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 142987 - Scale layer property
Scale layer property
Status: RESOLVED OBSOLETE
Product: GIMP
Classification: Other
Component: General
unspecified
Other All
: Low enhancement
: ---
Assigned To: GIMP Bugs
GIMP Bugs
: destructive (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2004-05-23 03:27 UTC by Calvin Spealman
Modified: 2018-05-24 11:05 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Calvin Spealman 2004-05-23 03:27:26 UTC
Layers should have an option to give them a scale, where they will be placed on 
the final image as a different size than they actually are. Thus, original 
detail could be retained in the ear-ring I draw on my character, and I can go 
back and make minor changes easily.
Comment 1 Sven Neumann 2004-05-24 07:12:28 UTC
This is a rather weird concept for an image manipulation program and I have
never seen another app doing this. I don't think we will add this feature.
Comment 2 Henrik Brix Andersen 2004-05-24 11:37:46 UTC
Sounds like a feature from an SVG editor. Not suitable for a raster based image
manipulation program, imo.
Comment 3 Sven Neumann 2004-05-24 13:55:14 UTC
Well, it's not completely insane and it's something that is supposed to be
doable with the framework provided by GEGL. That doesn't necessarily mean that
we will do it but we could keep it as an enhancement request. If only it had a
more meaningful summary...
Comment 4 weskaggs 2004-05-24 17:48:14 UTC
This fits in with the general idea of maintaining a transformation matrix for
each layer, so that one can build up a set of linear transformations (scale,
rotate, perspective, etc.) without losing detail at each step.
Comment 5 Calvin Spealman 2004-05-25 02:15:39 UTC
Yes, OK, the summery was a little vague. 
 
First then, weskaggs' mention of a full transformation mechanism is much better 
actually. The first step of this could actually be a transformation matrix, 
with more user-friendly interfaces for Scale, Rotate, etc. added layer, simply 
auto-generating the matrix needed for the requested transformation. 
 
I often work with many components to my image, and often I work in high detail 
and scale things down, but I hate to loose the detail in the case that I need 
to make more changes. This is why I want the feature, and rotation of layers 
would be wonderful as well. 
 
Henrik, yes, this does sound like a feature of an SVG, but that doesn't mean it 
isn't useful or wouldn't be a welcome feature. I very often require both vector 
and raster related tools for a single image, so this just seems natural to me 
and completely unnatural to separate them. 
Comment 6 Sven Neumann 2004-05-25 08:48:55 UTC
Why on earth do you scale things down in the first place? If you scale things
down in a pixel-based image manipulation program you have to expect that you
loose detail. That's the concept of pixel-based image manipulation.

Anyway, all these concepts (also the other enhancement requests you made) are in
the GEGL design so rest assured we've already put a lot of thought into this
direction.
Comment 7 Calvin Spealman 2004-05-25 16:35:50 UTC
OK, so, I'm confused. You're saying its a pointless, useless feature and there  
is already work towards it. So, is this a possible happening or not?  
  
And, I scale them down because somethings are just easier to work with at a  
larger scale, but they may need to be smaller relative to the rest of the  
image. This also goes along with my bug #143138 
Comment 8 Sven Neumann 2004-05-25 16:45:28 UTC
GEGL is supposed to support such things in it's rendering graph but that doesn't
necessarily mean that GIMP is going to make use of it. It might make sense to
discuss it later when the framework is in place. At the moment it's far from
being possible.
Comment 9 samuel 2006-04-04 17:07:22 UTC
I would like to second the motion..
One example of where this can be useful is where you're trying to fit a layer from one image into another .... lots of tweaking of location, rotation, scaling, shearing, etc.  

With the standard pixel-only translations, you lose information at each step.  If it's done as a matrix operation, you pretty much end up doing the same amount of work (just add in the (cheap) cost of the matrix math), but the resulting quality is far higher.

Once you start doing other changes (e.g. painting on the result), you might have to fix the image in pixels, but the end result quality for many operation sequences would be a world apart.
It might also make it a good bit cheaper for keeping track of history as well... All you'd have to store would be the initial image, the trasformaiton matricies and the final image (for speed).
You might, however, want a checkbox to turn the feature off, as some people might actually <i>like</i> the distortive effect of multiple pixelated translations.
Comment 10 samuel 2006-04-04 17:16:18 UTC
The one case where a pixelated result can be most wanted is when you're shrinking down a huge layer to a small graphic -- doing fruther  transformations based on the huge original can be expensive, so fixing after the initial scale down can be important.
Comment 11 Michael Natterer 2012-10-17 09:50:43 UTC
*** Bug 686275 has been marked as a duplicate of this bug. ***
Comment 12 bitbonk 2012-10-17 11:21:00 UTC
Please see Bug 686275 for another view on this subject.
Comment 13 GNOME Infrastructure Team 2018-05-24 11:05:39 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/78.