I haven’t seen any negative response to DLSS5 except by people who literally only see timeless, unerring holy art when they look at anything. IDK who thinks making games visually enhanced over the artist’ limited tools/time/abilities is a deterioration.
If it actually looked good sure but it just cranks up the contrast and puts a shitty Instagram filter over the characters. I don’t care how delusional you are I just can’t understand how you’d think that wouldn’t ruin whatever the character is supposed to actually look like. In one of their own examples the character just had a family member die, they aren’t wearing makeup for a reason, not that DLSS cares so yassified grieving woman go brrr
Referring to the slop filter as visual enhancement is being very generous. The AI doesn’t know shit about why things look the way they look. Can’t wait to play Deep Rock Galactic but with the lighting all fucked up, despite that being an extremely important game mechanic that the AI doesn’t know anything about.
Referring to technological progress as slop is being very predictably monotonous. AI isn’t exclusively a black box. There are tunings and imposed considerations that can be made so the AI DOES know about anything it needs to. The technology is only in preview form, and you seemingly expect immediate perfection, rather than incremental progress.
Honestly people are making a big deal out of this but I don’t get it, the time for the outage was when both AMD (2023) and Nvidia (2022) introduced frame generation into their respective upscalers. It was always gonna be downhill from there.
I mean, lots of people were and still are annoyed by frame gen too - especially with the deceptive way nvidia uses it in their marketing (claiming a 5070 will be as fast as a 4090 when it actually just makes more fake frames).
was dlss ever worth it? It just seemed to make games run worse without any significant improvement to fidelity.
Absolutely but it depends on what kind of DLSS we’re talking about, since nvidia uses the term to talk about multiple very different technologies. DLSS framegen can make things look smoother but increases latency and introduces visual artifacts. I would never turn it on for something like a first-person shooter, but I could maybe see myself using it in a game like Civ VII where a few extra milliseconds of latency isn’t a big deal.
The really important one though is DLSS upscaling. Back in the day if you ran a game at lower resolution (usually for performance reasons) and upscaled it it would look like shit. DLSS upscaling lets you do that and have it look almost as good as rendering it natively, as long as you don’t push it too far. It didn’t work super well when it was first released, but these days it looks great and can really improve performance.
The benefits of DLSS yassification are questionable.
I haven’t seen any positive response to DLSS5 except by people who literally only see shapes and colors when they look at art. IDK who thinks making random parts of scenes visually louder instead of the way the artist intended is an improvement.
Jensen Huang had to come out and tell people to stop making fun of DLSS5. AI isn’t going to fall has hard as NFT’s but it’s certainly losing popularity.
