I want to use my old renders again for advertising purposes. Better window reflections, better roads, sky, people and plants. I would love to create a working setup for this. There is a lot of potential in stable diffusion but it is a bit complicated. That's why your experiments on certain topics like this are very useful and inspiring. Revising high-resolution Render outputs seems a bit problematic to me. It is possible to do it in pieces, but areas such as the sky take up a large space in the picture and this 512x512 or 768x768 resolution limitation is a bit of a problem. But keep trying :) Thank you!
Thank you! For high-res renders, the best approach I have found so far is to split the image into smaller pieces manually, upscale each section individually using Img2Img in Stable Diffusion (with low denoising strength), and then stitch them back together in Photoshop (or another tool). This avoids memory issues that can cause crashes. The Img2Img script 'Ultimate SD Upscale' can also be used for upscaling directly in SD. Good luck with your project!
Couldn't get the Photoshop plugin to work with Automatic1111! Tried both the one-click installer and the unzip method, but no luck. Since it hasn't been updated in over a year, it might not be compatible with Photoshop 2025 anymore. It seems like the Photoshop-SD plugin now primarily works through ComfyUI, but I personally don't find ComfyUI very intuitive for workflow. Cheers @mae2309!
Yes! I'm working on something less 'nerdy' and more approachable right now. AI in architecture is a huge topic, so I’ll be covering it step by step-bite by bite. Stay tuned! Cheers, @oguzhanuzun7028!
Support the channel: paypal.me/ArchViz007/50 | buymeacoffee.com/ArchViz007 - Thanks!
I want to use my old renders again for advertising purposes. Better window reflections, better roads, sky, people and plants. I would love to create a working setup for this. There is a lot of potential in stable diffusion but it is a bit complicated. That's why your experiments on certain topics like this are very useful and inspiring. Revising high-resolution Render outputs seems a bit problematic to me. It is possible to do it in pieces, but areas such as the sky take up a large space in the picture and this 512x512 or 768x768 resolution limitation is a bit of a problem. But keep trying :) Thank you!
Thank you! For high-res renders, the best approach I have found so far is to split the image into smaller pieces manually, upscale each section individually using Img2Img in Stable Diffusion (with low denoising strength), and then stitch them back together in Photoshop (or another tool). This avoids memory issues that can cause crashes. The Img2Img script 'Ultimate SD Upscale' can also be used for upscaling directly in SD. Good luck with your project!
isnt better to do the stable diffusion photoshop plugin? thats what I use. with out the scaling dimension.. cheers
Couldn't get the Photoshop plugin to work with Automatic1111! Tried both the one-click installer and the unzip method, but no luck. Since it hasn't been updated in over a year, it might not be compatible with Photoshop 2025 anymore. It seems like the Photoshop-SD plugin now primarily works through ComfyUI, but I personally don't find ComfyUI very intuitive for workflow. Cheers @mae2309!
can you make a complete beginner ai architecture tutorial video?
Yes! I'm working on something less 'nerdy' and more approachable right now. AI in architecture is a huge topic, so I’ll be covering it step by step-bite by bite. Stay tuned! Cheers, @oguzhanuzun7028!