The “secret” to great images with Picture This… lies in the prompt you send to the Stable Diffusion AI engine. For Picture This… the prompt is the name of the Marker. Prompts are limited to approximately 70 words, and you can fit all that in a Marker Name!
The basic structure for a text-to-image prompt is:
- Subject: who or what is the focus of the image, e.g. middle aged man
- Doing: the attributes of the image. The subject can be doing anything, or it could be as simple as “looking to the distance”, or “wearing a scuba mask”
- The Look: the visual attributes of the image. Camera angle or shot type, e.g. Wide, View from above, etc. For Picture This… we recommend Medium Wide to Wide shots as you can zoom in to the generated images up to 200% without losing quality. You are staging the shot.
- Lighting style: “Cinematic”, “bright sunshine”, “foggy”, etc.
- Art Design: The style of the image. Here’s where you tell Stable Diffusion how to style the appearance of the image. For example “cartoon” “ultra real” “water color” digital painting” “surrealistic”etc.
- Color Treatment: Do you want the color to be vibrant, muted, sepia, “orange and teal”?
- Where it is: describe the environment. Are they in a “forest”? Is it a “cityscape” or “designer kitchen”?
- Quality: terms like “best quality” “intricate details” and “masterpiece” are often used as descriptors and they will have some positive impact on the images. We use “best quality” and “masterpiece” on almost every image we want to look realistic.
Putting it all together, a prompt will look something like these examples, which have had the generated square image scaled and/or composed for the widescreen ratio of the edit.
Picture This… uses the latest Stable Diffusion AI model so any article on creating prompts for Stable Diffusion will apply to creating prompts in Final Cut Pro and Premiere Pro for Picture This…. These are some we’ve found useful. Note that negative prompts are not currently available in Picture This…, but support is planned for a future update.
- Stable Diffusion Prompt Guide for Beginners https://aituts.com/stable-diffusion-prompts/
- How to come up with good prompts for Stable Diffusion https://stable-diffusion-art.com/how-to-come-up-with-good-prompts-for-ai-image-generation/
- How to use negative prompts https://stable-diffusion-art.com/how-to-use-negative-prompts/
Finally, here is a link to information about the uses, limitations and bias of the Stable Diffusion models used by Picture This…: https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0#uses