Gimp is the most advanced pixel picture editing program under Linux. However many beginners decide to remove it from the computer since it is not intuitive. It needs time to understand its concept and how to use it. Therefore the common pitfalls:


Gimp does not know objects, so you can not click to select. To select something the area of pixels need to be selected first and can then be modified. A picture is a single 2D array of pixels (however gimp has some features to deal with that efficiently!). Pictures containing objects have usually vector oriented formats as svg, so consider to not use gimp.


Gimp supports intensively layers to work efficient. Many frustrating experience of new Gimp users occur, since editing is just possible to the layer selected, but the new Gimp user wants to modify an other layer. An example is that when adding text to a picture, the text goes to an other layer to not damage the picture. Additional it is not intuitive to pop up the layer menu using Ctrl+L or via Windows > Dockable Dialogs > Layers


Layers are just supported when working xfc Gimp format, they get merged and are lost if you export to other formats as jpg, png.

Worth to install is also the gimp help that uses the F1 key to sensitively point into the manual. Under Gentoo it is a separate package:

emerge gimp-help

Daily work with Gimp

When a new file is opened, "Advanced Options" allow to select a picture with transparent background. This is much better than later erase the background. Erase to transparency can be done easily if the gimp vocabulary is known. Go to Layer > Transparency > Add Alpha channel. Note: not all file formats understand transparency.

To paint a straight line use the brush but press Shift before and while working with the mouse.

To fill an area use the gradient icon and select gradient as FG to BG and the shape as radial or linear.With the mouse click you select start and end points. Other niche effects is FG to transparency with white FG. This gives a reflection effect.

Copy and past work after an area of pixels has been selected (Ctrl+C and Ctrl+V). Past creates a new special layer (floating selection or pasted layer). Anchor layer Ctrl+H is then used to push it back to the original layer and overwrite it. An other option is Add New Layer to have the floating layer converted to a regular layer. Copying to other layers of the same image is not intuitive. Make just the destination layer and the pasted layer visible. The convert he pasted layer to a new layer (add new layer) and then do merge visible layers.

The array of pixels can have different meaning for its data: Gray scaled, RGB color or indexed color (the value of the pixel is an index in a color map, and picks from there the color). To add color to a gray scaled picture Image > Mode has to set to RGB.

If you make a drawing from scratch, do no draw on the background layer, add a second transparent layer and start there. If just an area is selected, it can just be edited inside of this are, rectangular select the whole picture if you want to get everything.

Animated Gifs with Gimp

Gimp supports well animated gif. When such a gif is opened, Gimp shows for every frame a layer that can be modified. The frames can have any name. Commonly the names are chosen similar as the Lowest layer in the layer window has the name "Background" and the layers above "Frame 1" and so on. The layer name (layer attribute) can contain additional information as "Frame 2 (100ms)", this tells how long the frame will stay visible and overwrites the default value that can be set when later exporting the gif file. Obviously when exporting to gif the option animated gif needs to be selected, otherwise all frames get merged into single picture. Under the menu Filters>Animation>Playback the result can be observed.

Icons with Gimp

A regular picture can be saved as *.ico (Microsoft Windows Icon). The size of the icons are typically 16*16 pixels, however it is also possible to have ico files that support multiple sizes. This is actually necessary since a 16*16 icon looks ok on the web browser corner, but looks ugly on the desktop. On the desktop icons of 32, 64, 128 pixels are used. To do this additional layers have to be added with the desired sizes. The icon can then be copied and scaled to those layers. To do this, different ways are possible. Mostly you start with the largest icon size 128*128, then you could save it in different sizes. After that you can open all individual files with open as layers, where you can select the other size images. A trick is here that the result gets the size of the largest picture. This can be controlled on the layers window and be influenced using the sequence the files have in the open dialog. At the end save it as ico.

Image maps for web pages

Image maps are used in html pages that contain an image with sensitive areas. The start is an image that can be opened in gimp. Filters > Web > Image Map opens then a second window, where the original window appears as well. Now the sensitive areas can be drawn on the original file and its actions defined. The result is not saved in the original file but in a *.map text file. This file contains the html code that has to be put into the web page. The code contains the map plus the reference to the original picture and the necessary declaration.

Linurs Servernest startpage