NVIDIA has developed a new compression algorithm for textures that looks very impressive.
VideoCardz spotted the paper from NVIDIA, which is called 'Random-Access Neural Compression of Material Textures,' and was also shared on Twitter by Bart Wronski, a researcher at Team Green who's part of the project.
The Neural Texture Compression (NTC) aims to tackle the spiraling requirements for high-res textures, allowing them to better fit into the available RAM. We're told that it delivers 16x more texels than Block Compression or BC (standard GPU-based texture compression).
You can see the quality of the results in the above tweet. In terms of speed, NVIDIA points out that GPUs with Tensor Cores aren't slow at the task - at least relatively speaking, at this early stage in the development of the tech.
In practical terms, though, a 4K render takes 1.15ms which isn't quick enough - compared to 0.49ms as measured on an RTX 4090 - so it'll need to be a good deal faster than this for NTC to work well enough to be usable.
Wronski admitted that:
"This requires future work, but it will only get better, and I'd argue that it is already in a practical realm."
So, given some time and refinement, this could be a key technology for achieving better textures. Or rather, fitting high-res textures into more modest VRAM loadouts (something of a hot topic right now, with the impending launch of the RTX 4060 Ti and RX 7600 with their rumored 8GB configuration in both cases).
We'll learn more about NTC and how NVIDIA plans to develop this tech at SIGGRAPH 2023 in August. It won't be riding to the rescue of current-gen GPUs, though, that's for sure - and it might be some way down the line before we see it realized. Mind you, Wronski sounds confident enough that it's maybe not as far off as we think...