Hi everyone, I find it very interesting to have AI in SketchUp, but I’m not getting the same results as before… attached you can see the same scene rendered with Diffusion about a year ago, and the same scene with the latest version of Diffusion.
Why is the result so different? (see attached image)
In both cases I used:
- “Photorealistic interior” preset
- Respect model geometry: 1
- Prompt influence: 0
The same thing happens in all models where I am using Diffusion. It is not a problem related to a single file.
I am using the version listed in the Extension Warehouse as 2.0.1, while once installed in SketchUp it is called 1.2025.9.28 (see attached image).
Hi Marco,
The version from about a year ago had a minimum value different from 0 for the prompt influence. I guess that would correspond to a value of around 0.35 now.
Best,
Jacques
Hi Jacques,
Thanks, I noticed that the scale now starts at zero, but even when I enter a value of 0.35, the result doesn’t improve… (see images).
Here in Italy, some people are writing to me asking for clarification, and I don’t know what to say 
The preset is called “Photorealistic Interior,” and even when using Autoprompt, a photorealistic render is promised (“…click Generate to create a photorealistic image of your model”).
I know that AI doesn’t work like a rendering engine, of course, but perhaps the use of this tool should be better contextualized, don’t you think?
For example, the description in the Extension Warehouse “Rapidly generate concept visuals and make design decisions faster” does not promise rendering…
