Rootax
Veteran
I can't seem how it's diminishing anything when it's inferior to TAA...
Well at least they have a solution to boost framerate now. Probably not as good, but it's out, and it's a big step for AMD I believe.
I can't seem how it's diminishing anything when it's inferior to TAA...
Well at least they have a solution to boost framerate now. Probably not as good, but it's out, and it's a big step for AMD I believe.
https://www.techpowerup.com/review/...olution-quality-performance-benchmark/10.htmlAccording to AMD, the FSR code will be open-sourced on their GPUOpen site "in mid-July." The shader code isn't publicly available at this time, and AMD wasn't willing to share it with me—I asked.
That video Alex, cant say I wasn't a bit disappointed by it. It felt like you went in with the intent to destroy it as much as possible. I understand nitpicking is what DF does and I dont have any problem with that, but little to no effort was made to say anything good about it on the other hand, which I think there is plenty to say on that front for a genuinely platform and engine agnostic technique like this.
It should be absolutely no surprise that FSR doesn't fare well with lower base resolutions so focusing on FSR Performance mode seemed a little....harsh. I also think you were too quickly dismissive of how 'bad' all modes below Ultra Quality looked, seemingly running with the idea that reconstruction techniques aren't worthwhile unless they essentially match native rendering, ignoring the huge performance gains you get in return(is performance king or not?). Quality definitely looks highly usable to me in all the examples I've seen.
The length of the video also felt a bit insulting. 14 minutes isn't super short, but y'all have spent 20 minutes talking just about the improvement of DLSS from 1.9 to 2.0 in one game before. FSR seems a bit too significant to just brush off so easily and I was definitely expecting something more substantial here.
I'd say overall, FSR is probably more impressive than most thought it was going to be. And could easily be a common inclusion in titles going forward if it's as easy to implement as claimed.
From GamersNexus video:Haven't had time to read articles yet.
Has anyone tested the actual cost of FSR?
Something like 1080p vs 4k ultra performance FSR?
It wasn't focused on performance mode. The majority of the Godfall comparisons were Ultra Quality, the rest of the video was showing all modes across a variety of resolutions, as you should do when evaluating any reconstruction tech.That video Alex, cant say I wasn't a bit disappointed by it. It felt like you went in with the intent to destroy it as much as possible. I understand nitpicking is what DF does and I dont have any problem with that, but little to no effort was made to say anything good about it on the other hand, which I think there is plenty to say on that front for a genuinely platform and engine agnostic technique like this.
It should be absolutely no surprise that FSR doesn't fare well with lower base resolutions so focusing on FSR Performance mode seemed a little....harsh. I also think you were too quickly dismissive of how 'bad' all modes below Ultra Quality looked, seemingly running with the idea that reconstruction techniques aren't worthwhile unless they essentially match native rendering, ignoring the huge performance gains you get in return(is performance king or not?). Quality definitely looks highly usable to me in all the examples I've seen.
14 minutes isn't 'brushing off', if anything I think DF's videos could all stand to be a bit more compact. DLSS 1.9 to 2.0 was a massive upgrade regardless, despite Nvidia just increasing the version number by .1 It basically made DLSS actually worthwhile.The length of the video also felt a bit insulting. 14 minutes isn't super short, but y'all have spent 20 minutes talking just about the improvement of DLSS from 1.9 to 2.0 in one game before. FSR seems a bit too significant to just brush off so easily and I was definitely expecting something more substantial here.
That video Alex, cant say I wasn't a bit disappointed by it. It felt like you went in with the intent to destroy it as much as possible. I understand nitpicking is what DF does and I dont have any problem with that, but little to no effort was made to say anything good about it on the other hand, which I think there is plenty to say on that front for a genuinely platform and engine agnostic technique like this.
It should be absolutely no surprise that FSR doesn't fare well with lower base resolutions so focusing on FSR Performance mode seemed a little....harsh. I also think you were too quickly dismissive of how 'bad' all modes below Ultra Quality looked, seemingly running with the idea that reconstruction techniques aren't worthwhile unless they essentially match native rendering, ignoring the huge performance gains you get in return(is performance king or not?). Quality definitely looks highly usable to me in all the examples I've seen.
The length of the video also felt a bit insulting. 14 minutes isn't super short, but y'all have spent 20 minutes talking just about the improvement of DLSS from 1.9 to 2.0 in one game before. FSR seems a bit too significant to just brush off so easily and I was definitely expecting something more substantial here.
I'd say overall, FSR is probably more impressive than most thought it was going to be. And could easily be a common inclusion in titles going forward if it's as easy to implement as claimed.
I think you're mixing up TAA with TAA U (U for upscaling). FSR works fine with TAA, and not all engines support TAA U.I think you are missing the main point, which explains why they were so dismissive: it gives worse results than TAA, which is also platform agnostic. If FSR would not require engine integration and worked out of the box with almost any game, then yeah it would be impressive for the results it gives. But if it requires integration why would a game or game engine implement it instead of TAA?
But it won't have temporal artefacts in the games that don't use/enable TAA.FSR will have the same temporal artefacts - FSR will use TAA in a game that uses TAA.
I think you're missing the point of FSR? You wouldn't use it instead of TAA only. You'd use it to get close to native and increase performance by a lot. If you don't need the extra performance then there's no need to turn it on.But if it requires integration why would a game or game engine implement it instead of TAA?
I think you're missing the point of FSR? You wouldn't use it instead of TAA only. You'd use it to get close to native and increase performance by a lot. If you don't need the extra performance then there's no need to turn it on.
Lastly there are people who really don't like TAA. There's a subreddit dedicated to hating TAA lol.
But it won't have temporal artefacts in the games that don't use/enable TAA.
I was referring to the question you pose at ~12m50s: "Why must a game only have basic upscaling?".
I fully agree that every game (at least every game using a deferred renderer?) should have a TAA U toggle, but the option to disable it should be there for people who notice and are bothered by temporal artifacts too much.
FSR likely needs the original frame to be AAed as good as possible.Lucky for them that FSR needs TAA to be even functionable
Thanks.From GamersNexus video:
Granted that's on a 5700G so it would be interesting to see the cost on proper desktop GPUs.
Watched HUB's video and yeah it's bad again. All they are saying there goes against every other review and image I've seen so far.
I don't know why you put a timestamp on terminator resistance but FSR doesn't need TAA to be even functionable. It works with any kind of AA.