Current Generation Games Analysis Technical Discussion [2023] [XBSX|S, PS5, PC]

[Remij]

Consoles having a fixed target makes them easier to focus on as a baseline. This isn't new phenomenon, especially when games got so complex every developer had to account for parity in quality of release just to account for how difficult it was accounting for all the platforms, and even then it's hard to do as we see.

Yeah. I can appreciate just how much work must go into figuring out baseline performance and visual targets across a multitude of platforms. Don't get me wrong, even from my very basic understanding of the process it's perfectly understandable why they do it the way they do. I just wish it didn't have to be that way.

PCs are evolving platforms. Devs could target hardware of the future and include settings which truly utilize art and assets of a visibly higher quality, with more objects, more detail, more physics, more particles, more variety, ect ect.. Things which actually separate it visually from the baseline. I know devs aren't going to build visuals that are generationally different for a tiny fraction of the overall market they're targeting, like I said, it makes sense.. But then you start to think of future consoles, and backwards compatibility.. and you know what? They're kinda like evolving platforms too now, right? How cool would that be if these improved assets, details, density and variety.. could then be enabled for a PS5 Pro when it releases down the line, or PS6?

You know what I mean? Imagine if they created higher quality assets for PC, scaled them down for use on the base consoles, then enabled them for the Pro/future consoles.

Anyway, it's just a dream lol.

 

[Shifty Geezer]

You know what I mean? Imagine if they created higher quality assets for PC, scaled them down for use on the base consoles, then enabled them for the Pro/future consoles.

That's generally how it's been done for years. Assets are created at 'art' level and then scaled down to the target platform. That simply covers poly counts, texture res, and amount of objects (NPCs, vegetation, etc) which ends up improved on PC, although we're at a point now where these are limited by hardware and you can't much improve a game by increasing asset quality (save object amounts).

The things that really make the difference are technologies being developed and applied. These have ended up tick-box toggles on PC.

I'm not sure there's a lot that can be done differently to make PC better. The constraining factors are, and have always been, economies of developing games for target hardware. Even without consoles, games since Crysis have been made targeting a sane target and just rendered at higher resolution and framerate on bigger GPUs. In fact, there are more options now to tailor the experience and enable/disable actual rendering features than ever before.

 

[Davros]

Devs could target hardware of the future and include settings

Given how long some games take to create they sort of have to.

 

[Remij]

That's generally how it's been done for years. Assets are created at 'art' level and then scaled down to the target platform. That simply covers poly counts, texture res, and amount of objects (NPCs, vegetation, etc) which ends up improved on PC, although we're at a point now where these are limited by hardware and you can't much improve a game by increasing asset quality (save object amounts).

The things that really make the difference are technologies being developed and applied. These have ended up tick-box toggles on PC.

I'm not sure there's a lot that can be done differently to make PC better. The constraining factors are, and have always been, economies of developing games for target hardware. Even without consoles, games since Crysis have been made targeting a sane target and just rendered at higher resolution and framerate on bigger GPUs. In fact, there are more options now to tailor the experience and enable/disable actual rendering features than ever before.

I think that's a trap people get themselves into believing because they're so used to the status quo.. which is to design for consoles, then add some flourishes on top for PC, because that's the cheap way to "take advantage" of more powerful hardware.. instead of actually designing a game for it.

 

[Seanspeed]

I think that's a trap people get themselves into believing because they're so used to the status quo.. which is to design for consoles, then add some flourishes on top for PC, because that's the cheap way to "take advantage" of more powerful hardware.. instead of actually designing a game for it.

The 'cheap' way is the only practical way. Developing games is super expensive already, and that's one of the biggest problems right now in the industry. Making it worse by asking them to develop more than one main 'technical branch' of a game just to suit the ever increasing standards of PC gamers would require so much extra resources and for very little actual financial benefit, it'd just be a terrible decision.

 

[sinTheta]

Devs could target hardware of the future and include settings which truly utilize art and assets of a visibly higher quality, with more objects, more detail, more physics, more particles, more variety, ect ect..

According to Steam HW survey only 1% of users have either 4080 or 4090. You could argue that those GPUs are the hardware of the future.

 

[Shifty Geezer]

I think that's a trap people get themselves into believing because they're so used to the status quo.. which is to design for consoles, then add some flourishes on top for PC, because that's the cheap way to "take advantage" of more powerful hardware.. instead of actually designing a game for it.

Designing a game for the high end means reducing your potential market to a tiny fraction of what it could be. Where's the economic argument to do that? To spend even more creating a game than you would targeting the mainstream for far less sales?

It's not a trap but a clear path that all people in the industry have found themselves following. Unless you can present an argument for a functional business model that operates differently to what we have, there is no argument.

 

[Remij]

Designing a game for the high end means reducing your potential market to a tiny fraction of what it could be. Where's the economic argument to do that? To spend even more creating a game than you would targeting the mainstream for far less sales?

It's not a trap but a clear path that all people in the industry have found themselves following. Unless you can present an argument for a functional business model that operates differently to what we have, there is no argument.

I stated multiple times that I understand why they do what they do. My argument is that it is possible for games to be superior on PC in a way that is far more tangible than just some extra resolution, or a few more objects here and there. Obviously budgets are an issue. Time, money.. ect. I'm not arguing the economics of it.

But, let's imagine that developers actually targeted high end PCs as a baseline, and scaled down to consoles instead of up. Perhaps PC game sales and sales of high end PC hardware would surge, creating a bigger market? You need to give people reasons to buy high end hardware.. which is why you have companies like Nvidia creating proprietary technologies to differentiate themselves from consoles and the competition. With no games truly differentiating themselves with the better technologies, visuals, ect... there's less reasons for people to purchase said hardware, and thus it becomes a self fulfilling prophecy where people don't buy the hardware because the games don't utilize it, and they don't utilize it because people don't have the hardware.

It's a trap. What annoys me about the situation, is that it impedes advancements in the field of 3d graphics, innovation in games, and the advancement of hardware technologies. Look how long PCs had SSDs for... before consoles actually got around to it? What innovations could we have had with developers targeting that, ages ago? Engines would be more advanced.. we'd be further along than we are. I get that's how it goes in tech and it's a fact of life, but I don't need to like it. I'd love to see tangible differences.. and see people push state of the art again on PC. There is so much untapped potential there.

 

[Scott_Arm]

Honestly I'm not convinced that if they started with the high end and scaled down that things would look much different. Besides path tracing, I don't think anyone is sitting on some really special rendering advancements that they can't use because they don't run well on consoles. I think we're mostly at diminishing returns, where most things scale poorly if you keep pushing quality and there's some "good enough" level you can scale them down to that runs vastly better. UE5 is a little different because it scales with resolution by design (per pixel geometry, shadows and lighting regardless of resolution pre-upscale), which is why it upscales so nicely. If you render a million particles you can scale it by rendering 500,000 or 100,000 instead, and it's probably not going to make the game look vastly worse etc.

 

[Remij]

Honestly I'm not convinced that if they started with the high end and scaled down that things would look much different. Besides path tracing, I don't think anyone is sitting on some really special rendering advancements that they can't use because they don't run well on consoles. I think we're mostly at diminishing returns, where most things scale poorly if you keep pushing quality and there's some "good enough" level you can scale them down to that runs vastly better. UE5 is a little different because it scales with resolution by design (per pixel geometry, shadows and lighting regardless of resolution pre-upscale), which is why it upscales so nicely. If you render a million particles you can scale it by rendering 500,000 or 100,000 instead, and it's probably not going to make the game look vastly worse etc.

I'm not asking for a game that looks unrecognizably different, just.. tangibly superior.. not simply cleaner and smoother. Obviously tech has converged, and due to the nature of having very similar tech and architectures across the board.. things aren't going to be radically different.. But I definitely believe that a game targeting high end PC, with all it's GPU grunt, memory capacity, and CPU power.. could be tangibly superior to a scaled down console version to run well across lower end hardware.

I know it's not exactly the same, but it reminds me of back in the day when you had Arcade boards out there which were designed to run the latest and greatest Arcade games pushing state of the art visuals and features. They had massive amounts of memory allowing for things never before possible.. along with insane GPU power with visuals and features not possible anywhere else. They were purpose built games, for purpose built hardware. Consoles would get a cut down version later down the line, sometimes altered drastically with completely different visuals or heavily paired back models and effects. It was important to make BOTH versions! Completely different architectures or not.. It was important to make the Arcade versions to push the state of the art, and then it was important to bring that and scale it down to a product that could be in everyone's home.

PC has massive amounts of memory, PC has massive amounts of CPU and GPU grunt. Give more objects physics, more animations, more geometry, more variety, more particles, more deformation. I dunno... do something with it that isn't just scale factoring up something that doesn't make much of a difference past a certain point.

 

[JPT]

PC has massive amounts of memory, PC has massive amounts of CPU and GPU grunt. Give more objects physics, more animations, more geometry, more variety, more particles, more deformation. I dunno... do something with it that isn't just scale factoring up something that doesn't make much of a difference past a certain point.

Come on, this is all about money, if there was more money in making a game for the PC and just make game again for console. Like you mention with those old arcade boards. They would do, but most likely the spreadsheet does not show that.

Same with Dampf's comment, about the mesh shaders, if they hit their targets without using mesh shaders on the PC why spend time and resources making it if its not needed? Unless it brings in a boatload of extra cash.
If you as a dev, hit your goals on time and on budget, you are gold from the "investors" point of view.

PC gamers complain when they can not get their 120/240/1 gazillion fps at 4k etc. But does any dev aim to hit those numbers or is that just a side effect of having a target of 60/30fps on console and trying to accommodate none bleeding edge PC hardware to?

 

[Dampf]

Same with Dampf's comment, about the mesh shaders, if they hit their targets without using mesh shaders on the PC why spend time and resources making it if its not needed?

Because it's more performance. You might not need it on your fancy new 4090, but people who do not run high end specs but use modest hardware, like the majority of PC users, that matters very much as the consoles are the baseline. A 2070 Super for example, would be on par with the Series X in normal circumstances, thus have great performance and image quality. Without Mesh Shaders however, it can no longer sustain that same performance target.

So it matters. Very much so.

 

[JPT]

Because it's more performance. You might not need it on your fancy new 4090, but people who do not run high end specs but use modest hardware, like the majority of PC users, that matters very much as the consoles are the baseline. A 2070 Super for example, would be on par with the Series X in normal circumstances, thus have great performance and image quality. Without Mesh Shaders however, it can no longer sustain that same performance target.

So it matters. Very much so.

But does it matter in regards to the amount of money that effort would earn them vs the cost?
I have no idea how Avatar runs on a range if PC hardware, but they are praised for the scaling, so are you saying it runs bad unless you have really powerful hardware? Like worse than XSS? Because its about hitting your targets on time and on budget that matters. Adding mesh shaders to pc is stupid if they run over budget or they have to drop something else. Especially when they claim their none mesh shader path is just as fast....

 

[trinibwoy]

Because it's more performance.

What's your evidence for this? The developers themselves claim mesh shaders weren't easy to get running faster than the vertex shader path on consoles.

 

[techuse]

When looking at game materials, what is the biggest limitation keeping them from looking photorealistic? I'm sure there are many reasons but what would be the biggest one? Is it purely an artist time/asset size issue? texture resolution? surface lighting calculations?

 

[Davros]

Designing a game for the high end means reducing your potential market to a tiny fraction of what it could be

Only if you take the short term view

 

[Deleted member 11852]

Only if you take the short term view

I'm not sure I follow. Are you thinking that publishers should invest hundreds of millions, or billions, in cutting edge AAA games for PC to sell it at launch for $70 to 1% of the user base, hoping that in 4-5 years when PC owners have upgraded their hardware, more will buy it for $10?

It feels like plenty of people already wait to get games in sales, so the only difference would be choosing to limit who can buy it at launch by requiring the highest performing hardware.

 

[Shifty Geezer]

I'm not asking for a game that looks unrecognizably different, just.. tangibly superior.. not simply cleaner and smoother. Obviously tech has converged, and due to the nature of having very similar tech and architectures across the board.. things aren't going to be radically different.. But I definitely believe that a game targeting high end PC, with all it's GPU grunt, memory capacity, and CPU power.. could be tangibly superior to a scaled down console version to run well across lower end hardware.

In what way? As I say, assets are already created at perfect quality and scaled down for all platforms. What else are you asking the devs to do? If they do something the consoles literally can't, then you have a devil of job porting - effectively have to rewrite the game.

Consoles would get a cut down version later down the line, sometimes altered drastically with completely different visuals or heavily paired back models and effects. It was important to make BOTH versions!

The level difference attainable wasn't also the difference between a $10 million console game and a $100 million arcade game. Games cost proportionally a lot more, the work of hundreds of people over years instead of a few over months. Nowadays you want games to reach the widest audience - including Sony now making cross platform versions of its flagship titles!! Even Sony won't limit itself to a niche despite having the biggest and most lucrative niche for it's AAA that'll drive hardware adoption.

 

[Seanspeed]

I feel a lot of this arguing is basically just traceable back to wish fulfillment("Man it'd be nice if developers were actually making games specifically for my super expensive PC first instead of consoles"), and then trying to desperately rationalize some world in which it might possibly make sense.

Y'all just need to enjoy titles like Star Citizen and Microsoft Flight Sim and be grateful such high end unicorn games still exist at all in a world where it simply doesn't make sense for the vast majority of devs/pubs.

 

[techuse]

In what way? As I say, assets are already created at perfect quality and scaled down for all platforms. What else are you asking the devs to do? If they do something the consoles literally can't, then you have a devil of job porting - effectively have to rewrite the game.

The level difference attainable wasn't also the difference between a $10 million console game and a $100 million arcade game. Games cost proportionally a lot more, the work of hundreds of people over years instead of a few over months. Nowadays you want games to reach the widest audience - including Sony now making cross platform versions of its flagship titles!! Even Sony won't limit itself to a niche despite having the biggest and most lucrative niche for it's AAA that'll drive hardware adoption.

How feasible would it be for devs to allow PC gamers the option to use some of the perfect quality assets? For example, offering the Drake model and extremely high res flora in the UC4 teaser trailer as an optional, high res asset download. Since PC games are all digital there is no need to worry about the storage limits of a disc. I guess it's just a matter of if the dataset becomes too large for the engine to handle in real time?