I tried colors heh…Did two tiny number labels on pants for some reason…
A wide oil painting depicting eight explorers, each dressed in one of the eight primary colors: red, orange, yellow, green, blue, indigo, violet, and black. This diverse group includes four women and four men of varying ages and ethnicities, styled as a mid-19th century American landscape painting. They are viewing a lush landscape from a mountain passage. Each explorer’s shirt has a numeral from 1 to 8 on the back, in the order: Man (1), Woman (2), Woman (3), Man (4), Man (5), Woman (6), Man (7), Woman (8).
Issues were encountered in going the opposite direction…
A group of two thousand explorers (comprised of women, men, and children of varying ages and ethnicities), on a frontier, who have just reached a point in a mountain passage where they behold a vast and beautiful land spread out before them, during the month of May. Each person of the 2000 has incredible detail paid to their appearance, clothing, and realistic representation of the human form down to minute details, every single person an accurate representation of a human with the artist’s highest attention and no expense spared in the time-consuming painting. Style: vibrant oil painting, Hudson River School movement, landscape, American frontier art.
We have numbers… kinda… lol… shows there’s no “reasoning” going on…
A wide oil painting depicting eight explorers, each dressed in one of the eight primary colors: red, orange, yellow, green, blue, indigo, violet, and white. This diverse group includes four women and four men of varying ages and ethnicities, styled as a mid-19th century American landscape painting. They are viewing a lush landscape from a mountain passage. Each explorer’s shirt has a numeral from 1 to 8 on the back, correctly numbered from left to right: Man (1), Woman (2), Woman (3), Man (4), Man (5), Woman (6), Man (7), Woman (8).
A wide oil painting in the style of a 15th century European artist depicting a large crowd of approximately 234 people gathered in Nuremberg, experiencing the first alien contact on Earth. The scene is solemn, with people of various ages and attire typical of the 15th century, showing mixed expressions of awe, fear, and curiosity. The setting includes medieval buildings and cobblestone streets, with a mysterious alien spacecraft subtly visible in the sky, casting an eerie glow over the scene.
Here’s the image depicting a large crowd in Nuremberg during the 15th century, experiencing the solemn moment of first alien contact on Earth. The scene includes various expressions and details typical of the era, with an alien spacecraft subtly visible in the sky. Let me know if there’s anything else you’d like to adjust or add!
The current Dall-E3 (which I have tried using with ChatGPT) does not seem to accurately represent the number of people, gender, or order, regardless of whether numbers are assigned.
Often, the number of people is incorrect, or even when numbers are assigned, they are randomly ordered.
A wide oil painting depicting eight explorers, each dressed in one of the eight primary colors: red, orange, yellow, green, blue, indigo, violet, and white. This diverse group includes four women and four men of varying ages and ethnicities, styled like a mid-19th century American landscape painting. They are viewing a lush landscape from a mountain passage. Each explorer’s shirt has a numeral from 1 to 8 on the back, correctly numbered from left to right: Man (1), Woman (2), Woman (3), Man (4), Man (5), Woman (6), Man (7), Woman (8).
When I try to make DALL-E 3 express something that is generally opaque as translucent, it appears to be wrapped in something translucent like plastic.
Maybe Sola (or DALL-E 4?) will overcome that problem.
A close-up of a translucent celery stalk in the foreground blending naturally into its surroundings. The background is a lush forest with various plants and trees, visible through the translucent celery stalk, creating a serene and magical atmosphere.
@_j about the style vivid and natural make a lot of difference and in my usecase mostly animating and or creating 3d Illustrations ‘vivid’ always is the best. Or In my case at least.
dalle.text2im({
size: “1024x1024”, // or “1792x1024” for a wide image, or “1024x1792” for a full-body portrait
prompt: “A detailed description of what you want the image to include”
});
For midjourney there is extensive documentation and people sharing their prompt templates on scribd and the web. you know what to type and to expect.
i created a prompt generator and enhancer service using the API implementation of openai, AS WELL as the GPTplus subscripton ( and have a Monica subscription with Dalle-3 included) , so i try all my prompts on multiple places but against the same model.
and on all four it is different on what you can expect. chatGPT4 has an API connection with Dalle-3 so that works , but you can also talk to Dalle-3, the chatBot can be a little pain concerning content policy blockers. if you try to generate an image in the style of Frank Stella looking like his artwork HarranII, then openAI Chatbot will block that request, but not via the API.
and Dalle and chatGPT will change your prompt if you use either too litte words or if you use words like drunk. this is changed into friendlier depictions of the drunk state…
the API is expensive, i create large amounts of images with the highest size and style ( that is 1792 * 1024 with style: vivid ) and that is about 1 cent for an image. then i also connected to an assistant and the normal chatbot for prompt enhancement and i spend easy 250$ month on the API alone. ( and for understanding other services i have temporary also Google Gemini, plus chatGPT+ plus Github Copilot + Monica + midjourney + runway + gencraft + stablediffusion. ) this is al temporary of course as i cannot keep on spending much on an experimental GAN Model
OpenAI enforces inclusivity by altering your prompt to include : hispanic, south asian, middle easy and African people in the image. and last week i had my first non-binary person addition in the image. nothing wrong with that but i did not request that.
and then the content policies will be stricter as they want to protect the intellectual property of their customers read: customers of MS)
i have tons of medioker prompts and images . if people want access to them than i can share them… i am a developer, not an artist
i guess that because i paid for it., sharing can do no harm.
I tried to generate a picture with as many people as possible, and it ended up looking like a “Where’s Waldo?(Where’s Wally?)” scene.
An enormous crowd of countless people, tightly packed together in a vast open area. Each individual is small but distinct, filling the entire space. The scene is teeming with diverse individuals, standing shoulder to shoulder, creating a sense of endless humanity in every direction.
I started from StanleyKrute’s first prompt:
“Imagine a group of 8 explorers, women, men, and children, of varying ages and ethnicities, on a frontier, lush oil painting style, in the manner of a mid-19th century American landscape painter of the West. They have just reached a point in a mountain passage where they see a vast and beautiful land spread out before them, during the month of May.”
To save your time, here is the summary of the concept:
Use ChatGPT Classic, tell it the image you want to create, and add at the end:
“{description of your image}
—-
The above is a picture I imagine. I know you can’t generate images, but I want you to describe it with topics including, but not limited to: Composition, Subjects, Color Palette, Light and Shadow, Types of Elements. Could you do that?”
Open a specialized GPT for this image generation process. This GPT should support DALL-E image generation. Its instruction is:
“This GPT specializes in creating images from user-provided descriptions. Users input details about the layout, main subject details, background, colors, and lighting. The GPT uses the DALL-E tool to create these images, strictly following the user’s requirements without adding its own interpretations. It interacts with users to gather all necessary details and clarifications to refine the image and meet the specified artistic requirements.”
Copy the feedback from ChatGPT Classic to the Drawing GPTs. Only copy the text related to the five major topics, other general conversation text does not need to be copied. And at the end write:
“{image description copied from ChatGPT Classic}
—-
Draw the above image, wide screen aspect.”
After you get the generated image, you can go back to the ChatGPT Classic chat room to tell it what you want to modify. To ensure it will always provide feedback in the format of the 5 topics, it’s recommended to add this after your request:
“{Your change request}
Friendly reminder, the description has to include: Composition, Subjects, Color Palette, Light and Shadow, Types of Elements.”
To get the above image, I have tried 4 times.
The first time I received:
I gave it a few tries and there were still more than 8 characters, so I asked for another modification:
Okay, I do get that 8 explorers now. But the background of the image has generated extra people to make the image rich. We have to state clear that no more people other then the 8 main character is allowed. Can you update the description please?
I consistently get 9 people. I guess the issue is the sixth character. Let’s just say she is a mid age women or somethings similar. Mention her as child’s mother may cause an extra child into the picture.
Then I got the final image, which I posted at the top of this post.
