-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
DLSS? No, TecoGAN...
- Thread starter Wrinkly
- Start date
Just a sizzle reel showing off bits of a paper on AI based upsampling/super resolution coming from Facebook reality labs. Looks pretty decent in the examples shown, also featured video I recognised from an earlier Facebook reality labs research sneak peek.
Training could be done in a day and a half on a single Titan V, so low barrier compared to any need for an Nvidia super computer... that said I think DLSS2.0 already removed that specific training requirement.
Makes sense that FB would be looking at it, it’s one of those areas that VR could massively benefit from when pushing higher and higher resolution and higher refresh rate displays... even more so with the limitations of mobile hardware. I recall abrash showing a combination of deep learning based upsampling of sparesely rendered areas and foveated rendering leading to up to a 20x reduction in rendered pixels.
Edit - important just to point out though this is all still just research, not a finished product.
Last edited: