Daily Generations: Not Castlevania

Created with: DALL-E 3

Simple style exploration today.

A good daily exercise to hone your generative imaging skills is to visualize an art style you’d like to create. Either in your mind or with Google images, Pinterest, or a visually appealing piece of media you saw recently.

Once you have the work in front of you or fixed in your mind’s eye. Analyze what makes the style distinct. Think about the mood, the color palette, the contrast. Is the line work rounded or angular, is there any line work at all? What shapes tend form in the piece? How abstract is the work? If the work is photographic, what do the shadows and lighting look like?

Once you’ve really dissected the piece. Try to describe it to yourself using the data you gathered. Focus on describing the style rather than the content.

Read your description back and see if you see your piece in it. Surprisingly often, you’ll notice the description could actually also apply to a totally different art style. For the purposes of this exercise, that’s actually a good thing. Bring up the other work and try to see what makes the two unique from each other.

Finally, put your updated style description into your favorite generative AI imager. Add content description as well, but it not too detailed. We’re focusing on style not composition. Compare the results and see how close you got.

Doing an exercise like this on a regular basis can really improve your prompt writing. As well as your general english language and visualization skills.

Pretty soon you’ll no longer have to rely on “Show me Castevania but not!”