Prompt 1:
movie still, close-up, a man in a futuristic coat in an orange sand patch, minimalist, hazy, foggy, red and amber, low contrast, muted tones
Prompt 2:
street photo of a man in a futuristic suit, neon, provia, hazy, foggy
Image 1:
The first prompt consistently yields results of a man in a desert, almost always with the same perspective and pose.
Image 2:
The second prompt yields quite varied results of a futuristic, foggy street with different characters. The pose is also quite different compared to the first prompt.
Image 3:
Let’s blend two prompts together using the ‘::’ operator. As you can see, the results took the pose from the second prompt. Now the city is mixed with the desert. The colors were taken from the first prompt.
Image 4:
But what if I want to make the city more prominent? I would add more weight to the second prompt using ‘::2.5’. Yes, you can use decimals. I added 2.5, since 2 had only a small effect and 3 was too big.
Image 5:
Why use '::' instead of a comma in a prompt? Let's see. With a comma, you can't control each part well; you might get a desert with a barely visible city. Changing word order helps, but it's not as easy as using '::'.
Image 6:
In the last experiment, let’s add more weight ‘::2’ to the first prompt. Now you get a desert, but the second prompt affected the pose of the main object. BTW, the results are quite similar to those of Image 5, where we used only a comma. This is logical, considering that the first part of the Image 5 prompt had more weight, just as when we added '::2'.
movie still, close-up, a man in a futuristic coat in an orange sand patch, minimalist, hazy, foggy, red and amber, low contrast, muted tones
Prompt 2:
street photo of a man in a futuristic suit, neon, provia, hazy, foggy
Image 1:
The first prompt consistently yields results of a man in a desert, almost always with the same perspective and pose.
Image 2:
The second prompt yields quite varied results of a futuristic, foggy street with different characters. The pose is also quite different compared to the first prompt.
Image 3:
Let’s blend two prompts together using the ‘::’ operator. As you can see, the results took the pose from the second prompt. Now the city is mixed with the desert. The colors were taken from the first prompt.
Image 4:
But what if I want to make the city more prominent? I would add more weight to the second prompt using ‘::2.5’. Yes, you can use decimals. I added 2.5, since 2 had only a small effect and 3 was too big.
Image 5:
Why use '::' instead of a comma in a prompt? Let's see. With a comma, you can't control each part well; you might get a desert with a barely visible city. Changing word order helps, but it's not as easy as using '::'.
Image 6:
In the last experiment, let’s add more weight ‘::2’ to the first prompt. Now you get a desert, but the second prompt affected the pose of the main object. BTW, the results are quite similar to those of Image 5, where we used only a comma. This is logical, considering that the first part of the Image 5 prompt had more weight, just as when we added '::2'.