Last week I wrote about spotting AI generated images (more like not being able to spot them). The generators are getting so good. How do YOUR AI images look? Do they need some work? Are you having trouble getting the results you want?
Let’s see if we can improve your images.
First, what AI says to do . . . write detailed, descriptive prompts that specify the subject, environment, and composition, and then add details about the style and mood. Being specific and avoiding contradictory terms will yield better results, and you can experiment with using certain keywords or even referencing existing images to refine your output.
Helpful, but not helpful. To get better images, you need to think back to when you were in school learning how to write creative stories.
These examples will help:
Be descriptive and specific
- Subject: Clearly define the main subject. Instead of “a woman,” try “a mid-30s woman with shoulder-length auburn hair”.
- Environment: Describe the setting, time of day, and weather. Instead of “outdoors,” use “in a sun-dappled forest clearing during golden hour”.
- Composition: Specify the camera angle, framing, or shot type. Use terms like “close-up,” “wide shot,” “overhead view,” or “shallow depth of field”.
Define the style and mood
- Style: Mention artistic styles, mediums, or specific artists. Examples include “cinematic,” “photojournalistic,” or “in the style of Van Gogh”.
- Mood: Use adjectives to convey the desired atmosphere, such as “ethereal,” “serene,” or “high contrast”.
- Use quality-enhancing terms like “high resolution,” “masterpiece,” “intricate,” and “best quality”.
Have fun with it. The more you practice the better you will get.