yellow, orange, blue etc) for b&w films : you did your shot, and that's it - there was no way to adjust the local fine-grained contrast distribution, just globally and/or manually(!).

pistikem (via www.gimpusers.com/forums)
2019-11-29 01:38:03 UTC (over 4 years ago)
postings
5

c2g in Gimp 2.10.12 not performing as in 2.8

i just installed the newest GIMP/GEGL pair 2.10.14/0.4.18 - and i am very pleased with the results : much faster, using all available cores, and then the output : c2g renders much more consistently ! no large white patches and such. though i still have to compare to pics generated with GIMP 2.6.

pistikem (via www.gimpusers.com/forums)