ONE ChatGPT Prompt to RULE THEM ALL - MidJourney/Stable Diffusion

LARRUZ shows how to use ChatGPT to write better prompts for MidJourney v5, Stable Diffusion, and DALL-E. The template includes subject, five keywords, camera type, lens, time of day, style, film, lighting, and aspect ratio. He walks through examples with urban scenes, portraits, and food photography to show how detailed prompts improve image results.
Chapters
In this video I want to show you a magic formula with the help of which you can get the Most out of ChatGPT. In this video we will look at an example application of using ChatGPT as a prompt generator for midjourney. But keep in mind you can apply the same logic to any application you are working on. It doesn't have to be related to image generation. The core idea behind this approach is very simple. If you know what makes a good prompt or good image in mid journey. So you can teach ChatGPT the structure of the prompt, you simply need to include what parameters you want in your prompt. Now the great thing is that this is not just limited to midjourney. You can use this with any image generators, e.g. stable diffusion or Dali. So using the simple formula, you'll be able to generate images of people, let's say portrait of places, even food photography, or if you want to run ad campaigns, so you can generate product images as well. Now the great thing about this is you're not going to be generating random images, but rather you will have a lot more control on what you want in your images. So let's get started. So here is our prompt. So we are telling chatgpt structure of a mid journey prompt. So it's supposed to have a subject of the image, then five descriptive keywords, camera type, then the type of lens, time of the day, then type of photography and then type of film to use realism level, best type of lighting, first the subject, then best aspect ratio for the subject of the image in this format. So we're defining the aspect ratio as a width and height. Then we're asking it based on the above structure, create a detailed narrative of the scene. So for version five, which is now the default model for mid journey, you need to have really detailed descriptive prompts. So we are asking it to kind of create a narrative for that and then generate different variations. So the first subject of the images, urban photography. So let's save and continue. Okay, so let's look at it, what it came up with. So the parameters, urban photography, gritty, vibrant, busy, dynamic, industrial. So I think it's going to be some industrial scene. The type of camera, then wide angle, golden hour, street photography. And it's going to be a color image. Then realism level is really high and then a soft light and aspect ratio is 9, 16 by 9. Okay, so based on these parameters it came up with this narrative. So as the sun begins to sit on the bustling cityscape, the urban landscape comes alive with gritty energy. The vibrant colors of the buildings and the neon sign contrast against the darkened sky. Okay, so it's a pretty, really nice narrative. Right. Just one thing I think I need to be careful about is when I'm copying this, it's simply not the dash dash ar, but need to be careful with that. Okay, so here's another one. Sleek, modern, minimalist, clean, futuristic, mirrorless. And here's the narrative and a blue hour. This it takes on the sleek and modern statics. Let's try a couple of these in my journey. So here are the results for the first one. So this is like industrial city. And the it's a golden hour and neon signs are here. So it looks pretty, actually pretty awesome. Especially look at the sky, the details. So here is just. It's. It's simply amazing. Look at it. And the second one is more futuristic looking city. So you see the skyscrapers. They're supposed to be contrasting lines and shadows. So you can see those details in here. It looks pretty good. All right, so next I asked it to create transfer a portrait for young woman. And let's say look at a couple of examples. So subject is that descriptive keywords, bold, energy, dramatic, mysterious, confident and mirrorless. Camera type. Then zoom lens type. It's a twilight and fashion portrait, black and white, medium, low key. Aspect ratio is 16 by 9. Let's look at another one. I guess like, I think it would be good to add the location as well. And here's the description based on that. So in the twilight hours, a young woman's bold and confident personality is captured in a fashion portrait shot on a mirrorless camera with a zoom lens. All right. Okay, let's try this. So again I'm going to type in slash imagined. And with all these I need to manually give the aspect ratio. Okay, look and select another one. Now one great thing I really like is it actually explains the port the why certain aspect ratio was selected. So for example, three by two aspect ratio is perfect. Showcasing the subject's natural proportions and adding a classic touch to the portrait. Okay, so I'm gonna try this one. So here are a couple of examples. This is pretty beautiful, especially the backlight in all these cases. It's really amazing. And that's because of the morning light mentioned in the prompt. Right. Then here's another one that is black and white. It actually adds a lot of dramatic effect. So it's pretty cool looking. So so far we simply looked at more natural scenes. Let's look at it. If it can come up with prompts for abstract art. And poster. So let's say a poster of Alps which is going to be interesting to see what it can generate. Okay, so majestic, breathtaking, snowy. It's going to be a drone shot, wide angle, sunrise, landscape and realism is high. Okay, that's pretty nice. The second one is going to be a medium format film camera, telephoto and early morning. Okay, that's black and white. I don't think it's going to look that great. But let's see. I actually like it. In each case it is suggesting to use different cameras. Okay, let's create a couple of them. Okay. These images are simply breathtaking. Look at this. This is I think the drone shot. So that's a drone camera with a wide angle lens and looks pretty awesome. The second one is early morning mist. So it actually put a black and white. But it's also pretty nice because you can see these clouds and mist in there. This is pretty awesome. And the third one is like more of a spring, a vibrant and colorful view. So there are these flowers and in the background you can see the mountains. These are pretty amazing. Okay, next prompt was related to food photography. So let's see what type of burger it can come up with. And then like the keyword it's coming up with juicy, cheesy, grilled, gourmet, malt watering type of lenses it's using is macro. So that's good. And it's going to be close up. Okay, a burger. Messy, unconventional, artistic, modern, creative. It's using a smartphone camera. Okay, this one. Film slr. Okay, these I think are going to be pretty interesting ones to look at. All right, so we have some mouth watering results. Look at these burgers and look at these. There looks simply amazing. Look at the attention to the details. If I open in the browser and look at this. This is simply amazing. Okay, so here's another one. Headphones with colorful smoke coming out of the earpieces. Let's see how it's going to look like based on this. A vibrant, colorful, psychedelic cereal, otherworldly and then playful whim, quirky fun. So I'm really interested in how the images are going to turn out. Okay, so here are the results. This is a poster and looks pretty neat. It's a little bit funky. A little bit funky. Then there was another poster that it added. And this looks pretty cool. I would say it's nice. And here is the last one. Really colorful. Alright guys, there you have it. Hope you found this useful. Let me know what you guys create with this. Thanks for watching. See you in the next one.










