I'm excited for new DLSS models (a new type of model, not just incremental refinement of the training data - much room for advancements but also maybe regressions). But really I want to know if the RTX Mega Geometry hype is real. Nanite for RT geometry rather than requiring fallback meshes? That sound huge.
Looks like we get some early examples of how it works and all the SDK releases around the end of the month. https://www.nvidia.com/en-us/geforce/news/dlss4-multi-frame-generation-ray-tracing-rtx-games/
=> More informations about this toot | View the thread | More toots from shivoa@mastodon.gamedev.place
text/geminiThis content has been proxied by September (ba2dc).