I finally got my hands on the new Google Gemini Image, it is definitely a step forward compared to all similar models. It is also the first contextual model that can truly replace Control Net. The first screenshot from Rhino shows the current project I am working on, the second one is a render from #comfyui.
❤7👍2
if you took our #grasshopper course on Stepik platform (it's a russian language course), then you are familiar with my project i used there as an example of NURBS modeling in Rhino. now I decided to try rendering it using Gemini 2.5 Flash Image. i'll show you what I got next.
❤5🔥1
These are renders, without any post processing, exactly as they came from #comfyui. To my taste, they are excellent. Obviously, some details are not rendered perfectly, but this can be solved by combining several renders with different seeds in Photoshop. also, it is possible to work with the surrounding in a same way.
A few interesting points:
- The new Gemini model is very cheap, several times cheaper than Flux.1 Kontext.
- It's also fast, usually returning results in a second. You have to wait for some generations, probably due to server load, but still not as long as with other models.
- I already mentioned that this is the first model of its kind that can truly preserve the context of the original image, which is extremely important for rendering. Sometimes this can even be a hindrance when you want to change details, in this case you have to mess around with prompts and inputs, but it's good that the model is inexpensive, so there's room for experimentation.
I remember that couple of years ago i thought that AI models cannot replace conventional rendering, but now we are close to that point, at least for early concepts.
A few interesting points:
- The new Gemini model is very cheap, several times cheaper than Flux.1 Kontext.
- It's also fast, usually returning results in a second. You have to wait for some generations, probably due to server load, but still not as long as with other models.
- I already mentioned that this is the first model of its kind that can truly preserve the context of the original image, which is extremely important for rendering. Sometimes this can even be a hindrance when you want to change details, in this case you have to mess around with prompts and inputs, but it's good that the model is inexpensive, so there's room for experimentation.
I remember that couple of years ago i thought that AI models cannot replace conventional rendering, but now we are close to that point, at least for early concepts.
🔥3❤2❤🔥1👍1
For fun, I made images with a plan view and a section. It was more difficult here, I had to experiment longer with the text prompt and input images. And the result, unlike the renders, of course cannot replace the real drawings, but it still looks interesting. Again, no additional processing was done; the model created all the annotations itself, and they are not entirely meaningless. Gemini understood where the bridge was, where the main space of the museum, and correctly placed the river. In short, it's cool.
❤5🔥2
Media is too big
VIEW IN TELEGRAM
Continuing the story with this project. I'm diving deeper and deeper into Google Flow, and I thought it would be good practice to make a video for this museum, especially since last week, all generations for Veo3 Fast were free if you have an Ultra subscription, so now you don't have to worry as much about spending all your credits and having to buy additional generations. It is worth mentioning that different features in Flow are available through different models, so it is not possible to make all parts of the video in Veo3 Fast; for many things, the service switches to Veo2 (which is not unlimited). In general, such a video takes about 4-5 thousand credits if you don't skimp and sometimes switch from Fast models to Quality. If you do skimp, you can easily get by with 1 thousand. Moreover, the difference between Fast and Quality lies in the models' structure, but in practice it is insignificant, and Fast generation can easily give better results simply because it knows or tries some other concept at the right moment. In total, you have 12,500 credits in the Ultra subscription (there was an announcement that they would be doubled, but this did not happen in my account; maybe they will update it from the next billing month).
Sound, captions, and editing in Adobe Premier Pro. Veo3 can generate videos with sound right away, but the audio won't be consistent from clip to clip, so you have to work with the sound separately to glue it together. I used free libraries from YouTube.
I thought out the plot of the generations in advance and made many attempts for each fragment to achieve the desired result, but in some places the AI model came up with interesting concepts that I left in. For example, the projection with the wolf turned out that way by accident. I didn't describe it in the prompt, I just asked for a video with a modern exposition about the northern lights and the Finnish forest.
Sound, captions, and editing in Adobe Premier Pro. Veo3 can generate videos with sound right away, but the audio won't be consistent from clip to clip, so you have to work with the sound separately to glue it together. I used free libraries from YouTube.
I thought out the plot of the generations in advance and made many attempts for each fragment to achieve the desired result, but in some places the AI model came up with interesting concepts that I left in. For example, the projection with the wolf turned out that way by accident. I didn't describe it in the prompt, I just asked for a video with a modern exposition about the northern lights and the Finnish forest.
🔥7👍5❤3
Media is too big
VIEW IN TELEGRAM
By the way, my second workshop on the PA Academy platform starts this weekend. Grasshopper, Rhino.Inside, Revit, and a bit of AI. We'll be modeling this building from the video. #education
https://paacademy.com/course/bim-rhinoinside-for-advanced-tower-design
https://paacademy.com/course/bim-rhinoinside-for-advanced-tower-design
🔥12
Adobe has added Flux.Kontext and Gemini 2.5 directly to Photoshop. This feature is currently only available in the beta version. It can be installed via Adobe Creative Cloud. All the pros and cons of the models are carried over with them. For example, Gemini can only output images in low resolution, even if your input is in a good quality, so you will have to do the upscaling separately. #ai
❤5