Making of "Villa Toscana"

2024-01-19T11:33:42.000Z

Exterior animation and the use of AI in architectural visualization.

The idea of creating exterior animation appeared to me quite a long time ago, I was waiting for a project on the basis of which it would be possible to create an animated video. Villa Toscana became such a project.

Hi everyone, my name is Makary and today I would like to tell you about visualization and animation of a new project, as well as the application of AI in archviz.

The project was carried out in a team @derevopark project architect - Daria Ragozina (https://www.instagram.com/dari.ragozina/)

There were several animated cameras on display for the project, generally ordinary spans that are not difficult. It was much more interesting to set the objects in the scene in motion.

For example, the fire was created in the simplest way without using any simulations. In 3D Max, we add the usual material CoronaLightMTL and throw in it a video footage of fire in AVI format. Throw the material on the plane, duplicate it a couple times and rotate it to create more volume. Thus in just a couple of clicks burning hearth is ready.

To add a light wind effect to the trees I used the Noise modifier, setting a small amount of noise and checking the animation noise checkbox.

You can also apply and animate the Twist modifier to add sway to the whole tree. The main thing is to specify small values so that the object does not sway like in a hurricane.

In conclusion, I would like to share my experience of using AI to improve the picture and add details to the visualization. The most pleasant tool is of course generative fill in the new Photoshop. With its help you can raise the realism of the frame at once during rendering. So, for example, on this frame I selected the fruit on the apple tree and generated more realistic than they were in the model. Also by selecting the sky with the help of generative fill I changed the clouds to more expressive ones.

And here, I added a lizard to the road by simply typing in the promt and selecting the desired area.

I also started running already finished renders through Stable Diffusion.

I simply put in a description of what is visible in the rendered promt, and add a standard negative promt for the realistic image.

Don't forget to load the image into the control net by selecting linearart in control type.

After generation, you can notice how parts of the picture have gotten better. Here, for example, I really like the detail of the white paint line.

And here's where I like the AI's rendering of the grass.

In photoshop through masks we show all the parts we like and get the final result.

Naturally such processing gives the visualization only plus one point, the main impression still depends on the overall picture that comes out of the program, but still it is always nice to improve your work at least a little, using new technologies.

The full version of the project can be found at the link: https://render.ru/ru/gallery/artwork/237314

I hope you found this article interesting, be sure to post your questions and thoughts in the comments. I wish you creative success and professional victories in your business!

Makary Lashin

Making of "Villa Toscana"