[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


Imagining and upscaling images is faster than constructing them.
>>
>>107813595
I have a 5090 and never turned on RTX or DLSS/FG, I only play games released before 2010.
>>
>>107813595
Approximating is faster than rendering, nyes.

It's all marketing slop. We should be getting better graphics but instead with have to paye 10 times more to render blurry, ghostery, smudgy, hallucinated slop made with 10 times less pixels than 10 years ago.
>>
>>107813625
why did you buy a 5090 then?
>>
>>107813595
>9 out of 10 frames are fake
there is no way it looks passable in motion
>>
>>107813641
The most accurate question. You can play older games with an rx 580
>>
>>107813625
>buy 5090
>never use more than 1% of it
Thank you for your service, goysoomer.
>>
>>107813641
>>107813663
>>107813741
Playing OG Crysis at 8k 480 FPS.
>>
>>107813648
That's what they said about 1-frame generation too.
And then they announced 3-frame generation and again they said there's no way it would pass, yet it looks perfectly fine.
>>
>>107813595
>Yeah bro just simulate every single particle render ever single vertice 60/144/270 times each second
It's surprising that no one came up with some interpolation technique before
>>
File: 1764637417282363.jpg (1.71 MB, 3408x2556)
1.71 MB
1.71 MB JPG
>>107813756

you payed 3000 USD to play a 20 year old game?

fucking tard
>>
>>107813770
It looks alright if you train yourself to ignore artifacts and pretend like they're part of the game and sign Nvidia hymns.
>>
>>107813648
>>107813770
Stationary it looks okay, motion blur and shitty screens hide a lot in motion.
The problem with FG is latency, upscaling alone doesn't introduce latency.
>>
>>107813756
>8k 480 FPS
>still doesn't look better than or equal to real lifr
>>
>>107813641
Why not? You still get more frames and resolution in older games
You don't need the dlss
>>
>>107813806
>upscaling alone doesn't introduce latency.
extra latency* that is
>>
>>107813641
For Blender things, mostly, also playing @ 12k & 14k resolution on a 8k screen just makes all older games cum-tier, shit looks so sharp it could brit milah your foreskin.

>>107813741
I used 100% of it, though.
>>
>>107813595
Not paying for your fake frames 2% performance improvement for a 30% price increase, Njudea.
I'm still on a 1080 ti and will run it till it's dead.
>>
>>107813806
I don't feel any latency when using it desu.
If I remember right, they even keep reading your mouse movement while it is generating the fake frames, and adjusts the frames depending on how you move the camera with your mouse.
Of course this is a machine learning approximation that uses depth information etc. to guess the camera movement, but it does feel pretty instant and I don't notice a difference.
>>
>>107813806
>The problem with FG is latency
This
>>
>>107813625
I got mine to play VR and none of those techs work in VR.
>>
File: 1767964883639632.png (1.75 MB, 1920x1080)
1.75 MB
1.75 MB PNG
Truly impressive technology.
>>
>>107813663
>You can play older games with an rx 580

You can play them with a 8600G even.
>>
>>107813806
>>107813831
Also let me add:
I DO feel a lot of altency when using this "lossless scaling" scam on Steam, or the AMD version.
But with native DLSS, it's no problem at all.
For context, I play on a 5120x1440 screen with 240 Hz, so I'm the target audience that needs these fake frames to max out their screen.
>>
>>107813831
But there is extra latency, Nvidia has even a tool to measure PCL.
They try to hide it with saying DLSS scaling will benefit the lantecy more then FG will increase it, but it's still bullshit and FG and even native is still higher latency then just DLSS upscaling.

>>107813851
Lossless Scaling is even worse in every way, even their upscaling had DLSS/FSR FG levels of latency increase. Absolutely useless outside of 1% niche use cases (i.e. Intel iGPU and running a game with hardcoded res and framerate on a modern display).
>>
>>107813851
>For context, I play on a 5120x1440 screen with 240 Hz, so I'm the target audience that needs these fake frames to max out their screen.
You'd probably have a better experience just using VRR at whatever FPS it runs without FG.
>>
>>107813880
It's not like I haven't tried that. FG still feels much better to me.
I can run like 80 FPS without FG and it looks choppy to the eyes. There's no tearing but I can tell it feels low-fps. Then with framegen I can get to 240 and it feels silky smooth. I don't notice an increase in latency, and the artifacts are minimal at best (only notice them when really looking for them).
>>
>>107813880
>>107813929
Also, VRR causes flicker with the background lighting of the screen, which is a much bigger issue, that's why I never turn VRR on anymore even if I don't use framegen
>>
>>107813945
That's a monitor issue. There's even OLEDs that don't' flicker or flicker in very specific scenarios (like low FPS that swinging heavily) while doing VRR.



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.