After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 767325 - Replace the Make Seamless plug-in
Replace the Make Seamless plug-in
Status: RESOLVED OBSOLETE
Product: GIMP
Classification: Other
Component: Tools
2.8.16
Other All
: Normal enhancement
: ---
Assigned To: GIMP Bugs
GIMP Bugs
Depends on:
Blocks:
 
 
Reported: 2016-06-07 10:17 UTC by Jo
Modified: 2018-05-24 16:25 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Result of Tile Seamless GEGL op with same source as the registry plugin (154.66 KB, image/jpeg)
2016-06-11 13:01 UTC, Jehan
Details

Description Jo 2016-06-07 10:17:50 UTC
the default "make seamless" plugin in gimp is nice but way to weak, i recommend to -add/ replace with- this plugin:
seamless_advanced_v1.16.py
Comment 1 Michael Schumacher 2016-06-07 11:32:02 UTC
And you got this from where?
Comment 2 Jo 2016-06-08 10:13:45 UTC
the link: 
http://registry.gimp.org/node/28112
Comment 3 Joao S. O. Bueno 2016-06-08 19:02:44 UTC
The given plug-in is int he public domain - if we can just factor out a second pdb call to that plug-in to preserve the current behavior, this seems a quite reasonable request.
Comment 4 Jehan 2016-06-11 12:14:09 UTC
If not mistaken, looking at our git log, this plugin seems to have disappear with commit b0645cf, and has been replaced with the gegl:tile-seamless operation (which I just tested on master and seems to work quite well!).

Since I don't think we will change a plugin in a minor 2.8 release, I propose we close this as OBSOLETE. Or am I missing something?
Comment 5 Michael Natterer 2016-06-11 12:27:44 UTC
Indeed, but who knows if we simply ported some simplistic shit to GEGL,
while that python plug-in produces much better results? Worth checking
at least...
Comment 6 Jehan 2016-06-11 13:01:07 UTC
Created attachment 329612 [details]
Result of Tile Seamless GEGL op with same source as the registry plugin

I tested both with the source image available as example in the registry page.
Attached is the result when using gegl:tile-seamless op (current master), then canvas doubled both in width and height and replicated 4 times. It works well and for this image at least, it has nothing to envy to the registry plugin IMO (cf. the other images in the link given in comment 2).

Other remarks:

- The python plugin has some deprecation warnings which would have to be fixed (but that's no big deal).

- It indeed has a lot more options than the gegl op (this one can only choose the input: selection or layer).
Unfortunately not being able to test them directly with the GEGL preview makes them a lot less interesting IMHO, because you'll have to just test, cancel, try and so on. And that's boring.

Also I am not familiar enough with tiling to really ponder how good are all these parameters. Is there anyone here with a lot of experience in image tiling to tell us what are really necessary tweaking options in your opinion?

In any case, I think — if these options make sense — it would be a lot more interesting to try and implement these in the GEGL op to have direct preview, because I found the test of the plugin extremely annoying because of the lack of preview. That's typically a kind of plugin where I would experiment for 20 min (without really understanding the parameters), then once I found OK values, I would just write them down and always use them forever from now on (hence missing the point of having parameters to tweak for every different case).
Comment 7 Jehan 2016-08-02 02:17:35 UTC
Someone was asking on IRC how to make an image seamless tileable in only a single direction. I guess that's at least one option our GEGL operation could have.
Comment 8 Jehan 2016-08-03 02:20:55 UTC
Another feedback for a useful feature: possibility to see the seamless image actually tiled with itself. Of course this one would not be on GEGL side, but on GIMP side.
I agree it would be interesting because sometimes even though the image is technically seamless with itself, it does not necessarily render well.
Comment 9 GNOME Infrastructure Team 2018-05-24 16:25:41 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to GNOME's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.gnome.org/GNOME/gimp/issues/909.