I am enjoying having the ability to prompt ChatGPT to create DALLE-3 images. There is one thing that I can’t seem to figure out how to do: I would like to create images to use for Teams meeting backgrounds. The trouble I am having is getting Dalle to create an image that won’t get “cut off” - I know the aspect ratio isn’t correct and I don’t know how to change it (or if I can change it) but if I could even just get Dalle to zoom out from the image I am trying to create as the main focus, that would be good enough. I have tried saying things like “using a zoom lens” “from 20 feet away” “with a wide angle lens” but it just seems to get ignored in the resulting image.
Thanks for any ideas.
Welcome to the forum.
I’ve been prompting text2image since VQGAN (et al)…
Can you share the prompt with us? Both what you put in and what ChatGPT makes the prompt? That would help you get squared away, I think. Thanks!
Here is an example I just tried - but the dog is always very close and takes up most of the image:
“photo of a poodle eating an ice cream cone taken with a wide angle lens” (also tried … “taken from across the street” and “from 20 feet away”
Here are four diverse and detailed prompts that I’ll use to generate the images:
- Photo of a spacious urban setting where, at a distance, a white poodle sits on the sidewalk, enjoying an ice cream cone. The wide-angle lens captures the vastness of the surroundings, with the dog as a noticeable but not dominating element in the scene. Buildings, trees, and pedestrians of various genders and descents can be seen, but they’re all secondary to the poodle’s delightful moment.
- Photo taken from about 20 feet away, capturing a park scene with children playing and birds flying. Towards the center, but not too close, there’s a black poodle licking an ice cream cone. The wide-angle perspective emphasizes the greenery and park elements more than the dog.
- Photo of a sunlit beach, where the horizon stretches far due to the wide-angle lens. People of different genders and descents are sunbathing and playing. Somewhere in the middle, but not too prominently, a brown poodle stands, holding an ice cream cone in its paws, looking as if it’s about to take a bite.
- Photo of a bustling city square in the evening, illuminated by streetlights. The wide-angle lens showcases the vastness of the square, with shops, cafes, and people of various descents and genders going about their activities. Not too close to the camera, a grey poodle can be spotted, standing by a fountain, nibbling on an ice cream cone.
With the new ChatGPT Plus DALLE3 interface, it can sometimes help to be a bit more verbose in what you want. Hope this helps!
Thank for the suggestions. This is very helpful.