https://jms55.github.io/posts/2025-12-27-solari-bevy-0-18/

[giscus[bot]]

https://jms55.github.io/posts/2025-12-27-solari-bevy-0-18/

Bevy, Rust, Graphics, etc

https://jms55.github.io/posts/2025-12-27-solari-bevy-0-18/

[klownie]

this is amazing. cant wait to experiment with it

[shanayyy]

Although I don't understand most of it, the result is fantastic.

[JMS55]

If you have questions, I'm happy to answer them!

Also, https://intro-to-restir.cwyman.org is a great resource for understanding ReSTIR.

[NoAuxCordAudi]

I've been playing around with this, and I'm running into an issue where some pixels come out with a drastically wrong color relative to their neighboring pixels. Most of them appear white, so they look like speckles in the rendered image. Do you have any suggestions, or would you need more information to provide a recommendation?

[JMS55]

By "this", do you mean bevy_solari? If so, awesome, glad you're giving it a try!

Sounds like it could be an issue with NaNs, but I'm not entirely sure. I would need the scene to be able to reproduce it.

Please reach out on discord https://discord.com/invite/bevy or open a github issue https://github.com/bevyengine/bevy/issues with more details!

[Vecvec]

• Markov chain resampling

I've found this improved my ray tracer quite impressively (though I used its predecessor for DI instead of restir DI), and so would highly recommend it (I haven't tried the others, so I'm sure they also give good improvements).

[JMS55]

Good to know! I'm experimenting with some other methods for improving DI/GI initial sampling - we'll see how it goes! Maybe I'll try markov chains too.

[Vecvec]

Good luck! I'm excited to see where solari goes next.