Line of Production/Editing

On this page, I would like to go into more detail about how I create and edit artworks. I divide the entire process behind the generation/editing into several steps, from the idea to the finished work.

STEP 1 - The Foundation

The first step is fairly simple: before anything can be generated or created, the basic theme of the image is determined so that the entire subsequent process has some kind of structure. In this case, I will use an example from my “Brittle and Broken” series. The basic idea behind this series is a melancholic atmosphere and sad symbolism. Since the foundation has already been laid, this image still needs a specific theme.

In this case: a desolate environment, a concrete floor covered in cracks, the symbol of a heart visible in the form of a cut-out in the floor. The whole thing is intended to give the impression that something has been taken away or is missing.

STEP 2 - Creating a Preference

Depending on how good your prompts are, you can already generate really great images. Because that's not always the case for me - or because I simply prefer to take advantage of all possible ways of generating images - I also like to work with preferences that contribute to the final result in the course of my work.

A preference is a rough generation, without too many details, on which the second stage of my work is based. It is more a matter of developing an idea in order to narrow down and specify the theme and rough design for the end product.

Typically, I set up preferences via perchance.org. It's free and fast - ideal, even if not all ideas come to fruition.

Example: 

The general idea: a sad scene, something with the symbol of a heart; with depressive energy

STEP 3 - Generating the Image

Now that the preference image can define the final result much more clearly, it can be used for the next step. In my case, that means Adobe Firefly. The preference image created in combination with a follow-up prompt for more details and even more precise direction provides even greater depth. Additionally, changes are made if necessary - both minor changes, i.e., minor adjustments to details, but also complete rearrangements.

Example: 

Reference Image + Prompt will turn into:

Now it depends on whether the resulting image meets the specified guidelines. In my case, that means: Does it capture the right kind of emotion? Does it fit in with the given theme? How good is the overall quality of the image? Typically, several preferences are created in order to test different approaches and directions. During the third step of processing, it becomes evident which approach will be used for the next step. Or, if any changes are made, it will now be clear how these will look and what exactly will be changed.

STEP 4 - Finalizing / Enhancing

If the desired result is achieved in step 3, the respective image is used for the final step: finalization and enhancement. The desired goal is to create even more detail and depth, making the image even more powerful. The tool of choice here is simply ChatGPT. In step 4, the final image from step 3 is taken and shaped into its (almost) final form.

Result: 

STEP 5 - End Processing 

The final step in this sequence consists of manual post-processing by filtering and adjusting the color saturation. This is usually done relatively quickly and is intended to give the generated image a little more personal flair. My tool of choice here is GIMP, but other editing programs such as Photoshop also work, of course.

I will explain the new process in GIMP itself, which I use for all images, in more detail another time. For now, here are the most important points: the entire image is softened using filters. Depending on the subject, the saturation is also increased or decreased to enhance the atmosphere of the image. Finally, the focus is adjusted using another filter.

Result: