The DF team were invited to GTC where the tech was unveiled and their first video was gushing over the technology while repeating Nvidia’s claims that the tech only modified lighting though there now seem to be some legitimate doubts about that. Also, they discuss receiving death threats over their initial positive coverage.
I have to wonder if they, like me, were considering the potential rather than what was immediately in front of them.
DLSS5 is, after all, an in-development project that currently runs on a second dedicated graphics card. Showing this off the way they did might have been a huge mistake, or maybe this reaction will course-correct what might have actually become a slop filter.
Lighting and shadow some are the most computationally intensive aspects of a game, they’re usually the first settings I dial back when I need to improve performance.
The possibility of running lower settings and having their visual fidelity improved as part of the upscaling pipeline is an appealing one for me. The key would be tuning it so that it can be an enhancement to what’s already there, not something that totally overrides it or makes it look like Gen-AI output.
In their first video they were openly gushing over the demos from Nvidia. That’s not “considering the potential.” They were straight up saying that it looked great.
I see - I don’t do well with the video format, I only read their article.