Fidelity Super Resolution in 2021

[Rroff]

We can say whatever we want but adding FSR to DOTA2 was a great idea from AMD, it is like they hired Jensen to make their strategy. That game has tens of times more users than any DLSS game at any time. And FSR will make a lot of them happy.

Personally I'll only accept any of these measures DLSS, FSR, TAAU whatever as a compromise for if a game implements a decent ray tracing model.

It's really simple, it's res scale vs FSR, both 50%.

He corrected himself later they were done at different render scale percentages and he messed up something else as well but doesn't say what and re-uploaded the images.

End of the day the context it is being portrayed in is misleading as it makes it look like FSR on produces a better image than native when actually it is showing off in-game resolution scaling (probably bilinear) vs FSR.

 

[kjaerligkatt]

Personally I'll only accept any of these measures DLSS, FSR, TAAU whatever as a compromise for if a game implements a decent ray tracing model.

Agreed, but that’s sort of stating the obvious. These techniques are all offering a compromise for some gain, and the trade-offs are a value proposition that will be specific to each individual. For some maybe it’s trying to run a less expensive/power demanding system, rather than the behemoth that allows anything to run at max res/quality, or for others trying to squeeze extra performance from their existing system by sacrificing a little image quality (especially in the current market conditions).

 

[ICDP]

Uh-huh, so what did Alex say that makes him an outlier? And why is that specific thing wrong?



But all you're doing is asserting it's negative without evidence. OK, I assert that it's positive without evidence. See how stupid this is?

How about if you believe that something is negative you give an example of why you think it's negative? I've asked this over and over, so far people have largely just stated objectively wrong things that I've literally just checked and verified are wrong.

So again this is just deflection and pivoting away from the point which is, provide some examples of why you think the review was "negative". I'm not interested in your personal emotional response and how you personally cope (or don't cope) with facts which you don't find palatable. That's just emotions occurring in your brain which have little or no relevance to the actual facts. You're confusing your own emotional reaction with the objective nature of the review. To be objectively negative in any way you'd have to have examples of where Alex was saying something verifiably incorrect, that makes the feature look bad where if he was honest it would look good. And so far no one has done that, people have tried but if what they're saying is wrong then the evidence doesn't support the claim.

I'm just asking for examples that support the claim, if what you're saying is true then this should not be hard.



ALL of the review sites regarding the 3080 launch (preview) were under strict NDAs about the specifics of what they could review, what games, what settings and what precise conditions. Those are conditions that Nvidia imposed on review sites if they wanted to get early access to the hardware in order to get reviews out for the launch day. DF are just being transparent about this to the end user so they have context for the review. It actually demonstrates open and honest engagement with users, they have zero control over the NDA Nvidia impose, they either agree or don't get early samples, the same as all other review sites.



Source?

I quite literally listed my evidence as to why DF have a poor reputation for being Nvdia paid PR and as such this makes whatever they say suspect. That is my evidence and it is enough for me and for most others to decide DFs FSR outlier review is not trustworthy.

Evidence and reasons have been posted to you mulitple times now and you keep ignoring it. That is your cognitive dissonance issue if you keep dismissing valid evidence.

 

[nvidiamd]

Personally I'll only accept any of these measures DLSS, FSR, TAAU whatever as a compromise for if a game implements a decent ray tracing model.

Yes but you are not the target, at least now in the beginning. If you read on reddit you will see a lot of people happy they can play games on integrated gpu's or old Nvidia cards. Especially on esports titles a feature like FSR is a blessing. And since it is lightweight, it is better than anything else because it has lower lag. This is how you build hype when you don't have a ton of money.

 

[Hotwired]

And zen you see this:
https://www.reddit.com/r/Amd/comments/o6skjq/digital_foundry_made_a_critical_mistake_with/

How awkward, DF wrote an entire paragraph using it as evidence that TAAU was much better than FSR and gets checked by reddit.

I tested this using Kingshunt, an Unreal Engine 4 game. I compared native 1080p (with something like a simple bilinear upscale to 4K) up against AMD FSR in performance mode, which is also working with an internal 1080p framebuffer. Looking at in-surface detail, there isn't much difference and sub-pixel detail doesn't scale particularly well. Edge detail, though, is much, much better. This is what FSR's design is about, not inner surface quality or aliasing strictly, but identifying edges and not blindly blending between them like a simple upscaler - it preserves an edge well. So FSR is better than a standard upscale. The thing is, games do not just use basic scalers - Unreal Engine 4 has a temporal AA upscaler in it. Thin transparent detail resolves to a much greater degree, while the quality inside surfaces - texture detail - is also elevated beyond the 1080p standard. This happens because TAAU (as it is called) is sampling previous frames into the 4K presentation, adding detail. And by default, edge enhancement comes as part of the same algorithm.

Aight DF time to walk it back about TAAU, that blur for the FSR images is the beta game you chose to test with being a beta game.

 

[StephenJ]

Looking forward to trying this when it gets added to more games

 

[RBImGuy]

And zen you see this:
https://www.reddit.com/r/Amd/comments/o6skjq/digital_foundry_made_a_critical_mistake_with/
How was Alex shilling? He is the best upscaling expert you can find on internet.

Cant trust DF anymore, it is what happens when your not skilled enough about the subject and paid by nvidia to do reviews it infects them perceptually what and how their judgement works.
Misleading people, 117803 showcases atm is really bad

 

[timorous]

How awkward, DF wrote an entire paragraph using it as evidence that TAAU was much better than FSR and gets checked by reddit.

Aight DF time to walk it back about TAAU, that blur for the FSR images is the beta game you chose to test with being a beta game.

Alex responded to the thread. Admitted the mistake in Kingshunt but doubled down on their conclusion by linking to some images he had on twitter doing the same comparison in several Godfall images.

Problem is they are still testing 50% TAAU with FSR Performance rather than testing the full gamut of settings to see if there is a crossover point. As already pointed out KitGuru tested FSR UQ vs 77% TAAU (so input res was the same) and found at this setting UQ was slightly ahead of TAAU due to the extra shimmering with TAAU in motion. So their TAAU > FSR conclusion seems invalid for the data they have provided. OTOH if they simply said for 50% render scales TAAU is better than FSR then their images back up that statement but they took that sample and generalised all the way up which is just really really dumb.

 

[Chuk_Chuk]

How awkward, DF wrote an entire paragraph using it as evidence that TAAU was much better than FSR and gets checked by reddit.



Aight DF time to walk it back about TAAU, that blur for the FSR images is the beta game you chose to test with being a beta game.

Ouch, their reputation is taking knock after knock.

 

[Hotwired]

Alex responded to the thread. Admitted the mistake in Kingshunt but doubled down on their conclusion by linking to some images he had on twitter doing the same comparison in several Godfall images.

Problem is they are still testing 50% TAAU with FSR Performance rather than testing the full gamut of settings to see if there is a crossover point. As already pointed out KitGuru tested FSR UQ vs 77% TAAU (so input res was the same) and found at this setting UQ was slightly ahead of TAAU due to the extra shimmering with TAAU in motion. So their TAAU > FSR conclusion seems invalid for the data they have provided. OTOH if they simply said for 50% render scales TAAU is better than FSR then their images back up that statement but they took that sample and generalised all the way up which is just really really dumb.

Yes, wonder what he's going to do about his video with extensive waffle of how TAAU was doing the job better than performance FSR now his data is proven to be false.

Leave it up?

 

[LtMatt]

Ouch, their reputation is taking knock after knock.

You love to see it.

 

[LtMatt]

The family of game titles supporting FSR has grown: enter the Nemestice + Battle Pass update for Dota 2, featuring AMD FidelityFX Super Resolution.

Nemestice Falls and New Powers Rise (dota2.com)

Image comparison
1440p FSR Dota 2 + RIS Test - Imgsli

 

[timorous]

Yes, wonder what he's going to do about his video with extensive waffle of how TAAU was doing the job better than performance FSR now his data is proven to be false.

Leave it up?

Obviously, who reads reddit anyway. Especially hardware and AMD since the post was removed from the NV and PC Gaming subs.

Nemestice Falls and New Powers Rise (dota2.com)

Image comparison
1440p FSR Dota 2 + RIS Test - Imgsli

I really like that there is an FSR scale slider so you don't need to use the presets. From what I have seen on reddit 90% FSR and above seems to offer a better than native appearance (probably due to the sharpening pass so I expect Native + sharpening would be better still) while provided a small performance bump. Pretty impressive.

 

[nvidiamd]

He is the ultimate shill, digging his own grave.
"Doing that does not at all change the conclusion our DF coverage - as if you look at the detail without DOF anyway, it is pretty easy to see how much better TAA U is."
Really if you look at the pics that guy posted, FSR wins by a landslide.
" I will update our article to reflect this and change the game comparison for TAA U from Kingshunt to Godfall which does not have DOF aperture affected by upsampling."
So because i was disproven on Kingshunt, i will remove it from the article and focus on Godfall since no one called me a liar on Godfall yet.

 

[nvidiamd]

"Something we also tested in GodFall but just did not include in the video. Images of God fall below at my twitter" Alex
https://twitter.com/Dachsjaeger/status/1407398823706701832

"You should be able to enable TAAU in all newer UE4 games by doing this: https://reddit.com/r/lowspecgamer/comments/i8bqhy/temporal_upsampling/… So you could compare it in Godfall as well, and also control the res it uses the match the performance/quality of FSR" Someone's tweet to Alex

"OMG - this would have been amazing for the video. I tried to get TAA Upsampling in GodFall using
@FransBouma
's amazing Unreal Engine Unlocker, but that game does not work completely with it. Thanks for the tip for the future! I will make sure to use this again." Alex answering two days ago.

So did he had any tests from Godfall when he made the video or he made them this morning when he noticed he was called a shill on Twitter?

 

[timorous]

And he is doubling down even further.

The exact same benefits are the same at all internal resolutions because TAA Upsampling is a temporal reconstruction solution and by its very nature will be surperior. It is actually adding in detail that a native image would have also had. FSR is taking an already aliased, mipped, and filtered lower resolution image and cannot generate detail, subpixel detail, or anything that that lower resolution image did not have. Sharpening or weighted edge enhancement is not generating new information in the same way, especially image sharpening which is just increasing local contrast, not adding in detail.

They are 2 very different things.

Does not even comment on ghosting or shimmering that temporal solutions can introduce and that can be far more distracting than a softer image or missing super fine details.

It really looks like he had a conclusion before starting and went picking to find comparisons that would lead to that conclusion.

 

[Hotwired]

I'd have thought that a professional reviewer would have noticed that FSR looked much like native while TAAU looked vastly better than both.

This is what he was looking at, taken from the comparison images from his review:

 

[NinjaCool]

Digital Foundry made a mistake in their SFR review and its good to see that they at least admit it but its weird their SFR conclusion will remain 100% the same despite this.

 

[LtMatt]

He has to double down now, no other choice. Live by the sword, die by it.

I don't understand why you would not use Ultra Quality in Godfall? 50% performance increase for near native resolution quality.

Godfall
4K Epic Settings (CAS off)
Native vs Ultra Quality FSR

FSR Native v Ultra Quality - Imgsli

What is not like to like about the 50% improvement in FPS? Please tell me.

 

[fs123]

Obviously, who reads reddit anyway. Especially hardware and AMD since the post was removed from the NV and PC Gaming subs.



I really like that there is an FSR scale slider so you don't need to use the presets. From what I have seen on reddit 90% FSR and above seems to offer a better than native appearance (probably due to the sharpening pass so I expect Native + sharpening would be better still) while provided a small performance bump. Pretty impressive.

Unless they have mixed up the sides, the FSR is indeed better looking in the DOTA comparison. Native also has the same 60% of RIS sharpening applied so it is a fair comparison.

https://imgsli.com/NTg3MTE

As I expected FSR looks better with additional sharpening compared to the default FCAS AMD launched it with.