HoloDust on 07 September 2026
So far, to me at least, faces pretty much always seem to look like AI slop with DLSS5. Environments are hit and miss. I find grading to be, in majority of cases, terrible compared to original image, like what modern Hollywood is pushing in last decade or so, making image look cheaper. Performance hit is insane. Yet, combined pipeline of RT and neural rendering is probably the future (at least near), unless some major shift happens (which it just might, if something like what Bolt is doing comes to fruition).
Last edited by HoloDust - on 07 September 2026
