Gimp-Forum.net - Tutorials and tips

Batcher and simple animation.

Sat, 18 Apr 2026 11:31:34 +0000

Looks like nobody is using Gimp anymore, it is all AI, however Gimp 3 can come in useful at times.

This is using the Gimp 3 plugin Batcher https://kamilburda.github.io/batcher/

(1) Removing a plain color background from an animation. It might be something downloaded from one of those animation websites. If it is possible using Color -> Color-to-Alpha then in Batcher these Gimp operations gimp-layer-add-alpha and color-to-alpha

  batcher-ani-01.jpg (Size: 87.94 KB / Downloads: 41)

(2) Distributing the layers across the canvas using the Gimp 3 Align tool and adding a background with Batcher.

  batcher-ani-02.jpg (Size: 100.9 KB / Downloads: 37)

Which gives this:

  walking.gif (Size: 289.13 KB / Downloads: 215)

Tool Presets

Sun, 05 Apr 2026 13:30:47 +0000

Most of the tools have presets that you can save. I only found out this today! You can save a desired preset's name for most tools that have a tool option etc and also edit it here. These show in the 'blue bucket icon with tools' for the different tools. If you click on the tool icon you can see all the different tool presets etc come up.

Text preset blue bucket.png (Size: 41.6 KB / Downloads: 47)

Setting a default directory to save files

Sun, 07 Dec 2025 17:30:06 +0000

You will have noticed that Gimp is hell-bent into saving everything by default in your Documents folder. It's no completely Gimp faults, it is the behavior of the GTK file dialog thayt Gimp uses.

The simple but incomplete solution is to add your favorite folder(s) to the Bookmarks panel on the left.

But you can also make Gimp use any directory by making it think it is your Documents folder.

The FreeDesktop standard defines several directories: DESKTOP, DOWNLOAD, TEMPLATES, PUBLICSHARE, DOCUMENTS, MUSIC, PICTURES, VIDEOS
A configuration file called user-dirs.dirs can be used to define what actual directory is used for these symbolic directories. This is typically used to give language-dependent names.
user-dirs.dirs is normally in ~/.config but you can tell the desktop code to use a different one by setting XDG_CONFIG_HOME environment variable to a directory that contains another version of user-dirs.dirs
You update the file with the xdg-user-dirs-update command, and you can check it use the xdg-user-dir command. These two commands will of course check the XDG_CONFIG_HOME environment variable.

So, all that is needed is to prepare a user-dirs.dirs pointing to the required directory, and set the XDG_CONFIG_HOME environment variable to point to the alternate user-dirs.dirs directory in a script before calling Gimp.

So for instance, setting your destination directory to ~/tmp:

Do once:

Code:

export XDG_CONFIG_HOME=~/tmp/               # directory for user-dirs file, and your Gimp files

xdg-user-dirs-update --set DOCUMENTS ~/tmp  # Creates/updates the user-dirs file

Before calling Gimp (so create/modify script):

Code:

export XDG_CONFIG_HOME=~/tmp/  # tell FreeDesktop where to look for directories definition

gimp                           # start Gimp (can also be "exec gimp" to get rid of one level of shell)

The exemple above keeps the user-dirs.dir file and the Gimp files in the same directory but using different directories is left as an exercise for the reader.

If any Windows or OSX users are reading this, I wonder if this applies in other OS as well. If you are missing the xdg-user-dirs-update command, user-dirs.dirs is a plain text file and it looks like this:

# This file is written by xdg-user-dirs-update
# If you want to change or add directories, just edit the line you're
# interested in. All local changes will be retained on the next run.
# Format is XDG_xxx_DIR="$HOME/yyy", where yyy is a shell-escaped
# homedir-relative path, or XDG_xxx_DIR="/yyy", where /yyy is an
# absolute path. No other format is supported.
#
XDG_DOCUMENTS_DIR="$HOME/tmp"

Astrophotography - Creating masks

Fri, 17 Oct 2025 20:18:55 +0000

Can anyone guide me to a video, link or post that explains how to edit and create mask for astrophotography images?

Windows Gimp 3.0 symbol entry

Mon, 09 Jun 2025 08:11:30 +0000

For adding a symbol to text in Windows Gimp 3.0 the previous unicode entry does not work (it is more-or-less ok with linux).

Two ways you might add a symbol

(1) Copy and paste into the Gimp text window, typically using a character map.

  01-char-map.jpg (Size: 75.12 KB / Downloads: 605)

(2) One of the Gimp developers came up with this for Windows (must be in the MS documentation somewhere), putting a symbol straight into Gimp text.

The Windows key + . (the period / full stop key) brings up this.

  02- symbols.jpg (Size: 138.93 KB / Downloads: 308)

There is a section for symbols, which narrows down selection a bit. Clicking on a symbol puts it straight into the Gimp text window, using the current font.

  03-symbols.jpg (Size: 147.24 KB / Downloads: 335)

Gimp Color Pixel Degrayscaler Tool

Wed, 09 Apr 2025 23:08:29 +0000

ghsv1.xcf (Size: 118.57 KB / Downloads: 349)
I made a color pixel degrayscaler tool for Gimp in 3.0.2/. It finds basic colors (R,G,B,C,M,Y) in hue,saturation & value of pixels prior to grayscale conversion
by looking at value of grayscale pixels & matching them to non-grayscaled pixels in above mentioned colors so they can checked for RGB & LCH channel values.

Gimp Color Pixel Degrayscaler Tool

What It Does

Finds basic colors (R,G,B,C,M,Y) in hue,saturation & value of pixels prior to grayscale conversion/full
color desaturation by looking at value of grayscale pixels & matching them to non-grayscaled/non-desaturated
pixels in above mentioned colorsfor checking the RGB & LCH channel values,plus how to colorize grayscale images using this tool.

File Format

Image is in XCF format.

How To Use

Download or save file,then open in Gimp as XCF under the ghsv1.xcf file name (grayscale,hue,saturation,value for short)
After opening in Gimp,make two duplicate images & close original. Select second duplicate
image,click Image & convert to grayscale,copy grayscale duplicate & paste it
as layer in place to first duplicate image. Zoom to 1600%,scroll fully up
then fully left with eye on background copy layer off. Use horizontal
& vertical position scales in image window (top & left edges) & in bottom left of image window,
pixel position numbers with px selected. Image should appear as shown as follows:
Beginning at 0 vertical position,grayscale value stripe,(gray value of 0 to 255)followed by two hue stripes,
RGBCMY stripes for saturation,RGBCMY stripes for value. Color picker tool should be set
to example,foreground color with use info window checked. Use select by color &
click on position (horizontal image scale) of grayscale stripe value pixel matching
the position to locate value of pixel in your grayscale image as shown after
selecting background copy as highlighted. Make certain that select by color tool
is set to composite & threshold set to 0.0 with other options such as antialiasing
for example unchecked before doing above. Scroll right until highlighted selection areas
in stripes are seen.(ex.91,3 referring to horizontal & vertical positions
of selected areas) & select background layer if you're colorizing your grayscale image
after converting it to RGB,then use color picker in the mentioned highlighted selection areas.

Colorizing with Gimp Color Pixel Degrayscaler Tool

I devised a way to colorize grayscale images after duplicating & converting them to RGB.
It uses a checkerboard pattern in a 3 by 3 pixel image. Use File,New & set the vertical
& horizontal number of pixels to 3. Zoom to 1600% & scroll to upper left corner.
The three rows in the checkerboard pattern is follows: top row: RGB,
middle row:GBR,bottom row:BRG. Don't forget to set pencil/brush tool to pencil & pixel size to 1 for each of the nine pixels in the pattern.
You can copy the pixel colors from the top row with the color picker,then pencil, & use it with the pencil
for the middle & bottom rows of pixels. Next,set the ellipse/rectangle tool to rectangle & draw a 3 by 3 pixel
square around the pattern. Copy it then use the pattern image in edit to transfer it to the highlighted selection
areas in the RGB converted duplicate of the grayscale image you're coloring. To begin coloring the mentioned grayscale image,
zoom to 1600% as before & again scroll to the upper left corner & select the first pixel with select by color
& use the pattern image in edit as above. This is for Gimp 3 but can be used in the 2.10 versions.
To separately change the red,green & blue pixels in the pattern,for example, click with color select on the red
pixels in the mentioned checkerboard using "Use select by color & click on position(horizontal image scale)..."
as above in the tool image & repeat the procedure for the green & blue pixels in the checkerboard. (forgot to mention
you need to use bucket fill set on fill whole image/selection when separately changing the red,green & blue pixel colors
in the checkerboard.

I hope someone can speed up this coloring procedure by automating what parts of it can be automated & doing things manually
for the rest of it. This is not an AI tool. For example I uploaded a solid red image converted to grayscale & some AI's,even online don't properly convert the image
to the correct color for R255,G0,B0 values.

Converting the API doc to Python

Thu, 03 Apr 2025 09:11:25 +0000

A little bit of C.

In C you have two kinds of data types.

The "native" ones: int and floats of various lengths, and pointers (memory addresses)
The "complex" ones (anything else, "structs" or arrays)

When you call a function:

native types are passed "by value": their value is copied in the call stack
complex types are passed "by reference": a pointer to them is copied in the calls stack. The fact that this is a pointer is indicated by *
the function can only "return" a single native type
the function can also update memory locations to which a pointer has been passed, this is how extra return values can be used

In addition:

Except for constructors, in the C description the first argument of a class method is an object of the class
When used with Python object syntax, this first argument becomes the object on which the method is called (with of course, one less argument).
So for instance the method described in the doc as do_something(SomeClass, arg1, arg2) is called in Python as instance_of_SomeClass.do_something(arg1,arg2)

For instance with this "C" prototype from the Gimp.Drawable description:

Code:

gboolean

gimp_drawable_get_offsets (

  GimpDrawable* drawable,

  gint* offset_x,

  gint* offset_y

)

Where GimpDrawable is a "complex" type, and gboolean and gint are "native" type (just names for machine-size integers).

* the function returns a boolean
* the first argument is a pointer/reference to GimpDrawable. Since GimpDrawable is a complex type it is a simple "input" argument.
* as stated above, since this is a method on Gimp.Drawable in python it will be called on the object which is the first argument.
* offset_x and offset_y are references to native types, and so are outputs to the function. In Python your inputs are immutable, so this is converted to extra return values, and the method returns a tuple.

So this would be called as:

boolean_success, offset_x,offset_y= get_offsets(layer: Gimp.Drawable)

But since the first argument is of the type described, in Python this method call on a Gimp.Drawable object, so it is passed implicitly when calling the method.

boolean_success, offset_x,offset_y= layer.get_offsets()

The returned boolean here is a bit artificial. This is probably because the method has to use output arguments to return X and Y, so this is the only way to check that the memory locations passed for the offsets have been updated. This can still be ignored in the vast majority of cases by writing:

_, offset_x,offset_y= layer.get_offsets()

where _ is the Python convention to indicate that you don't really care about this value.

Slightly harder:

Code:

GimpLayer*

gimp_image_get_layer_by_name (

  GimpImage* image,

  const gchar* name

)

The function takes a reference to a GimpImage, so in this case this is the implicit reference to the object on which the method is called.
The function takes a reference to gchar. This isn't an output for a single char, this is a reference to an array of characterss (arrays and pointer are equivalent in C, an array is defined by the address of its first element). And an array of gchar is... a string. What also indicates that this is an input-only argument is the const modifier.
The call returns a pointer to a GimpLayer, so in Python, a plain Gimp.Layer

So this defines a Gimp.Image method that would be type-hinted as:

def get_layer_by_name(name: str) -> Gimp.Layer

Raising the bar:

Code:

GimpLayer **

gimp_image_get_selected_layers (

  GimpImage* image

)

Here the call return a pointer to a pointer to GimpLayer... this is really a pointer to an array of GimpLayer, so for Python just a list of Gimp.Layer. So this is just:

def get_selected_layers() -> list[Gimp.Layer]

Gotcha:

Note that some calls are defined as functions and not as object methods. For instance:

Code:

gboolean

gimp_file_save (

  GimpRunMode run_mode,

  GimpImage* image,

  GFile* file,

  GimpExportOptions* options

)

This is a plain function in the Gimp namespace that takes 4 explicit arguments.

transform tools in action

Tue, 18 Mar 2025 09:57:47 +0000

How to use transform tools. Here PS but possible in gimp.

https://www.youtube.com/watch?v=HAL_W6DGTUs

Non-destructive editing in Gimp3

Tue, 11 Mar 2025 13:20:22 +0000

Gimp V3 introduces "non destructive editing" (aka NDE).

When you apply filters to layers, instead of altering the layer pixels immediately (as done in V2), the filter can be added to a stack of filters which is applied on the fly when Gimp uses the layer contents. When this is the case there is a "fx" icon next to the layer, and if you click on it you get the stack of current filters, that you can further edit.

FilterStack.jpg (Size: 97.23 KB / Downloads: 517)

This can be useful. For instance, in the image above, you can tweak the denoising and not lose the Exposure correction or the sharpening (and given that the trick is usually finding the right blend of denoising and sharpening this makes the workflow a lot more efficient). You can also decide that you will do Exposure before everything else by moving it down in the stack.

Whether the filter is applied immediately or added to the stack is set by a "Merge filter" check box in the dialog:

MergeButton.png (Size: 33.54 KB / Downloads: 3374)

When the box is checked, the filter is applied immediately. This box only shows when you start the filter from the regular menus. If you edit the filter in the filter stack it is implicitly considered as "unmerged".

If you apply the filter immediately it is applied before the filters in the stack.

LUT's - lookup tables : load, or make, and use

Wed, 25 Sep 2024 17:52:24 +0000

It is easy to correct or change the colours from a set of pictures with LUT's.
What are LUT's or 'lookup tables', some information: https://nofilmschool.com/what-is-a-LUT
Free LUT's, one of the many free sites to download LUT's : https://freshluts.com (register and download or post)
Can one make LUT's? Yes, with G'Mic-Qt.
- load a picture and make a layer copy
- make a new layer and correct or change the colors in the layer copy
- use G'Mic-Qt / Colors / CLUT from After - Before Layers
- set the 'Output Mode': Save CLUT as cube or pngFile (use cube)
- fill in 'Output Folder' (make a folder where you put all your 'cube' files)
- fill in 'Output Name' (a name that clarified what the LUT does - better for later use)
- click on the 'OK' button
The LUT can now be used for other pictures.
- load a picture
- G'Mic-Qt / Colors / Apply External CLUT
- choose the 'HaldCLUT Filename (folder where you saved the cube file)
- done
Example (I used a LUT I made before from another corrected photo)

  Roland Van Campenhout.jpg (Size: 129.64 KB / Downloads: 698)
The Lut I used (cube file compressed):
  dran noa lee02.7z (Size: 665 KB / Downloads: 542)

G'Mic-Qt Target Color Spot

Thu, 05 Sep 2024 15:56:29 +0000

I jus discovered G'Mic-Qt Target Color Spot. Nice to play with colours.

GMicQt Target Color Spot s.jpg (Size: 95.95 KB / Downloads: 419)

What is the gamma encoding and why are my color computations wrong?

Tue, 16 Jul 2024 19:25:50 +0000

So, you tried to predict or emulate Gimp's behavior with math and failed. And you wonder why. You have come to the right place, take a seat and fasten you seat belt. Also, make sure you check all images at their actual size (click on them to see them in a separate tab, and make sure your browser is exactly 100% zoom) otherwise some examples could seem to demonstrate exactly the opposite of what is stated in the text.

The root cause is that our eyes are not linear devices. If you double the power output (for instance, switch on a second light), you don't perceive the room twice as bright. Somewhat brighter, yes, but certainly not twice as bright. Furthermore, our eyes are more sensitive to small differences between dark tones that between light ones.

So, when pixel color values are encoded on a byte, encoding the physical output directly (which is what Gimp calls Linear light) wastes precious bytes values encoding bright nuances we can't perceive, and leaves a handful of values to encode all the subtle darkness nuances we can distinguish.

To avoid this, the values are encoded using a power law that has two nice advantages:

The perceived "medium gray" is about half-scale
There are many more values to encode low luminosity levels

The theoretical law is encoding = luminosity ** 2.2 (where a luminosity of 1.0 is maximum brightness) and then scaled to the [0 .. 255] range. So for instance physical medium gray (0.5) becomes (0.5 ** (1/2.2) ) * 255 = 186 (#BA) (the 2.2 value is what is known as the "gamma").

In practice images are encoded using a slightly enhanced law called the sRGB color space where the "physical" medium gray is #BC

GammaCurve-Annotated.png (Size: 68.84 KB / Downloads: 446)

So far, so good. So, where is the problem?

The problem is that if the gamma-encoded values are a good way to store the values, they cannot be used directly to compute things. For instance, the average color of an area with an equal number of black (#00) and white (#FF) pixels is not #80. This area emits towards your eyes half the power of the fully white area, so its average color is the "physical" medium gray, not the "perceived" medium gray as demonstrated by the picture below (that should be checked at full scale(*)):

Gamma.png (Size: 14.93 KB / Downloads: 487)

Which side is closer to the middle? In fact, making the square on the right indistinguishable from the one in the middle is a good way to tune your display...

The rule is: when doing computations that are based on physical models (spoiler alert: most are), channel values should always be converted to physical values (Gimp's Linear light) before any math is applied to them.

This applies equally to the three color channels. For instance if you do a red-green gradient, you would expect the midpoint of the gradient to be orange.

In Gimp 2.10, it is so, because the middle of the gradient is where half the power is red and half the power is green. In other words, each color channel emits half the full scale, so the middle color is #BCBC00:
Red-Green-2.10.png (Size: 2.55 KB / Downloads: 1747)
In Gimp 2.8, working directly on the gamma values, the middle color is #808000, which is a much darker brown:
Red-Green-2.08.png (Size: 642 bytes / Downloads: 1745) .

As an exercise for the reader, here is the averaging of a pattern with three colors. Click for full scale display otherwise your browser is showing you the wrong colors!

ThreeColorAverage.png (Size: 61.35 KB / Downloads: 1697)

If you use Filters > Blur > Pixellize on the pattern, you should obtain the same color.

All these examples show that the direct computation in the sRGB space produces results that are too dark.

Some more notes:

Gamma-encoding is mostly used in low-precision modes (8-bit (from most image formats) and 16-bit integer (from PNG)
Use of sRGB of course assumes that the images uses sRGB, if you load an image with another color profile, you have to use that color profile for the conversions.
If you use the Pointer dialog or the Sample Points dialog you can ask for the Pixel representation, which in high-precision images is the actual [0.0 .. 1.0] linear light value.

Spreadsheet:
GammaSpreadSheet.zip (Size: 126.67 KB / Downloads: 853)

ODS format for maximum compatibility
Contains macros, so macros have to be enabled
Shows formulas to convert from sRGB to linear and vice-versa
Also contains functions for the same (LINEAR2GAMMA and GAMMA2LINEAR) that take [0 .. 1] values and return [0 .. 1] values (in other words you still have to scale to from [0 .. 255] values, but as a bonus it works with 16-bit values too...)

Python functions for same:

Code:

import sys,math

def srgbToLinear(v255):

    v=v255/255.

    return v/12.92 if v <= 0.04045 else math.pow((v+0.055)/1.055,2.4)

def linearToSrgb(linear):

    srgb=linear*12.92 if linear < 0.0031308 else 1.055*math.pow(linear,1/2.4)-0.055

    return 255*srgb

(*) Browsers scale images using the sRGB values directly, so scaled images may end up darker than the original image. The three-color average example above is a good example. At native size the linear average (right square) is identical to the pattern at the top, but if the image is scaled by your browser (just zoom in/out) your browser will gamma-average the pattern at some point and make it look like the bottom left square.

Fill by line art with perfect edges on a B&W image or vice-versa

Sat, 13 Jul 2024 05:58:25 +0000

No method is wrong, but some time I see convoluted methods to bucket fill with line art detection when there is a BackGround, where it can be extremely simple to bucket fill with line art detection and get perfect edges even with a BG.
Thus I'll show you one simple way without deleting any BackGround, no selection, no mode change in your tool settings.

GIMP has a layer mode in which we are interested, it's the Multiply mode for white BG or its opposite Screen mode for black BG

Get you image in GIMP, then:

Create a new white layer (if the BG is white), and put it at the bottom
Create a transparent layer above the white layer
Put the top layer (original image) in Multiply mode (Screen mode for black BG)
Bucket fill by line art the transparent layer

Done, perfect edges!

Moving a dock in 4 easy steps

Tue, 28 May 2024 17:28:09 +0000

DraggingDocks.png (Size: 112.09 KB / Downloads: 453)

(click for full size)

Getting active image and layer in Python scripts

Thu, 02 May 2024 07:32:44 +0000

If your script is correctly written, in particular its "registration" code, it gets the active image (and the active layer) as parameters:

Code:

#!/usr/bin/python

# -*- coding: UTF-8 -*-



from gimpfu import *



def theCode(theImage,theDrawable):

    pdb.gimp_message('Running on image "%s" (%d) on drawable "%s"' \

        % (theImage.name,theImage.ID,theDrawable.name))

    

register(

    'the_registration_atom',

    'Example registration of a python plugin (short description, pops up in menus)',

    'Example registration of a python plugin (long-winded help that shows in the PDB browser)',

    'Author','Copyright owner','2021',

    'Test menu item',

    '*', # works on all image types

    [

        (PF_IMAGE, "image", "Input image", None),           # Active image if called from the UI

        (PF_DRAWABLE, "drawable", "Input drawable", None),  # Active drawable if called from the UI

        # Other parameters would come here

    ],

    [

        # if you want to be really nice and your script creates image/layers, you can return

        # then and descrive their type here. Makes it easier for other scripts calling your script.

    ],

    theCode,

    menu='/Test/',

)



main()

registration.png (Size: 135.2 KB / Downloads: 584)

Notes:

An important part is the use of the menu named argument. This makes the registration use the new variant. The older variant where the 7th argument is a full path to the menu item (which is still found in some examples) behaves a bit differently.
The "drawable" can be a layer (usual case) but also a layer mask or a channel or even a layer group. Some scripts can work on all types, others may want just a layer. You can use a PF_LAYER instead of a PF_DRAWABLE but the Gimp error message isn't too user-friendly. Better check in your code and react appropriately.
Since the script works on a drawable, the menu item is disabled if there is no active drawable in the image (this can happen after you deleted all layers). You can of course remove the drawable argument if the script doesn't require a specific drawable (runs on all layers, or generates new ones)
Since the script works on an image, the menu item is disabled if there is no image in Gimp. You can of course remove the image argument if the script doesn't require a specific image (runs on all images, or loads/generates new ones)

An example script "core" with long-winded comments:
regdemo.zip (Size: 2.2 KB / Downloads: 866)

Easy quick photo type painting

Sun, 31 Mar 2024 22:26:08 +0000

Heres a tutorial I put together, if u want to make art like me, its kinda fun and quick.

https://www.youtube.com/watch?v=mWq9ca9jGQo

Dynamics

Sun, 03 Mar 2024 16:34:45 +0000

Dynamics are a way to change some brush charactertistic(s) in the fly using more hints than just the pointer position. The hints are:

Pressure: the pressure on the stylus (this requires an adequately configured tablet)
Velocity: how fast the pointer moves
Direction: where the pointer goes
Tilt: inclination of the stylus, on advanced tablets
Wheel/rotation: I don't know what that is but I assume this is also for advanced tablets
Random: not a physical input, just a random value between 0 and 100%
Fade: the distance since the beginning of the stroke, as a percent of the Fade length configured in the paint tool (this repeats)

In practice with a mouse the only usable ones are Direction, Random, and Fade (and Velocity if you are really skilled).

These hints can be applied to the following characteristics:

Opacity: from 0 to 100% of configured opacity
Size: from 0 to 100% of the configured brush size
Angle: from 0° to 360°: warning: the median setting is 180°, which is an upside-down brush, so for some mapping it is useful to use an upside-down brush.
Color: a color is sampled from the active gradient
Hardness
Force
Aspect ratio: from "0" (circular brush) to whatever is configured for the brush
Spacing: from 200% spacing to the configured spacing
Rate
Flow
Jitter

You define all this by editing/creating a Dynamics (as with other Gimp resources, you cannot edit the standard one, you have to make a copy first).

The first thing to do in a Dynamic is to define the mapping matrix. To control a given characteristic using a given hint, you just check the box that is at the intersection of the "hint" and the "characteristic". The example below associates Size with Random (so the size is random), and the Color with Fade (so the color is taken from a gradient as the pointer advances). A hint can apply to several characteristics (so you could have random size and color) and a characteristic can depend on several hints (but good luck predicting how this behaves).

Dynamics1.png (Size: 156.91 KB / Downloads: 486)

You can define more finely how the hint is applied to change the characteristic by changing the mapping curve:

Dynamics2.png (Size: 111.92 KB / Downloads: 499)

For instance, in the image above, the red-green part of the gradient is mapped to smaller part of the input range than the green-blue part, while the blue-red is not mapped at all, so the output favors mostly the green-blue part.

Gimp comes if a collection of pre-set dynamics, a very useful one is Track direction, where the brush angle is rotated according to the stroke direction, so the brush keeps the same orientation relative to the stroke instead of remaining upright all along.

Dynamics3.png (Size: 45.43 KB / Downloads: 656)

Why is the Move tool moving the wrong layer?

Tue, 23 Jan 2024 08:43:18 +0000

By default, the Move tool auto-picks the layer of the top most-non transparent pixel where you click. So, starting with such an image:

Move-source.png (Size: 32.74 KB / Downloads: 445)

Depending on where you click the Move tool will pick a different layer:

Move-perspective .png (Size: 36.65 KB / Downloads: 1030)

This can be disconcerting at times because if you click at random in text, you could be clicking on a transparent spot (for instance, inside the D or O above) and so selecting a layer below (the background or the green spot), which will make the move tool move that layer and not the text layer.

So, to move text, you have to click on the characters themselves (or on any opaque part of the layer, but this is what we are naturally doing most of the time for non-text layers).

In case of dire need, you can make the Move tool act on the active layer:

Move-options.png (Size: 27.9 KB / Downloads: 1015)

But instead of checking the tool option, you can also just start the drag with a shift-click. It is usually a bad idea to set the tool permanently in the Move active layer mode, because it makes arranging several layers very cumbersome.

Finally, remember that in the Text tool, you can move the layer you are editing by Alt-Control-drag (independently of where you click).

What your mom never told you about PNG compression levels

Mon, 06 Nov 2023 12:59:56 +0000

The PNG compression level is that setting in the PNG options that is set by default to 9 and that everyone keeps that way.

But what is its purpose anyway since the compression is lossless? Once decompressed the image is still the same whatever the setting? So why is it an option?

It happens that the PNG compression algorithm is quite CPU-intensive so this setting tells the algorithm how hard it should try to find things to compress. Unlike JPEG where you trade file size for image quality, in the PNG format you trade file size for CPU during export, and on big images it makes quite a difference. So lets see what happens with four different images (16Mpix each except the photo which is 14Mpx):

A plain text (8 lines in a text layer with a large font), which is expected to compress very well (PNG is designed to be efficient on computer-generate images)
A Simplex noise which is random but with a large proportion of pure black pixels (about 50%)
A Plasma noise which is also truly random but is made slightly less random by applying a small median blur to it.
A photo of a gorgeous French village.

Each image was exported (without alpha channel) in the 10 levels of compression, measuring the final file size and execution time. Then the results were plotted, using the size relative to the uncompressed version and the relative time to compress:

CompositePlot.png (Size: 100.48 KB / Downloads: 1846)

For instance, in the Simplex case, the final size with compression is about 20% of the uncompressed size, and for level 7 the export took about 5 times longer than the uncompressed export.

These curves make it clear that in all cases, some compression is better than no compression at all. However, past levels 4-5, there is a very substantial increase in compression time for little or no gain on the file size. In absolute numbers, given that the uncompressed export took a good half second, level 9 is nearly 18 seconds for the photo, when level 6, that produces a file which is less than 3% bigger, takes less than 4 seconds.

Choose wisely

Get back the [Tool Options] in one click

Fri, 08 Sep 2023 11:18:15 +0000

Recently on reddit someone gave a tip in a very simple way to show back the Tool Options...
Why I never ever thought or tried this, it's something I still don't understand...

Anyway, there is no need to go to the top menu Windows > Dockable Dialogs > Tool Options => just double click on the tool you are using in the ToolBox!
If the Tool Options dialog is not open or you don’t find it, just Double click on the tool you’re using to bring it up. The dialog will flashes if it's opened or the dialog will open and flashes.
Yes as simple as that!

No need to say that I tried it, I did close the Tool Options, double click on a tool in the toolbox, and Voilà... the Tool Options appears on the right side (yeah just need to transfer it to its proper place, anyway if you pass by the top menu it's the same > appears on the right side )