Bo_Hazem
Banned
We only saw bad examples so far, that's why. But we'll see how it'll pan out.
Last edited:
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
- Status
We only saw bad examples so far, that's why. But we'll see how it'll pan out.
To me it`s look like a proper PS5 game... can`t get much better, doubt the will be a PS5 version.... maybe for marketing reasons only.
ethomaz
Banned
I can understand because they sell more decals on iPhone than Xperia phones.Sony showing the decals on iPhone instead of Xperia.
Until now they did not have a hit with Xperia.
That is what it is... even so some Xperia phones in the past had amazing builds and features.
Last edited:
onesvenus
Member
No, in fact uses Tier 2 VRS, I was wrong. When looking for the AMD announcement video I watched it again and it says it's using VRS Tier 2.
It's obvious we are missing something because when it's well used (see the gears 5 image) it shows almost no perceivable differences. In fact you know those artifacts could be due to a number of different things. I don't know why we are assuming it's from VRS
SlimySnake
Flashless at the Golden Globes
I played it on pc at max settings.b
Sony Launches “PlayStation 5 Launch Collection Merchandise Bundle;” Includes Water Bottle, Hat & More
PlayStation 5 Launch Collection Merchandise Bundle Contents:
- Merchandise approved by PlayStation for high quality Standards and Child Safety
- Dad hat polyester-wool blend with embossed adjustable clasp
- 17 oz. Stainless steel water bottle double wall insulation, keeps drinks hot/cold for hours, wide-mouth and leak-proof, bpa Free
- Knit socks – high quality polyester-spandex blend, One size fits most
- Tech Decals — Premium 3M Material, Scratch and water Resistant, Leaves No Residue
Sony Launches "PlayStation 5 Launch Collection Merchandise Bundle;" Includes Water Bottle, Hat & More
Sony has launched the "PlayStation 5 Launch Collection Merchandise Bundle,"& it includes a water bottle, phone case and more.mp1st.com
Number one best seller on Amazon under “Video Games”, “New Releases”.
Those color coded buttons...
Tier 2 is Not used on Dirt Series X as far as I know . Maybe Tier 1, Tier 2 is used on PC, as showcased .
onesvenus
Member
Well, if you discard the best implementation we have, in Gears 5, it's obvious we only have bad examples.
Bo_Hazem
Banned
Because even PS4 Pro has better far distance graphics, I believe XOX as well.
So AMD and Dirt 5 devs confirming it as the "holy grail" Tier 2 makes it pretty embarrassing to talk about VRS as a "feature".
Last edited:
ethomaz
Banned
If you read the own MS article about VRS Tier 2 it is a very case by case and the results can variate a lot... so the devs needs to see if it is worth or not.
Moving Gears to Tier 2 Variable Rate Shading - DirectX Developer Blog
The team at The Coalition haven’t stopped innovating since bringing Tier 1 VRS to Gears Tactics, and have brought Tier 2 VRS support to both Gears 5 and Gears Tactics. The team saw similarly large perf gains from VRS Tier 2 –
devblogs.microsoft.com
What the hot take?
- Only see performance improvement in higher resolutions. It is in most cases useless for 1080p or below resolutions.
- It only see improvement in GPU bound games.
- It is very limited where you can use it and it not improve frametime in all places.
About the look... while in some cases it is very close to the original imagine you can still spot reduced resolution where VRS Tier 2 was applied (even in Gears 5 you can see it) but if you get the performance boost in frametime it is worth of use.
Last edited:
Hendrick's
If only my penis was as big as my GamerScore!
You mean in static screen captures? VRS is not for 600X magnification scrutiny. Luckily, no one sees that while actually playing.
Bo_Hazem
Banned
If you have a bad vision, yes. If you are a normal person with healthy vision it's pretty obvious. Look at my post above yours.
Hendrick's
If only my penis was as big as my GamerScore!
Your still image of a single frame? Yeah, my point exactly. No way anyone notices that while playing. That also may have nothing to do with VRS.
Last edited:
Bo_Hazem
Banned
sircaw
Banned
Just look at the eyes of your Avatar, Goddammit.
ethomaz
Banned
Guys the point is not about the quality of the VRS, it indeed degrade the overall image.
VRS was made to gain frametime and that it really do a great job in some cases and it is worth the degradation of the image.
It is a trade off option for devs.
I can see a 4k 30fps Quality Mode with VRS turned off and a 4k 60fps Performance Mode with VRS turned on on console hardware.
Now 1080p Performance modes won't have gains with VRS.
VRS was made to gain frametime and that it really do a great job in some cases and it is worth the degradation of the image.
It is a trade off option for devs.
I can see a 4k 30fps Quality Mode with VRS turned off and a 4k 60fps Performance Mode with VRS turned on on console hardware.
Now 1080p Performance modes won't have gains with VRS.
Last edited:
LiquidRex
Member
Tomb Raider Chronicles - SQUARE ENIX TO MARK 25 ANNIVERSARY OF TOMB RAIDER
www.tombraiderchronicles.com
SSfox
Member
The best thing would be a remake ala Crash or demon Soul, at least for the first 5 games.Tomb Raider Chronicles - SQUARE ENIX TO MARK 25 ANNIVERSARY OF TOMB RAIDER
www.tombraiderchronicles.com
In the topic of remake i'm suspecting RDR1 remake can happen (and also it's something i want personally)
Last edited:
almost exactly, but I must correct the last sentence and the beginning (I just have to
... don't ask me why):- it will not degrade the overall image quality, it should only degrade parts which are not "visible" (due to darkness or general blurred out details), if implemented correctly
- 1080p won't have a huge gain on a big GPU, yes. The big GPU will most of the time not be limited by the resolution of shading. But the smaller the GPU is, the higher is the gain from that
It is just a feature to reduce the work to be done, so the saved cycles can be used elsewhere (e.g. more frames).
Dabaus
Banned
Went through this twitter. Nothing but vague teasing like "big things happening at X.' Im over these so called insiders and their empty teases that dont end up happening anyways.
DeepEnigma
Gold Member
That's a nifty little site, thanks for sharing.
It would be nice if they would patch The Last Guardian to unlock the framerate so you can keep HDR or use the digital version.
Panajev2001a
GAF's Pleasant Genius
Because it seems to indicate the artefacts you would expect from essentially "localised lower resolution rendering": I am not saying it cannot be well used, just that the dev in this case may have been a bit aggressive and "oh who is going to zoom in and worry about that when the game is 120 Hz, SHIP SHIP SHIP!".
To get that effect it would have meant more draw calls as Tier 1 sets rasterisation rate per draw call only ( https://devblogs.microsoft.com/directx/gears-vrs-tier2/ ) which would seem a bit difficult in a 120 Hz framerate target.
Thanks for the correction btw, I appreciate the link and updated notes.
Last edited:
Flippo
Neo Member
Pay in the sense of money..
Because Bo needs the new Sony Alpha 1 camera
7300 euro, dang... 4K/120 8K/30
Last edited:
Neo Blaster
Member
Going from a free 60fps upgrade to another full buy would definitely suck, we've already got scammed by Capcom (DMC5) and 505(Control), not to mention Sony itself(Spiderman).
Shmunter
Member
I’d be #1 whiny bitch poster
Dabaus
Banned
Not that id use it but im surprised Sony didnt try to do a new rebooted PS Home or whatever it was called. Not that i want them to nick and dime people but the potential to sell Microtransactions for your own personal hub seems like a no brainer when your brand awareness is is at an all time high and have like a 120 million installed base.
Shmunter
Member
Would be nice. Unfortunately they must’ve run it through the computer, and concluded taking time away from games, their dlc and micro transactions is less profitable.
Hope I’m wrong
LucidFlux
Member
I get what you're saying but the asset(s) in question aren't zipping by the player's field of view where the loss of image quality wouldn't be easily noticed. These are prominent areas of focus that hang right in the middle of the screen for long stretches of time.
Now, I agree whatever is causing this may in fact have absolutely nothing to do with VRS and if it does it's just a horrible implementation.
ethomaz
Banned
A feature that reduce the work to be done at the cost of degraded image.almost exactly, but I must correct the last sentence and the beginning (I just have to
- it will not degrade the overall image quality, it should only degrade parts which are not "visible" (due to darkness or general blurred out details), if implemented correctly
- 1080p won't have a huge gain on a big GPU, yes. The big GPU will most of the time not be limited by the resolution of shading. But the smaller the GPU is, the higher is the gain from that
It is just a feature to reduce the work to be done, so the saved cycles can be used elsewhere (e.g. more frames).
It is a trade off feature... or lossy if you want to give a name.
But it is what the industry is moving for... same happened with AA techs... Dynamic and/or temporal retrotemporal resolutions, etc.
Last edited:
TheThreadsThatBindUs
Member
I don't know if Dirt 5 is using VRS, but the type of artefacting exhibited on the screens taken from the 120Hz mode are certainly indicative of such. Usually if it quacks like a duck, it's probably a duck.
That said, it would seem like a poor implementation. Everyone needs to remember that VRS usage among developers is still quite a fresh new thing. And given there's no one, all encompassing implementation (devs have to roll their own unique implementation), there's very likely to be some iteration and exploration devs will do until they get it right.
So a few poor examples of VRS Tier 2 implementations doesn't mean VRS Tier 2 is shit. It just means it'll take some time for devs to figure it out. As has been said many times from the beginning, VRS Tier 2 conceptually SHOULDN'T perceptively degrade IQ. If is it, the dev is doing something wrong. And at this stage, that's ok and we should expect that.
If we're still getting weird macro-blocking artefacts on texture detail for large polys 4yrs into the gen, then we need to start asking some deeper questions about the API or hardware setup for MS's VRS implementation. At this stage, however, teething issues are and should be expected.
ethomaz
.
Yes the cost of the performance optimization is degradation of IQ, but the question is and should be, "by how much". If it's perceptible, then throw the optimization in the bin and look for something else. If it's not, let's go bitches!
VRS Tier 2 SHOULD NOT produce a perceptible degradation in IQ, if implemented as intended. At least for now, Gears 5 provides a clear positive data point for an implementation that achieves the stated goal of the VRS technique. So nobody should be making sweeping statements about the viability of the VRS technique as a whole based on examples of bad implementation.
VRS isn't like some quick and dirty AA method where the dev just flips a switch in the renderer to enable it. The Dev needs to write their own unique algorithmic approach to it's utilization. Therefore, some implementations by some devs will be worse or better than others.
That said, it would seem like a poor implementation. Everyone needs to remember that VRS usage among developers is still quite a fresh new thing. And given there's no one, all encompassing implementation (devs have to roll their own unique implementation), there's very likely to be some iteration and exploration devs will do until they get it right.
So a few poor examples of VRS Tier 2 implementations doesn't mean VRS Tier 2 is shit. It just means it'll take some time for devs to figure it out. As has been said many times from the beginning, VRS Tier 2 conceptually SHOULDN'T perceptively degrade IQ. If is it, the dev is doing something wrong. And at this stage, that's ok and we should expect that.
If we're still getting weird macro-blocking artefacts on texture detail for large polys 4yrs into the gen, then we need to start asking some deeper questions about the API or hardware setup for MS's VRS implementation. At this stage, however, teething issues are and should be expected.
This is a bit of a gross over-simplification,
ethomaz
.Yes the cost of the performance optimization is degradation of IQ, but the question is and should be, "by how much". If it's perceptible, then throw the optimization in the bin and look for something else. If it's not, let's go bitches!
VRS Tier 2 SHOULD NOT produce a perceptible degradation in IQ, if implemented as intended. At least for now, Gears 5 provides a clear positive data point for an implementation that achieves the stated goal of the VRS technique. So nobody should be making sweeping statements about the viability of the VRS technique as a whole based on examples of bad implementation.
VRS isn't like some quick and dirty AA method where the dev just flips a switch in the renderer to enable it. The Dev needs to write their own unique algorithmic approach to it's utilization. Therefore, some implementations by some devs will be worse or better than others.
Last edited:
onesvenus
Member
Would you say that checkerboarding or temporal reconstruction is also lossy 4K?
Just checking if that goes both ways
Shmunter
Member
Temporal reconstruction generates more detail. It’s opposite in function.
ethomaz
Banned
Yes... I always stated that.
The fact I don't like CBR of the PS4 Pro already tells you a lot.
Actually it can't fill all parts and the reconstructions fell off due artifacts and blurry that makes the image less detailed.
Last edited:
TheThreadsThatBindUs
Member
Lossy isn't inherently bad.... I think is the point,
onesvenus
is trying to make.
GPUs internally have been using lossy texture compression formats for decades. Should they throw them in the bin because they're lossy? No... they're valuable because their freaking fast to compress/decompress.
There's always a trade-off between performance vs quality with any lossy technique, and lossless techniques are often an order of magnitude more costly in terms of performance than lossy methods.
The important question is never, "is it lossy". It's "by how much".
onesvenus
is trying to make.GPUs internally have been using lossy texture compression formats for decades. Should they throw them in the bin because they're lossy? No... they're valuable because their freaking fast to compress/decompress.
There's always a trade-off between performance vs quality with any lossy technique, and lossless techniques are often an order of magnitude more costly in terms of performance than lossy methods.
The important question is never, "is it lossy". It's "by how much".
Last edited:
ethomaz
Banned
Yeap I agree... lossy gives beneficies in rendering time.Lossy isn't inherently bad.... I think is the point,
GPUs internally have been using lossy texture compression formats for decades. Should they throw them in the bin because they're lossy? No... they're valuable because their freaking fast to encode/decode.
There's always a trade-off between performance vs quality with any lossy technique, and lossless techniques are often an order of magnitude more costly in terms of performance than lossy methods.
The important question is never, "is it lossy". It's "by how much".
It is a trade-off like you said.
It is a dev by dev choice of what is worth for them.
The point is a lot of people try to pass CBR, DLSS, VRS, etc as something that gives better IQ and that is false... it gives a close enough IQ with better performance.... devs choose it in most cases that the performance results are bigger than the lose in IQ.
Last edited:
Shmunter
Member
I don’t know, there are some stellar results out there. If reconstruction served no purpose, it wouldn’t exist.
In all honesty any dev not doing reconstruction and wasting resources on native resolutions is doing it wrong.
onesvenus
Member
It doesn't generate more detail, it uses already existing detail to fill in the missing detail. Can you point me to a temporal reconstruction that uses full 4k images? You won't find it because temporal reconstruction is done using partial 4k images, and using the previous ones to create a complete 4k image.
It falls completely in
ethomaz
definition of lossy: "A feature that reduce the work to be done at the cost of degraded image."
saintjules
Member
Shmunter
Member
re-construction constructs a higher resolution image from a lower resolution one. Whatever the underlying technique.It doesn't generate more detail, it uses already existing detail to fill in the missing detail. Can you point me to a temporal reconstruction that uses full 4k images? You won't find it because temporal reconstruction is done using partial 4k images, and using the previous ones to create a complete 4k image.
It falls completely in
Lossy compression takes a top source image and removes things to compress.
It’s direct opposite.
Last edited:
TheThreadsThatBindUs
Member
CBR and DLSS are fundamentally different quantities than VRS. CBR/Temporal Reconstruction/DLSS improve IQ by increasing resolution. By definition you're generation new information (new pixels). Yes the IQ is degraded compared to a native 4K but the techniques aren't using a native 4K framebuffer as an input in the first place, they're using one or more lower resolution buffers. So they do indeed increase IQ by increasing pixel density.
VRS is a pure performance optimization that reduces shading work in areas where you can get away with a lower resolution shading rate (thus degraded IQ). You're not creating new information and so it's by definition lossy and thus very different from the former.
It doesn't generate more detail, it uses already existing detail to fill in the missing detail. Can you point me to a temporal reconstruction that uses full 4k images? You won't find it because temporal reconstruction is done using partial 4k images, and using the previous ones to create a complete 4k image.
It falls completely in
Of course they generate new detail. You're creating whole new pixels, increasing the resolution. Neither CBR/Temporal Reconstruction/DLSS fall into your definition, as none of them reduce work (you are doing more work than simply rendering the base resolution image), and all of them increase IQ by increasing pixel density (resolution) from the base image.
The quality of the new pixels generated is immaterial. The IQ isn't degraded from the base image, as the original base image pixels are unchanged, and added to a collection of newly generated pixels to increase resolution. Instead of comparing to a native 4K image, if you compare with a basic 4k upscale using some shitty bilinear or bicubic function, it's more evident that these methods increase IQ over the base image.
It's the opposite of lossy, as has been said.
I don't know what the base resolution is for each game, neither have I looked all that closely at the Gears 5 implementation of VRS to see if or where any artifacts may appear. However, from the Dirt 5 screens it's pretty clear the issue isn't base rendering resolution. It's that VRS is applied to reduce shading rate on brightly lit, moderate contrast areas of the screen. This is the polar opposite of how and where it should be applied.
VRS should ONLY be applied to low contrast, high color uniformity, areas of the image; e.g. areas in shadow, and area with high color uniformity for example a clear blue sky.
The base rendering resolution will impact the performance improvement gained by using VRS more than the quality of the result, provided VRS is implemented properly. Again I must stress, provided VRS is implemented properly.
Last edited:
ethomaz
Banned
You are right.
My mistake was to take in mind that most people compares the resolution techs with the native resolution it is supposed to replace (the target resolution)... so the native option has better IQ and more details.
But what you said is right... 1800p base resolution with Resolution Tech to reach 4k will indeed looks better than 1800p native... not just look better it will be more expensive to render for the hardware... of couse it doesn't change the fact that target resolution (4k) will have better IQ and be more hungry in processing power than that 1800p + ResTech.
It is a trade off when compared target resolution.... but it is a improve if you compare the base resolution.
Lossy is really not the best word here.
Last edited:
ethomaz
Banned
I believe the focus is to release R&C in March/April.
After hey will probably have two more first party for this year... GT7 in October? And a big one on holidays.
But that is all guesses.
I’m convinced next time they talk about GT7 will be with the release date.
Last edited:
Three Jackdaws
Banned
Sony were drip feeding PS5 info throughout last year, and the fans just about managed to put up with it, as
HeisenbergFX4
pointed out, this was part of their strategy to “bleed” Microsoft out.I think they’ll be doing something similar this year with the big game announcements and trailers/gameplay because likewise Microsoft should have some ammunition this year which they’ll be ready to fire, I don’t think that’ll be in the form of triple A releases but likely triple A exclusive announcements, gameplay and trailers. I cannot wait for Hellblade 2!
Dabaus
Banned
Id say with the besthesda acquisition Microsoft bled sony out, and im not even a microsoft fan.Sony were drip feeding PS5 info throughout last year, and the fans just about managed to put up with it, as
I think they’ll be doing something similar this year with the big game announcements and trailers/gameplay because likewise Microsoft should have some ammunition this year which they’ll be ready to fire, I don’t think that’ll be in the form of triple A releases but likely triple A exclusive announcements, gameplay and trailers. I cannot wait for Hellblade 2!
Three Jackdaws
Banned
Sales of the consoles would strongly suggest otherwise.
DarkMage619
$MSFT
Are console sales the sole metric MS uses to gage success?
- Status