[EastCoast]

When last we meet we discussed how DF used a game called Kingshunt to compare FSR using an app for the unreal engine called unreal unlocker to enable TAAU. Giving TAAU the impression that it was better then FSR. However, based on the investigation by the original poster TAAU disabled something called DepthOfFiedlQuality. Which sharpens the image. Thus giving the illusion that TAAU is "better then native". When in fact it is not. When DepthOfFieldQuality is disabled and compared, properly, FSR is superior to TAAU!!!!

Link: [Reddit] Digital Foundry made a critical mistake with...

This time, it appears that something similiar is happening with their review of FSR 2.0 IQ:
1. Using Amd's competitor gpu
2. Using DX11 game instead of using a DX12 game with FSR 2.0
3. Altered the default sharpening setting of 0.3. to 0.
4. The implementation of fsr 2.0 in GOW was a custom in house job with no involvement from AMD.
5. The blurriness/ghosting shown in motion is looking to be a results of a combination of Depth of Field and TAA. Modded versions of GoW don't exhibit this behavior.

In which they are claiming that FSR 2.0 can be inferior to TAA used in the game. By showing ghosting and other IQ imperfections. However, there were some odd caveats emerged. This is looking, once again, like an issue with how the game uses DoF and TAA just as found in the prior review using Kingshunt to compare FSR 1.

JuxtaposeJS Embed

JuxtaposeJS Embed

cdn.knightlab.com

JuxtaposeJS Embed

JuxtaposeJS Embed

cdn.knightlab.com

Original Vs TAA Off

JuxtaposeJS Embed

JuxtaposeJS Embed

cdn.knightlab.com

Original Vs TAA/Sharpening/DoF/TAA/Vignette Off

(Links above could take 60 seconds or more to fully download. If not refresh the page)
Original Link Here

Below you will find the DF video. You can skip ahead to the 13:00 mark where comparison videos, using a 6900xt, contrast DF findings of ghosting. In which no ghosting was found in the water nor the snow as found in the DF video.

Digital Foundry Review of FSR 2.0 Vs DLSS (newest version).
Skip Ahead to 13:00 mark to compare the water/snow to the videos below.

As you can see there is no ghosting found using FSR 2.0 using a Radeon RDNA 2 video card. It's inconclusive if that's the result of tweaking of DoF as found in Kingshunt. IQ differences between FSR 2.0 used on RDNA 2 vs RTX.

Edit:
It was brought to my attention that Alex was interviewed about a week before publishing his FSR 2.0 vs Dlss IQ review. Where he was told that FSR 2.0 was not compatible with dx11.

Journalistic integrity comes into play when this kind of information is known. As this insight should have been a disclaimer to his review. And made plain to see in the title and video thumb nail. But it was not.

There is no mention of this in the review in the title nor thumbnail. IMHO this invalidates the IQ review entirely and wholisticaly. Integrity is the key here. Just as the user jon provided me the video showing Alex being told about FSR 2.0 not being compatible with dx11. So should the disclosure of this be made available in his review. In a clear, precise manner that sets the tone of the review as a more acedemic then factual endeavor

52:42


Edit 2:
Original

Original

VS

AA/Sharpening/DoF/TAA/Vignette Off

AA/Sharpening/DoF/TAA/Vignette Off

 

[UltraMega]

I watched this df video. Didn't seem like they said TAAU was better than FSR 2.0 to me. The take away I got from it was that it goes DLSS, FRS2, then TAAU as far as quality goes. Seems like they have mostly good things to say about FSR2, but they do correctly point out that DLSS is still superior though to a much more narrow extent vs FSR1.

 

[Imglidinhere]

I mean... the fact that FSR 2.0 works with either team's GPUs is why it's better imo.

 

[HyperMatrix]

I mean... the fact that FSR 2.0 works with either team's GPUs is why it's better imo.

From the perspective of the 80% market share of new GPU sales, which is Nvidia, I have to disagree with you. FSR 2.0 isn’t better than DLSS 2.0 due to wider availability. The wider availability with regard to compatible hardware is good for those people who previously had no other option available to them. That’s as far as that goes. For example, FSR 1.0 was also available on all hardware. That didn’t make it better than DLSS. In fact, even in games that only supported FSR 1.0 (like far cry 6) I didn’t use it. Because the fact that it exists and the fact that it works on my card, doesn’t mean it’s worthwhile. I also didn’t use DLSS 1.0 because of the image quality issues.

However, it’s good that FSR 2.0 exists as it requires the same type of work that would go into implementing DLSS or XESS. So hopefully it will result in more developers integrating all technologies into their games.

 

[EastCoast]

I really do wonder if DoF was tweaked again? Or, something about the 3090/drivers they used that causes a conflict?

 

[EastCoast]

Another video
Taken from the 7:35 mark of the DF video. No shimmering found as the axe is thrown

 

[UltraMega]

From the perspective of the 80% market share of new GPU sales, which is Nvidia, I have to disagree with you. FSR 2.0 isn’t better than DLSS 2.0 due to wider availability. The wider availability with regard to compatible hardware is good for those people who previously had no other option available to them. That’s as far as that goes. For example, FSR 1.0 was also available on all hardware. That didn’t make it better than DLSS. In fact, even in games that only supported FSR 1.0 (like far cry 6) I didn’t use it. Because the fact that it exists and the fact that it works on my card, doesn’t mean it’s worthwhile. I also didn’t use DLSS 1.0 because of the image quality issues.

However, it’s good that FSR 2.0 exists as it requires the same type of work that would go into implementing DLSS or XESS. So hopefully it will result in more developers integrating all technologies into their games.

From the perspective of usability, there are way more GPUs out there that can run FSR vs DLSS since only the rtx Nvidia cards can do DLSS but older Nvidia cards can do FSR. The geforce 10 series is still by far the most common GPUs out there.

I think the differences between FSR2 and DLSS are definitely clear in a break down video like DF provides but at the same time they both seem good enough that in normal gameplay the differences would be very hard to notice. FSR doesn't ever have to be fully on par with DLSS, it just has to be close enough that the differences are negligible and we're probably already at that point.

 

[UltraMega]

Just to add some more thoughts; one thing that is clear is that FSR lacks in it's ability to take full advantage of temporal information and additionally lacks from not being supported by any machine learning, and I think DF is correctly pointing this out while still offering good praise for what FSR is able to accomplish.

When you see examples of things moving quickly across the screen and leaving a sort of trail of less detailed visuals behind moving objects, that happens because FSR can only reconstruct from what is on the screen. It takes a few frames for FSR to be able to catch up and re-reconstruct visual data in these situations. I think DLSS essentially works in the same way but with two main differences. One is that it has dedicated hardware support so it's able to reconstruct faster, and the other is that is relies on machine learning templates to help mitigate this issue.

FSR may be able to improve on this with updates, but until there is more of a hardware and/or machine learning aspect added in, I expect lower quality trails following fast moving object will continue to be one of the main differences between FSR and DLSS.

All and all I'm definitely impressed that FSR is able to get so close to DLSS without the same kind of extras that DLSS relies on.

 

[HyperMatrix]

From the perspective of usability, there are way more GPUs out there that can run FSR vs DLSS since only the rtx Nvidia cards can do DLSS but older Nvidia cards can do FSR. The geforce 10 series is still by far the most common GPUs out there.

I think the differences between FSR2 and DLSS are definitely clear in a break down video like DF provides but at the same time they both seem good enough that in normal gameplay the diffences would be very hard to notice. FSR doesn't ever have to be fully on par with DLSS, it just has to be close enough that the differences are negligible and we're probably already at that point.

83% of discrete graphics sales are Nvidia at the moment. The last Nvidia generation to launch without RT/Tensor cores was released over 6 years ago. The projected performance increase on the 4000 series are 2x over 3000 series. The RTX 3060 at $329 is already roughly the same as a 1080Ti in rasterization. By the end of this year, you should be able to get 50% higher performance than 1080Ti at around $250. My contention is that people who are running $200 or lower GPUs, aren't going to be the biggest spenders on new titles coming out.

I'm genuinely happy that FSR 2.0 is available as it is far better than FSR 1.0 which was absolute garbage. But just because something is commonly available, doesn't make it better at the function it's supposed to perform. At this point we're literally advocating for people who are running PCs far less powerful than consoles. And while I'm happy that they're getting something that will increase their performance, the lowest common denominator shouldn't be determining what is and isn't good. By that logic, you could make the argument that pre-baked lighting is better than ray tracing, because it's far more accessible. But I'm sure you'd agree that we wouldn't want to go down that road.

The entire point of upscalers, which have always existed in various forms, is to get "free performance" so to speak. Their entire worth is based on how close to native resolution they can get. If you're getting a noticable drop in image quality, how's that different from just dropping quality settings or lowering resolution? And if FSR 2.0 is better than DLSS 2.0 because of reach/availability, wouldn't that make FSR 1.0 better than FSR 2.0 since it can be implemented in so many more games? And how about when AMD's next gen GPUs come out with a new hardware accelerated/ai version of FSR that works similarly to DLSS 2.0? Would FSR 2.0 be better than the new FSR at that point because it can run on more cards? I clearly say no. The best upscaler is the one that provides the best image quality, as that is its only job. And the better it does that job, the more value it has. Not everybody having access to it, doesn't make it not the best.

 

[Imglidinhere]

From the perspective of the 80% market share of new GPU sales, which is Nvidia, I have to disagree with you. FSR 2.0 isn’t better than DLSS 2.0 due to wider availability. The wider availability with regard to compatible hardware is good for those people who previously had no other option available to them. That’s as far as that goes. For example, FSR 1.0 was also available on all hardware. That didn’t make it better than DLSS. In fact, even in games that only supported FSR 1.0 (like far cry 6) I didn’t use it. Because the fact that it exists and the fact that it works on my card, doesn’t mean it’s worthwhile. I also didn’t use DLSS 1.0 because of the image quality issues.

However, it’s good that FSR 2.0 exists as it requires the same type of work that would go into implementing DLSS or XESS. So hopefully it will result in more developers integrating all technologies into their games.

I disagree entirely. Anything pre RTX series, cannot run DLSS, which immediately invalidates a whole host of potential technologies that would otherwise benefit. For people who have to play on medium/low settings, FSR 1.0 is actually pretty good. I would've loved that kind of feature in any game back when I was a kid.

83% of discrete graphics sales are Nvidia at the moment. The last Nvidia generation to launch without RT/Tensor cores was released over 6 years ago. The projected performance increase on the 4000 series are 2x over 3000 series. The RTX 3060 at $329 is already roughly the same as a 1080Ti in rasterization. By the end of this year, you should be able to get 50% higher performance than 1080Ti at around $250. My contention is that people who are running $200 or lower GPUs, aren't going to be the biggest spenders on new titles coming out.

I'm genuinely happy that FSR 2.0 is available as it is far better than FSR 1.0 which was absolute garbage. But just because something is commonly available, doesn't make it better at the function it's supposed to perform. At this point we're literally advocating for people who are running PCs far less powerful than consoles. And while I'm happy that they're getting something that will increase their performance, the lowest common denominator shouldn't be determining what is and isn't good. By that logic, you could make the argument that pre-baked lighting is better than ray tracing, because it's far more accessible. But I'm sure you'd agree that we wouldn't want to go down that road.

The entire point of upscalers, which have always existed in various forms, is to get "free performance" so to speak. Their entire worth is based on how close to native resolution they can get. If you're getting a noticable drop in image quality, how's that different from just dropping quality settings or lowering resolution? And if FSR 2.0 is better than DLSS 2.0 because of reach/availability, wouldn't that make FSR 1.0 better than FSR 2.0 since it can be implemented in so many more games? And how about when AMD's next gen GPUs come out with a new hardware accelerated/ai version of FSR that works similarly to DLSS 2.0? Would FSR 2.0 be better than the new FSR at that point because it can run on more cards? I clearly say no. The best upscaler is the one that provides the best image quality, as that is its only job. And the better it does that job, the more value it has. Not everybody having access to it, doesn't make it not the best.

Nvidia isn't going to manage doubling performance unless they crank the power draw, which is already projected to be absurd even compared to Ampere.

 

[UltraMega]

83% of discrete graphics sales are Nvidia at the moment. The last Nvidia generation to launch without RT/Tensor cores was released over 6 years ago. The projected performance increase on the 4000 series are 2x over 3000 series. The RTX 3060 at $329 is already roughly the same as a 1080Ti in rasterization. By the end of this year, you should be able to get 50% higher performance than 1080Ti at around $250. My contention is that people who are running $200 or lower GPUs, aren't going to be the biggest spenders on new titles coming out.

I'm genuinely happy that FSR 2.0 is available as it is far better than FSR 1.0 which was absolute garbage. But just because something is commonly available, doesn't make it better at the function it's supposed to perform. At this point we're literally advocating for people who are running PCs far less powerful than consoles. And while I'm happy that they're getting something that will increase their performance, the lowest common denominator shouldn't be determining what is and isn't good. By that logic, you could make the argument that pre-baked lighting is better than ray tracing, because it's far more accessible. But I'm sure you'd agree that we wouldn't want to go down that road.

The entire point of upscalers, which have always existed in various forms, is to get "free performance" so to speak. Their entire worth is based on how close to native resolution they can get. If you're getting a noticable drop in image quality, how's that different from just dropping quality settings or lowering resolution? And if FSR 2.0 is better than DLSS 2.0 because of reach/availability, wouldn't that make FSR 1.0 better than FSR 2.0 since it can be implemented in so many more games? And how about when AMD's next gen GPUs come out with a new hardware accelerated/ai version of FSR that works similarly to DLSS 2.0? Would FSR 2.0 be better than the new FSR at that point because it can run on more cards? I clearly say no. The best upscaler is the one that provides the best image quality, as that is its only job. And the better it does that job, the more value it has. Not everybody having access to it, doesn't make it not the best.

Saying one is "better" than the other is subjective. A good argument can be made that FSR is "better" because it works on a much larger percentage of GPUs that are actually out there in the market being used right now. A quick look at a steam hardware survey shows that non-RTX cards are still massively more prevalent than GPUs that can run DLSS. I don't think anyone is arguing that FSR is superior at image quality, but arguably better due to massively higher compatibility.

Steam Hardware & Software Survey (steampowered.com)

 

[HyperMatrix]

Nvidia isn't going to manage doubling performance unless they crank the power draw, which is already projected to be absurd even compared to Ampere.

Have you seen any of the leaks from verified sources? They're going from Samsung 8nm which was just a modified 10nm, to TSMC 5nm. Core count is up, as are clocks significantly. On top of that they've massively increased the cache. Not as much as AMD, but still many many times larger than current gen.

I disagree entirely. Anything pre RTX series, cannot run DLSS, which immediately invalidates a whole host of potential technologies that would otherwise benefit. For people who have to play on medium/low settings, FSR 1.0 is actually pretty good. I would've loved that kind of feature in any game back when I was a kid.

That doesn't make FSR better. It makes it available. As the only option. The only way one can be better than the other, is if you have a choice of which to use, and pick one over the other. "I'm happy it's available for low end GPUs" is not the same as "FSR is better than DLSS," which is what you claimed. And I quote:

"I mean... the fact that FSR 2.0 works with either team's GPUs is why it's better imo."

No. Not better. Accessible to more people. Not better. It is in fact worse. It has worse image quality, and worse performance on average. If you got to pick which to use, you'd pick DLSS. Because DLSS is "better."

Saying one is "better" than the other is subjective. A good argument can be made that FSR is "better" because it works on a much larger percentage of GPUs that are actually out there in the market being used right now. A quick look at a steam hardware survey shows that non-RTX cards are still massively more prevalent than GPUs that can run DLSS. I don't think anyone is arguing that FSR is superior at image quality, but arguably better due to massively higher compatibility.

No. The definition of better is very clear. FSR isn't "better" because it works on a larger percentage of GPUs. You can say FSR is more readily available. You can say it has better hardware support. Or that it provides access to better upscaling methods than is available to non-RTX cards. But you can't say FSR is better without changing the implication of the word. And to do so, you'd need to state it as such. For example:

• A Toyota Camry has better fuel mileage than a Ferrari
• A Toyota Camry has a better price than a Ferrari
• A Toyota Camry does a better job of transporting a family of 5
• A Toyota Camry is better when it comes to maintenance costs

But none of those things allow you to say a Toyota Camry is better than a Ferrari. When you compare similar things and use the general term "better," it means you're comparing them based on how well they do what they're supposed to do. The Ferrari does almost everything a car is supposed to do a lot better. The fact that the price puts it out of the reach of most people doesn't make the Toyota a better car. It just makes it a more affordable and attainable car. And as a side note...generally more attainable items are considered "less good" based on human psychology...but let's not get into that.

So what are DLSS and FSR? They're upscaling techniques. They're supposed to improve performance, and maintain image quality. That is their only job. Which provides the best image quality? DLSS. Which provides the best performance improvement on average? DLSS. Therefore, you can't say that FSR is better than DLSS unless you're disputing those claims. Do you happen to like FSR more because it's available to more people? That's a fair statement and a fair reason to like and support it. You can say I think AMD is better than Nvidia for releasing FSR and making it available to everyone. You can even go so far as to say that you think open technologies like FSR, with wider reach, are better than closed ones with limited compatibility like DLSS. Because in that statement, you're qualifying the where and why of the claim of one being better than the other. But based on the English language, and the definition of the word "better," it is something that can't be attributed to FSR based on the reasons you've stated without additional attributions.

It's simply a matter of language, and not a matter of subjective opinion.

 

[UltraMega]

Have you seen any of the leaks from verified sources? They're going from Samsung 8nm which was just a modified 10nm, to TSMC 5nm. Core count is up, as are clocks significantly. On top of that they've massively increased the cache. Not as much as AMD, but still many many times larger than current gen.


That doesn't make FSR better. It makes it available. As the only option. The only way one can be better than the other, is if you have a choice of which to use, and pick one over the other. "I'm happy it's available for low end GPUs" is not the same as "FSR is better than DLSS," which is what you claimed. And I quote:

"I mean... the fact that FSR 2.0 works with either team's GPUs is why it's better imo."

No. Not better. Accessible to more people. Not better. It is in fact worse. It has worse image quality, and worse performance on average. If you got to pick which to use, you'd pick DLSS. Because DLSS is "better."



No. The definition of better is very clear. FSR isn't "better" because it works on a larger percentage of GPUs. You can say FSR is more readily available. You can say it has better hardware support. Or that it provides access to better upscaling methods than is available to non-RTX cards. But you can't say FSR is better without changing the implication of the word. And to do so, you'd need to state it as such. For example:

• A Toyota Camry has better fuel mileage than a Ferrari
• A Toyota Camry has a better price than a Ferrari
• A Toyota Camry does a better job of transporting a family of 5
• A Toyota Camry is better when it comes to maintenance costs

But none of those things allow you to say a Toyota Camry is better than a Ferrari. When you compare similar things and use the general term "better," it means you're comparing them based on how well they do what they're supposed to do. The Ferrari does almost everything a car is supposed to do a lot better. The fact that the price puts it out of the reach of most people doesn't make the Toyota a better car. It just makes it a more affordable and attainable car. And as a side note...generally more attainable items are considered "less good" based on human psychology...but let's not get into that.

So what are DLSS and FSR? They're upscaling techniques. They're supposed to improve performance, and maintain image quality. That is their only job. Which provides the best image quality? DLSS. Which provides the best performance improvement on average? DLSS. Therefore, you can't say that FSR is better than DLSS unless you're disputing those claims. Do you happen to like FSR more because it's available to more people? That's a fair statement and a fair reason to like and support it. You can say I think AMD is better than Nvidia for releasing FSR and making it available to everyone. You can even go so far as to say that you think open technologies like FSR, with wider reach, are better than closed ones with limited compatibility like DLSS. Because in that statement, you're qualifying the where and why of the claim of one being better than the other. But based on the English language, and the definition of the word "better," it is something that can't be attributed to FSR based on the reasons you've stated without additional attributions.

It's simply a matter of language, and not a matter of subjective opinion.

Better is a subjective term when comparing the overall variables for each. You can write an essay about it if you want, but ultimately it's still a subjective term. I'm not personally arguing that either is better, merely that the term is subjective when comparing the two overall so you cant correctly say it's wrong when it's subjective. What's better to you isn't what's better to everyone.

 

[HyperMatrix]

Better is a subjective term when comparing the overall variables for each. You can write an essay about it if you want, but ultimately it's still a subjective term. I'm not personally arguing that either is better, merely that the term is subjective when comparing the two overall so you cant correctly say it's wrong when it's subjective. What's better to you isn't what's better to everyone.

Your lack of understanding of basic definitions and grammar has no bearing on my acceptance of your statements. What you're typing doesn't convey what you're trying to say. Unless you're claiming that you'd use FSR over DLSS if given a choice, then the answer is quite clear. You're redefining words to back up a certain narrative. I'm free to use either technologies and have no reason not to use the one that is better. You have to stick to FSR, which objectively has inferior image quality and performance gains on average. So it's not all that surprising that you'd be biased and try to defend the only option you have. It's basic human psychology. But fanboyism doesn't help anyone.

 

[dagget3450]

This is why i have been complaining about upscaling since its become the new fad for the gpu makers. The waters will be blurred and muddied and lots more room for errors when testing.(no pun intended) This is the kind of crap i expect to become normal now, at least when benchmarking and comparing.

Any feature that requires game dev/implementation is going to fall flat on its face in the long term. Looking at the history of "features" added that require extra work from a money making production isn't going to happen. UNLESS the gpu brand dumps their own effort and money to make it happen.(temporary at best)

Why in the world is top end gpus being used for upscaling testing anyways. Marketing still trying to convince people its a top end feature? Save the upscaling garbage for the low end hardware where its actually usable and has a meaning. I personally think its all a joke and snake oil outside that use case scenario.

 

[UltraMega]

Your lack of understanding of basic definitions and grammar has no bearing on my acceptance of your statements. What you're typing doesn't convey what you're trying to say. Unless you're claiming that you'd use FSR over DLSS if given a choice, then the answer is quite clear. You're redefining words to back up a certain narrative. I'm free to use either technologies and have no reason not to use the one that is better. You have to stick to FSR, which objectively has inferior image quality and performance gains on average. So it's not all that surprising that you'd be biased and try to defend the only option you have. It's basic human psychology. But fanboyism doesn't help anyone.

I already said DLSS is better when it comes to image quality. FSR is better when it comes to compatibility. Which one of those things matters most to any particular person is subjective. It's just that simple.

 

[HyperMatrix]

I already said DLSS is better when it comes to image quality. FSR is better when it comes to compatibility. Which one of those things matters most to any particular person is subjective. It's just that simple.

The first part of your statement above is grammatically correct. Because you’ve stated which specific area you believe FSR is better in. That being compatibility. The second part isn’t. It’s not up to each person to decide which of those things matters most to them (image quality vs. compatibility). Because you either have access to DLSS or you don’t. If you have access to it, you’ll use it. If you don’t have access to it, then you’re not really choosing FSR over DLSS for compatibility, or otherwise. Because there is no other option available to you.

To reiterate, my issue isn’t with FSR at all and I’m genuinely happy that FSR 2.0 came out. And I’m quite unhappy with many of Nvidia’s practices. I’ve just seen an enormous amount of trolling and attacks and fanboyism online regarding this whole matter where DLSS was garbage blur and then FSR 1.0 came out and it was “better than DLSS” and then FSR 2.0 and the same people who hated on DLSS were now talking about how amazing FSR 2.0 is.

I’m only interested in one thing. Facts. Technical information. And that’s why I clicked the original post. I was curious to see the alternate findings and to see where DF may have made a mistake as I generally do see them as being biased in favor of Nvidia. So I apologize for being very strict with wording. I'd like this community to remain one where we can discuss various technical features openly and with the intention of learning, as opposed to simply defending our positions.

FSR 2.0 is a better upscaler at doing more with less. Nvidia intentionally limited DLSS to RTX cards for marketing purposes. They started with the same thing FSR 2.0 does, then used the tensor cores to improve that. But they could have easily released this lesser version of it for previous gen cards. Obviously poor image quality isn't a big concern for them considering they released the absolutely horrendous DLSS 1.0. So it was purely greed and marketing.

We're not in complete disagreement here. Just a matter of contention over the broad definition of "better" which would seem to imply either superior image quality or performance gains, though I understand that this is not your intent.

 

[UltraMega]

The first part of your statement above is grammatically correct. Because you’ve stated which specific area you believe FSR is better in. That being compatibility. The second part isn’t. It’s not up to each person to decide which of those things matters most to them (image quality vs. compatibility). Because you either have access to DLSS or you don’t. If you have access to it, you’ll use it. If you don’t have access to it, then you’re not really choosing FSR over DLSS for compatibility, or otherwise. Because there is no other option available to you.

To reiterate, my issue isn’t with FSR at all and I’m genuinely happy that FSR 2.0 came out. And I’m quite unhappy with many of Nvidia’s practices. I’ve just seen an enormous amount of trolling and attacks and fanboyism online regarding this whole matter where DLSS was garbage blur and then FSR 1.0 came out and it was “better than DLSS” and then FSR 2.0 and the same people who hated on DLSS were now talking about how amazing FSR 2.0 is.

I’m only interested in one thing. Facts. Technical information. And that’s why I clicked the original post. I was curious to see the alternate findings and to see where DF may have made a mistake as I generally do see them as being biased in favor of Nvidia. So I apologize for being very strict with wording. I'd like this community to remain one where we can discuss various technical features openly and with the intention of learning, as opposed to simply defending our positions.

FSR 2.0 is a better upscaler at doing more with less. Nvidia intentionally limited DLSS to RTX cards for marketing purposes. They started with the same thing FSR 2.0 does, then used the tensor cores to improve that. But they could have easily released this lesser version of it for previous gen cards. Obviously poor image quality isn't a big concern for them considering they released the absolutely horrendous DLSS 1.0. So it was purely greed and marketing.

We're not in complete disagreement here. Just a matter of contention over the broad definition of "better" which would seem to imply either superior image quality or performance gains, though I understand that this is not your intent.

This isn't a grammatical issue. Saying it over and over in longer and longer ways doesn't change anything. Subjective opinions are subjective, period. You don't have to agree with it.

 

[HyperMatrix]

This isn't a grammatical issue. Saying it over and over in longer and longer ways doesn't change anything. Subjective opinions are subjective, period. You don't have to agree with it.

I don't agree that something is better than something else when it is objectively worse at doing what it is supposed to do. You don't have to agree with facts.

 

[jonRock1992]

I HIGHLY doubt that Alex is trying to make FSR 2.0 look bad lol.