Next gen PS5 and next Xbox launch speculation - Secret sauces spicing 2019

[Albert Penello]

Thanks! My thought was that if conceptualized from the beginning, you could downgrade higher SKU CPUs to lower tier SKU CPUs. The cost equation would be whether it was worth it to make the lower tier SKU package accept a higher tier "reject" die and its own customized die, but it doesn't surprise me the economics don't work on that.

Click to shrink...

There are a lot of reasons why this doesn't work but I'll give a simple example.

For binning to make sense, first the performance increase has to be enough more for people to care. So debate #1 (using Scorpio) is 6.6tflops enough more power to charge more money than 6? Your mileage may vary on this answer.

Second, you now have to build a case that can reliably handle 6.6tflops. This will not be the same case that can handle 6. So question #2 is - do you build a different case, power supply, etc. for this version, or do you make every version of the console capable of handling the higher-performing chips?

Both solutions are expensive - you either pay for an entirely new design, or you over-pay on every single one.

So right there the concept sort of fails. In the case of the Scorpio dev-kit, that was already going to be a much different design, it's a small-run product with a lot of changes and is very, very expensive to build which is typical for bespoke development kits. So it was a no-brainer for those consoles but would make for bad economics at mass scale.

 

[VX1]

In the case of the Scorpio dev-kit, that was already going to be a much different design, it's a small-run product with a lot of changes and is very, very expensive to build which is typical for bespoke development kits

Click to shrink...

Oh yes,probably the best looking dev kit ever! Plus that oled display...
I bet that whole thing was damn expensive!

 

[Pheonix]

No. But that exact strategy that was used for the Project Scorpio Dev Kits which is why they have 6.6tflops. They were binned with the better parts. For a large-scale console release, no way that would make sense as it's far too unreliable at scale.



Hm. There isn't a one-size-fits-all answer to this because there haven't been that many E3's where next-gen consoles have been presented, and the debate on what was shown, and when, differed greatly each time.

I guess I can say this. First, there are probably less than 25 people within Xbox that know the exact plan. Phil and his directs as well as a small handful of others. People are brought in for their individual sections, but exactly what is being shown in what order is very, very tightly controlled and many things are code-named so if someone sees a list they don't know what the titles are. I always question when anyone says they "know" or "have heard" what someone is doing at E3 because the people that know aren't talking.

But it's not super hard to reverse-engineer what might be shown if you're following the industry, looking at all the different developers and triangulating what they are working on. It's why E3 bingo is so much fun.

As for consoles and what might be shown? It has a lot to do with balancing what action you want people to take, what you think you get strategically vs. waiting, and what you think the competition is doing. Since I'm not there, I don't know what the current team is thinking in this regards.

So your guess is as good as mine on this one. Scorpio reveal was pretty far out so they could do that level of unveil easily.

I'm personally not holding out for much in the way of a console news. There simply isn't enough reason to do it, and there are only downsides. If you're Sony, numbers are great, you're not at E3 anyway, and your next event is later in the year so why risk stalling any holiday sales.

For Xbox, because Sony isn't at E3 there is literally 0 risk of them taking any news cycles away, and like Sony, Xbox sales are good, people already know they are doing something since Phil said it last year, so I think they focus on games and xCloud.

If there is anything I'm guessing it will be something teaser-y like Scorpio if at all. And they can make the call to do something like that pretty late in the game.

Like I said - I have ZERO knowledge of the plan. They may have other reasons to do more so I'm not taking a hard stance they won't do more, but it would surprise me.

Click to shrink...

Thanks.... need you to help clear up something.

Being that both companies are probably planning to release within 6 months of each other, and likely going to have even more similar or closely matched APUs this time around, what is the likelihood of one coming in at $100 more than the other and what real difference will that $100 make?

Lastly, do you believe his time around MS will be more inclined to match whatever price sony is targeting rather than cost more?

 

[VFX_Veteran]

I also think it's worth noting that we can do limited raytracing to get the visual impact we want at minimal cost.

Click to shrink...

This was somewhat done with SSR and it broke the visual continuity as reflections would disappear or would fade to quickly. I wouldn't count on any RT at this point as it would be extremely difficult to justify the ms cost.

Getting better indirect lighting on nearby objects and characters while introducing fallbacks at certain distances could allow us to use full raytracing more selectively.

Click to shrink...

See above for my concern.

As as example, we could focus on getting characters' clothing reflecting correctly off of their skin and vice versa, giving reflective objects like eyes, swords, mirrors, guns, buttons, buckles, etc. a more realistic representation of 10 meters around them with 4x the voxels we've used on PS4/XB1. If done carefully and designed around, you can do a lot to impress and, well, trick gamers without going all the way. We'll also have better implementations of contact shadows, better skin simulation, and lots of baked, interpolated cloth and hair simulation to make those close-ups even more believable in other ways.

Click to shrink...

We still don't have proper solutions for materials now for in-game footage. I can count on one hand how many in-game characters use actual SSS (the proper implementation - not the hacky masking texture trick).

It's gonna' be dope as fuck even if we're talking the lower end of the power spectrum, guys. The worst-case scenario for next-gen is still pretty damn good when you give the reigns to competent artists and engineers.

Click to shrink...

TBH, I'd really just like to see the next gen consoles implement some of the higher-end PC options (i.e. POM being standard, or more displacement mapping) in multiplat games for their standard. We still have a long ways to go with just the basic stuff implemented on consoles let alone deal with the RT "dream".

 

[Kyoufu]

do you think we'll get Bully 2 before that?

Click to shrink...

No chance imo. They'll most likely be fully focused on GTA6. Besides, I feel like if Bully 2 was going to happen it would have been a thing by now. It's a very old IP.

 

[TheThreadsThatBindUs]

CPUs have been used for RT for years, at least in the film industry.

Click to shrink...

The film industry =/= realtime rendering.

CGI is pure software rendering. They use no fixed function GPU hardware at all, just CPUs and LOTS and LOTS of RAM. RAM is king for offline rendering.

Games and real-time rendering is completely different. The two have entirely different (almost diametrically opposed) priorities.

There are a lot of reasons why this doesn't work but I'll give a simple example.

For binning to make sense, first the performance increase has to be enough more for people to care. So debate #1 (using Scorpio) is 6.6tflops enough more power to charge more money than 6? Your mileage may vary on this answer.

Second, you now have to build a case that can reliably handle 6.6tflops. This will not be the same case that can handle 6. So question #2 is - do you build a different case, power supply, etc. for this version, or do you make every version of the console capable of handling the higher-performing chips?

Both solutions are expensive - you either pay for an entirely new design, or you over-pay on every single one.

So right there the concept sort of fails. In the case of the Scorpio dev-kit, that was already going to be a much different design, it's a small-run product with a lot of changes and is very, very expensive to build which is typical for bespoke development kits. So it was a no-brainer for those consoles but would make for bad economics at mass scale.

Click to shrink...

Thanks Albert. You just confirmed a lot of my thoughts regarding next-gen console development.

 

[Dragon's Game]

hopefully we get a video of Sam Houser talking about the PS5 like he did with the PS3

https://www.youtube.com/watch?v=SO2ZyHmDpCE

 

[anexanhume]

There are a lot of reasons why this doesn't work but I'll give a simple example.

For binning to make sense, first the performance increase has to be enough more for people to care. So debate #1 (using Scorpio) is 6.6tflops enough more power to charge more money than 6? Your mileage may vary on this answer.

Second, you now have to build a case that can reliably handle 6.6tflops. This will not be the same case that can handle 6. So question #2 is - do you build a different case, power supply, etc. for this version, or do you make every version of the console capable of handling the higher-performing chips?

Both solutions are expensive - you either pay for an entirely new design, or you over-pay on every single one.

So right there the concept sort of fails. In the case of the Scorpio dev-kit, that was already going to be a much different design, it's a small-run product with a lot of changes and is very, very expensive to build which is typical for bespoke development kits. So it was a no-brainer for those consoles but would make for bad economics at mass scale.

Click to shrink...

Ah, I wasn't meaning closely spec'ed systems. I was talking about a system where you're disabling half the GPU cores, as an example.

 

[Albert Penello]

Thanks.... need you to help clear up something.

Being that both companies are probably planning to release within 6 months of each other, and likely going to have even more similar or closely matched APUs this time around, what is the likelihood of one coming in at $100 more than the other and what real difference will that $100 make?

Lastly, do you believe his time around MS will be more inclined to match whatever price sony is targeting rather than cost more?

Click to shrink...

Not really going to comment lest someone pick it up as "news" in terms of speculation on price.

I will say that I'm guessing the next-gen consoles are more divergent from each other than convergent, on the whole. Meaning the choice (and tradeoffs) between the two brands will be more complicated than the simple power controversy between XBO and PS4. I think we're gearing up for different machines in more interesting ways.

Of course, last time I predicted something it didn't go so well for me so hopefully this post ages better than that one.

 

[Pheonix]

Not really going to comment lest someone pick it up as "news" in terms of speculation on price.

I will say that I'm guessing the next-gen consoles are more divergent from each other than convergent, on the whole. Meaning the choice (and tradeoffs) between the two brands will be more complicated than the simple power controversy between XBO and PS4. I think we're gearing up for different machines in more interesting ways.

Of course, last time I predicted something it didn't go so well for me so hopefully this post ages better than that one.

Click to shrink...

Haha.... I can understand your caution.... thanks for that little snippet. Thats very interesting.

 

[Barnabas Reynolds]

RT isn't Nvidia exclusive tho


Ray tracing is gonna be the PBR of next gen.

If this is achieved on Vega 56

Click to shrink...

That's a carefully planned tech demo, made to produce the least expensive reflections possible and without any of the overhead of a real game. I hope the consoles push RT in some form as it will help along the technology, but my expectations are very low. Hearing rumors of AMD merely equaling a 2080 isn't helping either, as I assume said rumors are talking about dGPUs.

Not really going to comment lest someone pick it up as "news" in terms of speculation on price.

I will say that I'm guessing the next-gen consoles are more divergent from each other than convergent, on the whole. Meaning the choice (and tradeoffs) between the two brands will be more complicated than the simple power controversy between XBO and PS4. I think we're gearing up for different machines in more interesting ways.

Of course, last time I predicted something it didn't go so well for me so hopefully this post ages better than that one.

Click to shrink...

That does sound interesting. Are you thinking they'll diverge on services, input devices, hardware tiers, or something else entirely?

 

[PetyrBaeless]

The one that comicbooks.com is reporting on?

Click to shrink...

Yeah, this is the one apparently: https://pastebin.com/PY9vaTsR/

 

[Kyoufu]

Yeah, this is the one apparently: https://pastebin.com/PY9vaTsR/

Click to shrink...

This is old but I still laughed. GTA6 in 2020 hahahaha. The whole thing reads like a badly written fan-fiction tbh.

 

[Kuro]

GTA6 one month exclusive and $250 wireless PSVR2 had me rolling

 

[Lashley]

RT isn't Nvidia exclusive tho


Ray tracing is gonna be the PBR of next gen.

If this is achieved on Vega 56

So RTX2080 level performance on PS5/Xbox 4, along a Zen 2 cpu, 24-32GB memory, daaymn next gen gonna be lit.



I agree that we'll see more 60fps games next gen, but it will be as part of a Performance Mode, which I expect to be mandated.

Click to shrink...

lmao no, you won't be getting 2080 level performance

 

[Anthony Hopkins]

RT isn't Nvidia exclusive tho


Ray tracing is gonna be the PBR of next gen.

If this is achieved on Vega 56

So RTX2080 level performance on PS5/Xbox 4, along a Zen 2 cpu, 24-32GB memory, daaymn next gen gonna be lit.



I agree that we'll see more 60fps games next gen, but it will be as part of a Performance Mode, which I expect to be mandated.

Click to shrink...

Ray tracing is not as big a leap as PBR.

 

[Barnabas Reynolds]

Ray tracing is not as big a leap as PBR.

Click to shrink...

Yeah, I mean, pffft, it's just the Holy-fucking-Grail of real-time rendering. No bigs

 

[Lashley]

Ray tracing is not as big a leap as PBR.

Click to shrink...

lmao

oh shit, you're serious?!

 

[7thFloor]

Ray tracing is not as big a leap as PBR.

Click to shrink...

You are wrong.

 

[Kyoufu]

Ray tracing is not as big a leap as PBR.

Click to shrink...

wat

 

[Anthony Hopkins]

Yeah, I mean, pffft, it's just the Holy-fucking-Grail of real-time rendering. No bigs

Click to shrink...

lmao

oh shit, you're serious?!

Click to shrink...

Yeah I've also seen the nvidia conferences talking about how it's the holy grail of real time computer graphics.

And I'm not denying it's not a needed and impressive tech, but the visual leap of no RT vs RT on is not as great as when games did not use material based shaders vs when they did.

For some reason Ray tracing has really screwed up people's perception of graphical improvement.
People act like Ray tracing is like ps2 to 360 level of difference, when in reality its just a more realistic way of doing lighting + shadows, which has its greatest benefit with in direct lighting.

quadrupling the polycount will be a bigger "leap" then ray tracing.

 

[Barnabas Reynolds]

Today's mood is, bewilderment.

 

[Lashley]

Stop the car, I wanna get out

 

[SharpX68K]

What did I miss today?

 

[Kyoufu]

What did I miss today?

Click to shrink...

Absolutely nothing.

 

[Barnabas Reynolds]

The lesson learned is this: If a nascent technology doesn't immediately bring muh graphics up to OVER 9000, it's not that big of a leap. I mean, who cares if we're only seeing a bit of what's going to be possible with improvements to scope and complexity?

 

[Anthony Hopkins]

The lesson learned is this: If a nascent technology doesn't immediately bring muh graphics up to OVER 9000, it's not that big of a leap. I mean, who cares if we're only seeing a bit of what's going to be possible with improvements to scope and complexity?

Click to shrink...

Oh, so we are talking about speculative visuals now?

Why don't we just stick to the games and tech demos which have very good implementations of the technology already.

 

[PetyrBaeless]

This is old but I still laughed. GTA6 in 2020 hahahaha. The whole thing reads like a badly written fan-fiction tbh.

Click to shrink...

"Ghost of Tsushima remastered" is the icing on the cake.

 

[Anthony Hopkins]

Some next gen AAA games with no ray tracing will look better the some AAA games that do use ray tracing.

 

[bsigg]

Some next gen AAA games with no ray tracing will look better the some AAA games that do use ray tracing.

Click to shrink...

Well, yea. Art direction is hugely important to the look of a title. That said, ray tracing in the long term should really help improve lighting in games overall.

 

[Barnabas Reynolds]

Oh, so we are talking about speculative visuals now?

Why don't we just stick to the games and tech demos which have very good implementations of the technology already.

Click to shrink...

Really, Mr Quadruple The Polygons is complaining about speculation? It's pretty obvious what the gains will do to real-time graphics, as we have the Quake 2 RT conversion to look at.

 

[Anthony Hopkins]

Well, yea. Art direction is hugely important to the look of a title. That said, ray tracing in the long term should really help improve lighting in games overall.

Click to shrink...

Ray tracing is the superior lighting tech, but until devs have unlimited resources there will be pros and cons to using ray tracing.

 

[Clowns]

Yeah, this is the one apparently: https://pastebin.com/PY9vaTsR/

Click to shrink...

"third party small developer form Europe"
"here's a list of all these titles from Ubisoft, EA, Warner Bros, Take-Two, and Sony that somehow I'm in a position to know the existence of"

 

[Anthony Hopkins]

Really, Mr Quadruple The Polygons is complaining about speculation? It's pretty obvious what the gains will do to real-time graphics, as we have the Quake 2 RT conversion to look at.

Click to shrink...

What exactly are you arguing against?

You think Ray tracing is a bigger visual leap then PBR?

You are aware of deminishing returns right?
Before PBR materials looked like coloured unrealisticly lit textures, PBR was a huge leap.

While Ray tracing simulates light realistic ly its not like there were no emulation of light, reflections and shadows before Ray tracing.

 

[SharpX68K]

Absolutely nothing.

Click to shrink...

Haha, I suspected as much. Still, we're in-season for rumors, leaks, etc.

Any day now, the mother of all next-gen PS5/Xbox leaks could happen.

 

[Iced_Eagle]

What exactly are you arguing against?

You think Ray tracing is a bigger visual leap then PBR?

You are aware of deminishing returns right?
Before PBR materials looked like coloured unrealisticly lit textures, PBR was a huge leap.

While Ray tracing simulates light realistic ly its not like there were no emulation of light, reflections and shadows before Ray tracing.

Click to shrink...

Don't forget the amount of things developers explicitly chose not to do, or simply couldn't do because it exposed the issues with lighting, reflections, refractions, shadows, etc. There's so much faking built on top of the lack of ray tracing to fool the player, and because developers can control things fairly well, they didn't expose the flaws often.

With ray tracing, so much creative potential is once again unleashed because things will "just work" as they expect them to. Obvioiusly it will take time for hardware to get it to reasonable performance, and for developer's to figure out how they balance ray tracing versus rastierzation. It's somewhat irrelevant which one you view as "better" than the other, as both are inevitable in the end and both are extremely valuable to advancing gaming graphics.

Here's a video from John Carmack where he talks about lighting and rendering, with lots of talk on ray tracing, and also PBR (video is from 2013 -- relevant timestamp linked): https://www.youtube.com/watch?v=IyUgHPs86XM&t=1h06m53s

*edit* added an appropriate timestamp to the video.

 

[BreakAtmo]

Ah, I wasn't meaning closely spec'ed systems. I was talking about a system where you're disabling half the GPU cores, as an example.

Click to shrink...

Or lowered clock speeds. Like if the high-end console has 72CU at 1400MHz, the low-end uses lesser versions of that same chip at 800MHz. But it seems that's not feasible.

lmao no, you won't be getting 2080 level performance

Click to shrink...

Agreed. What I could see (if the rumors about Navi clocks being much higher is true), is a Navi card hitting that performance with an 1800-2000MHz boost clock, and a console chip using that GPU at only, say, 1400MHz.

 

[Anthony Hopkins]

Don't forget the amount of things developers explicitly chose not to do, or simply couldn't do because it exposed the issues with lighting, reflections, refractions, shadows, etc. There's so much faking built on top of the lack of ray tracing to fool the player, and because developers can control things fairly well, they didn't expose the flaws often.

With ray tracing, so much creative potential is once again unleashed because things will "just work" as they expect them to. Obvioiusly it will take time for hardware to get it to reasonable performance, and for developer's to figure out how they balance ray tracing versus rastierzation. It's somewhat irrelevant which one you view as "better" than the other, as both are inevitable in the end and both are extremely valuable to advancing gaming graphics.

Here's a video from John Carmack where he talks about lighting and rendering, with lots of talk on ray tracing, and also PBR (video is from 2013): https://www.youtube.com/watch?v=IyUgHPs86XM

Click to shrink...

I agree, I am just saying visually Ray tracing is not as big a visual leap as PBR.

 

[7thFloor]

Here's a video from John Carmack where he talks about lighting and rendering, with lots of talk on ray tracing, and also PBR (video is from 2013): https://www.youtube.com/watch?v=IyUgHPs86XM

Click to shrink...

I remember watching this video when it was new just to fantasize about the future of real-time graphics.

I agree, I am just saying visually Ray tracing is not as big a visual leap as PBR.

Click to shrink...

You're only saying this because we're not seeing fully ray-traced games yet, this is just the beginning of a transition to an entirely different rendering method.

 

[Deleted member 12635]

Albert Penello
Thanks for stopping by and having shed some light on E3 planning. Confirmed my stance that anyone who thinks he knows the plans is just doing educated guesses as we all do at this moment.

 

[TheTruthofKayle]

Ray tracing is not as big a leap as PBR.

Click to shrink...

Having true full raytracing Is the biggest leap that real time graphic will have until we will not change the formula on how 3d graphics willl be generated ,))

 

[chris 1515]

• I agree, I am just saying visually Ray tracing is not as big a visual leap as PBR.

  Click to shrink...

I agree too. This a great told but many things are needed to progress:

• Much more geometry(probably end of normal maps and parrallax mapping). Only tesselation and displacement. Much better hair and foliage, end of silouhette problem. Require a huge change of current rendering hardware. It would reduce the time to create assets.
• continuous LOD
• Better IQ, temporal is pretty good but when thing moving not perfect. Decoupled sampling from visibility?
• Much better motion blur and Depth of field effect.

Global illumination raytracing can be added to the list but is not the only things to improve.

 

[TheThreadsThatBindUs]

Ray tracing is not as big a leap as PBR.

Click to shrink...

I agree.

lmao

oh shit, you're serious?!

Click to shrink...

You are wrong.

Click to shrink...

wat

Click to shrink...

Anthony Hopkins is right -- at least from the perspective of visual fidelity in games.

Current rasterization technology is excellent and is still developing as well. It's a moving target.

RT is nice for dealing with the edge cases where rasterization and its plethora of many lighting and shadowing techniques breakdown. In the best implementations of the latter, and with a carefully designed game, those will be few and far between.

RT's biggest wins are on the developer workflow side. That's where the most of the (legitimate) hype and excitement comes from. And from what I've read from realtime rendering devs vs. CGI devs, the biggest wins are on the offline rendering side as opposed to games development.

From the perspective of the end product, the difference will be minimal.

If you look at many of the NVidia RT demos, you'll notice that the "RT-off" parts of the videos use the most obnoxiously basic stencil shadows, cube maps and other such rudimentary techniques. Whereas raster-tech in AAA has moved beyond that.

You only need to sit and play a couple of hours of RDR2 on a PS4Pro or XB1X to see how rasterization at its very best can produce some profoundly stunning results.

A fully RT'd game of the equivalent visual fidelity will require so much more compute performance (even with RTX and Turing/Tensor cores) that I personally struggle to see why it's worth it.

Also, there are a number of cases that absolute fuck over RT performance, e.g. geometry with small holes like fences and such. There still needs clever workarounds to get around these (that don't equate to "remove fence from game"). Fully RT'd games across the board aren't there yet and won't be for a little while yet.

 

[chris 1515]

I agree.





Anthony Hopkins is right -- at least from the perspective of visual fidelity in games.

Current rasterization technology is excellent and is still developing as well. It's a moving target.

RT is nice for dealing with the edge cases where rasterization and its plethora of many lighting and shadowing techniques breakdown. In the best implementations of the latter, and with a carefully designed game, those will be few and far between.

RT's biggest wins are on the developer workflow side. That's where the most of the (legitimate) hype and excitement comes from. And from what I've read from realtime rendering devs vs. CGI devs, the biggest wins are on the offline rendering side as opposed to games development.

From the perspective of the end product, the difference will be minimal.

If you look at many of the NVidia RT demos, you'll notice that the "RT-off" parts of the videos use the most obnoxiously basic stencil shadows, cube maps and other such rudimentary techniques. Whereas raster-tech in AAA has moved beyond that.

You only need to sit and play a couple of hours of RDR2 on a PS4Pro or XB1X to see how rasterization at its very best can produce some profoundly stunning results.

A fully RT'd game of the equivalent visual fidelity will require so much more compute performance (even with RTX and Turing/Tensor cores) that I personally struggle to see why it's worth it.

Also, there are a number of cases that absolute fuck over RT performance, e.g. geometry with small holes like fences and such. There still needs clever workarounds to get around these (that don't equate to "remove fence from game"). Fully RT'd games across the board aren't there yet and won't be for a little while yet.

Click to shrink...

Raytracing biggest advantage is on the production side. It would save lighting artist time. But some effect are impossible to do without it or have some problem like screenspace effect.

 

[PLASTICA-MAN]

Raytracing biggest advantage is on the production side. It would save lighting artist time. But some effect are impossible to do without it or have some problem like screenspace effect.

Click to shrink...

I started a featured map in Fortnite yesterday which had futuristic buildings which are made of glass and are supposed to reflect everything and are supposed to be reflected on the ground too. Your character reflection isn't well made because it isn't a planar one and even with SSR, when you are facing the buildings, you only see a part of them reflected on the ground but as soon as you look a little on the ground, the buildings reflections get completely vanished which is really very bothersome and doesn't need a tech expert to notice that and everyone can notice it.

Such kind of limitations in screen space will always force artists to do tricks to avoid putting reflective surfaces whether huge ones like the buildings or even tiny ones like mirrors. If next-gen will come such lame trend of broken glass, this would be completely outdated and meaningless. That is why we really hope that somehow both MicroSoft and Sony will include some kind of ray-tracing hardware in thein next-gen consoles or at least the devs need to try to circumvent such limitations voxelizing the scene to add dynamic GI, indirect shadows and world space reflections.

 

[StraightToHell]

Another PBR > Ray Tracing when it comes to visual fidelity here, if you had to choose between the two. To recap, Ray Tracing allows the path of a photon to be traced, but PBR replicates how photos are scattered when they hit a surface - which can be and has been done without RT.

Now, take a look at RT examples in older titles that don't use PBR. You get nice shadows and reflections, but surfaces still look like :flat textures:

https://www.youtube.com/watch?v=mtHDSG2wNho

https://www.youtube.com/watch?v=XVZDH15TRro

https://www.youtube.com/watch?v=CS3RvLhfBCM

 

[SharpX68K]

Albert Penello, welcome back.

I had never seen this exact video before (from E3 2017) but it's a joy to watch and listen to the questions and your answers about Xbox One X.

 

[PLASTICA-MAN]

Another PBR > Ray Tracing when it comes to visual fidelity here, if you had to choose between the two. To recap, Ray Tracing allows the path of a photon to be traced, but PBR replicates how photos are scattered when they hit a surface - which can be and has been done without RT.

Now, take a look at RT examples in older titles that don't use PBR. You get nice shadows and reflections, but surfaces still look like :flat textures:

https://www.youtube.com/watch?v=mtHDSG2wNho

https://www.youtube.com/watch?v=XVZDH15TRro

https://www.youtube.com/watch?v=CS3RvLhfBCM

Click to shrink...

The thing is that PBR is not tied to resources and ccould be applied even in last-gen games ( MGS5, Remember Me) while RT requires enough power to be correctly achieved. PBR is just the correct use physical coefficients and properties (dunno why it took so long before it became widespread) while RT requires calculating the path of rays and light in general (it has been there for a long time but still waiting for the improvement of hardware to get widespread).

 

[StraightToHell]

The thing is that PBR is not tied to resources and ccould be applied even in last-gen games ( MGS5, Remember Me) while RT requires enough power to be correctly achieved. PBR is just the correct use physical coefficients and properties (dunno why it took so long before it became widespread) while RT requires calculating the path of rays and light in general (it has been there for a long time but still waiting for the improvement of hardware to get widespread).

Click to shrink...

That's what they do, but from a player / user perspective (which is where this sub-topic started), many of us are saying that the jump in visual quality from pre-PBR to PBR is going to be more impact than the introduction of RT in future.

To take this a step further, I think the combination of more accurate PBR models plus further advancements in direct lighting, shadows, reflections and dynamic GI will be what sets the next gen apart from the current, not the introduction of RT into one or more of those areas.

 

[Gamer17]

Some new info from Reddit suggests Sony is using hbm2 combined with ddr4 and is extremely pleased with outcome for games,12gb ddr 4 and 8 gb hbm2 is reserved for games and 4 gb gddr4 for os :

24 GB RAM in total (20 GB usable by games)

8 GB in form of 2 * 4-Hi stacks HBM2

Sony got "amazing" deal for HBM

in part due to them buying up bad chips from other customers which can't run higher then 1.6 Gbps while keeping 1.2v.

HBM is expected to scale down in price a lot more than GDDR6 over the console lifetime

Samsung, Micron and SK Hynix already shifting part of their capacity towards HBM due to falling NAND prices

Sony will be one of the first high volume customers of TSMCs InFO_MS when mass production starts later this year (normal InFo already used by Apple in their iPhone)

InFO_MS brings down the cost compared to traditional silicon interposers - has thermal and performance advantage as well

InFO_MS allows them to drive their 1.6 Gbps chips @ 1.7 Gbps (435 GB/sec.) without having to increase the voltage above 1.2v

HBM is more power efficient compared to GDDR6 - the savings were invested into more GPU power

additional 16 GB in form of DDR4 @ 256 bit for 102.4 GB/sec.

4 GB reserved for OS, the remaining 12 GB usable by games

memory automatically managed by HBCC and appears as 20 GB to the developers

HBCC manages streaming of game data from storage as well

developers can use the API to take control if they choose and manage the memory and storage streaming themselves

memory solution alleviates problems found in PS4

namely that CPU bandwidth reduces GPU bandwidth disproportionately

2 stacks of HBM have 512 banks (more banks = fewer conflicts and higher utilization)

GDDR6 better than GDDR5 and GDDR5x in that regard but still less banks than HBM

at the same time trying to keep CPU memory access to slower DDR4

very satisfied with decision to use two kinds of memory for price to performance reasons

allowed them to go below ~50 GFLOPs per GB/sec. bandwidth but still keep above 40 GFLOPs per GB/sec.






He also mentions ps4 refresh is coming around September for 199 on 7nm but not pro.pro stays the same .

https://www.reddit.com/r/PS5/comments/bdfarm/rumor_playstation_5_price_leaks_and_it_aint_cheap/