AMD FidelityFX Super Resolution 2.0 Source Code Published

[aufkrawall]

watched fully at 4k and i have played the damn thing, if the video dude don't zoom, freeze, capture specific sequences and loop them and give an explanation is damn fucking hard to notice

It has obvious breakage in a lot of scenes, no zoom needed to notice this.
They need to fix it badly, how it falls apart with disocclusion is kind of unique among TAA/TAAU solutions.

[neoe]

is better for 99% of the people on most common situations tho, nobody is going to play and zoom 3 freaking kms away just to see if a pixel is missing in the next block building, realistically speaking

for VR users , the image is "zoomed" , and every pixel make more sense. and also performance hungry. so such things are so needed by VR.

[jrch2k8]

It has obvious breakage in a lot of scenes, no zoom needed to notice this.
They need to fix it badly, how it falls apart with disocclusion is kind of unique among TAA/TAAU solutions.

again, not saying is perfect but i didn't notice them while normal gameplay or at least was not noticeable enough to affect me enough to notice it on gameplay.

Sure, if you walk around slowly, carefully checking every frame, stoping motion then moving violently i'm sure issue will appear and that is true as well for DLSS tho to a lesser extent depending on the game BUT most ppl won't do that.

Also i do understand there are ppl with eyes sensible enough to notice such thing like there are ppl with ear sensible enough to notice distortion on sound quality bitrate/equipment and i'm not saying just have to use FSR2. Just saying that for most user doing normal gameplay is barely noticeable and it is good enough.

It need improvements? i'm sure it does and will get better over time but for now is good enough to not require nVidia RTX class hardware to have decent upscaling for the average joe gamer, for those that need a bit more they can put the extra cash and go with DLSS and we all happy

[theriddick]

I wonder if the community will be able to build support into games themselves for older titles (using injection method of something).
This of cause will require analysis texture/vector data and such I know that. Maybe it can be done?

The thing that made FSR2.0 not as good as DLSS2.x was certain transparency scenarios as medium to far distance such as a character moving its hands around or a fence grating that has transparent sections.

Some games will look worse the others depending on how many instances of these transparent+animated+in-motion objects exist. I hope someday someone can put some code into FSR2.0 to fix this shortcoming....

Digital Foundry did a mini-deep dive into FSR2.0 (only 1 game) while back and pointed out these issues.

[smitty3268]

FSR 2.0 doesn’t prove anything. It has the same draw backs of all the other temporal upscalers I’ve seen. Only thing AMD did right was make it open source. While all the other upscalers are game engine specific. DLSS and other AI upscalers that utilize inference dedicated hardware like tensor cores/avx-512 or matrix cores will also run faster which is why we see a 5-8% increase using DLSS compared to FSR on the same GPU. Isn’t that the main purpose of using an upscaler? To gain more performance instead of buying more expensive hardware?

The performance differences between the 2 techniques are pretty minimal. NVidia even released a version of DLSS once that didn't use the tensor cores, so it's a pretty good bet it's not that deeply tied into using them. More likely NVidia just wanted to be able to be able to market the fact you needed to buy newer hardware from them in order to get any benefits.

[brucethemoose]

I wonder if the community will be able to build support into games themselves for older titles (using injection method of something).
This of cause will require analysis texture/vector data and such I know that. Maybe it can be done?

The thing that made FSR2.0 not as good as DLSS2.x was certain transparency scenarios as medium to far distance such as a character moving its hands around or a fence grating that has transparent sections.

Some games will look worse the others depending on how many instances of these transparent+animated+in-motion objects exist. I hope someday someone can put some code into FSR2.0 to fix this shortcoming....

Digital Foundry did a mini-deep dive into FSR2.0 (only 1 game) while back and pointed out these issues.

Motion vectors are not readily available like the depth buffer. AFAIK it would have to be modded into a game, rather than injected via something like ReShade or vkbasalt.

You can't efficiently generate them either.

[WannaBeOCer]

The performance differences between the 2 techniques are pretty minimal. NVidia even released a version of DLSS once that didn't use the tensor cores, so it's a pretty good bet it's not that deeply tied into using them. More likely NVidia just wanted to be able to be able to market the fact you needed to buy newer hardware from them in order to get any benefits.

Again the performance difference is between 5-8% between FSR and DLSS. I don’t use any of them since they both degrade graphics. The release of DLSS 1 didn’t utilize Tensor cores was horrible. When ever there was motion you could visually see the AI scrambling trying to restructure the image causing flickering. It’s the reason why everyone bashed DLSS when it was released.

https://youtu.be/YWIKzRhYZm4

[theriddick]

[QUOTE=WannaBeOCer;n1330256]
I don’t use any of them since they both degrade graphics. /QUOTE]

I think aspects are improved actually. But you could say in specific instances up-scaling can cause problems if not properly calibrated. But DLSS-2 is quite far along now and DLSS3.0 is probably not far off.

My next card was going to be a 4080 because I want DLSS and I run at 4k-120hz (and HDMI2.1) where it can actually be put to good use. However I'm still a bit concerned about the power consumption of NVIDIA's next cards, if they do chug 450w power in casual gaming examples then I may not bother!

[aufkrawall]

It need improvements? i'm sure it does and will get better over time but for now is good enough to not require nVidia RTX class hardware to have decent upscaling for the average joe gamer, for those that need a bit more they can put the extra cash and go with DLSS and we all happy

As I have no intention to bad-mouth FSR 2.0, I wish this were true. But as with other artifacts (e.g. those of DLSS), it's hard to un-see them once you are aware of the issue.
And in case of FSR 2.0, the issue is that screen regions that suddenly change, or were occluded before, look pixelated and noisy for long enough to be noticeable. This also applies to semi-transparent effects like ground fog seen in GoW, grass behind it looks clotted and noisy at the same time. This is far worse than the DLSS smearing issue with thin shapes in front of some contrasts/patterns (at least with recent 2.3.9/2.4.3, which GoW doesn't ship by itself). So it's a bit disappointing that after such a long wait, it's still kind of flawed.

We can be optimistic and hope that it gets improved soon. Though, it also took Nvidia long to make DLSS look good. I consider it enjoyable only since 2.2.6 version released about a year ago, previous 2.1.x versions were still quite ugly in terms of motion blur etc. I suspect the disocclusion issue of FSR 2.0 is similarly hard to fix/improve as is the DLSS smearing issue.

[tildearrow]

Typo.

This temporal uoscaling solution