The Middle Ground Approach: PC, PS4, XB1

[Scott_Arm]

Sacred 2 is developed by a real small developer. I don't know if you can extrapolate their setup to others. They also have one guy working on the PS3 renderer. That doesn't mean they don't have a larger staff working on the rest of the game/game engine for PS3. The renderer is just one piece.

Edit:
I also just did a quick search for "PS3 lead platform" and I'm getting hits for Modern Warfare 2, Rockstar Games, Metal Gear Solid: Revengeance (lol), Bioshock Infinite, Hitman Absolution. I know there was a lot of talk about the switch to PS3 as lead part way through last gen, because it required the most effort and many of the memory optimizations and work put into multi-threading job systems would benefit xbox360 as well.

 

[Billy Idol]

Sacred 2 is developed by a real small developer. I don't know if you can extrapolate their setup to others. They also have one guy working on the PS3 renderer. That doesn't mean they don't have a larger staff working on the rest of the game/game engine for PS3. The renderer is just one piece.

True. But it is not a proof of 2/3 up to 3/4 dev budget for the PS3 either...which was the question I raised in my post.

 

[Scott_Arm]

True. But it is not a proof of 2/3 up to 3/4 dev budget for the PS3 either...which was the question I raised in my post.

Yah, I don't know about that, but I think a lot of devs did use PS3 as lead, at least part way through last gen.

 

[temesgen]

I'm not seeing why it would be hard to optimize games to the relative strengths of the platform. Most games will be developed on Unreal or Cryengine or some other middleware so why wouldn't it be possible to scale the performance based off the power of the GPU?

Granted I'm mostly talking about resolution here but there should be graphical effects can be turned on for the PC and PS4 version which might be absent for the XB1 even if the PS4 version runs a lower setting.

In many ways this is already how PC games work, there are graphic settings including framerate which users 'dial in' based off their GPU. The developer can code to a reference and then based on the strength of the platform various features and resolution should be able to be increased, decreased or turned off.

If anything it should be easier this generation than ever before to produce titles that scale relative to performance. There might be some additional development time needed to address the memory setup on the XB1 but shouldn't the biggest multi-platform resource sink be due to Open GL versus DirectX with artist assets?

 

[Scott_Arm]

I think the easiest thing is to make them look pretty much identical but scale resolution. It makes the life of the artists and level designers a lot easier. For them, it's going to look the way they design it and it's not going to change based on platform settings. You might get some slight differences in quality of AA, or something like that, but otherwise things remain the same.

 

[temesgen]

I think the easiest thing is to make them look pretty much identical but scale resolution. It makes the life of the artists and level designers a lot easier. For them, it's going to look the way they design it and it's not going to change based on platform settings. You might get some slight differences in quality of AA, or something like that, but otherwise things remain the same.

Again, I think we'll see effects on the PC which are present but of lower quality on the PS4 and XB1. Developers will likely make the game (PC high settings) and then adjust for the consoles (PC middle settings).

If I'm a developer I want to use all the latest and greatest techniques so I stay relevant as well as a desire to showcase my skills. Most will likely do their best and let the engine scale it down if need be.

 

[Prophecy2k]

True. But it is not a proof of 2/3 up to 3/4 dev budget for the PS3 either...which was the question I raised in my post.

2/3 to 3/4 of the dev budget for the PS3 version of a MP game is silly and can't hold any basis in reality.

That would imply well over 3/4 of the dev team being engine and tools coders, which simply isn't going to be true. On the avereage PS360 game dev team, you'll have a small number of senior software engineers responsible for the engine, tools, rendering etc. And a fairly small sized team working under them.

The majority of the dev team however will be made up of artists, animators, gameplay programmers etc, whose work producing content comprises the lions' share, and doesn't need to be duplicated for each targeted platform. These guys will make up the majority of the dev team, and thus dev budget, and their work will be used across all versions of the game.

 

[Prophecy2k]

Again, I think we'll see effects on the PC which are present but of lower quality on the PS4 and XB1. Developers will likely make the game (PC high settings) and then adjust for the consoles (PC middle settings).

If I'm a developer I want to use all the latest and greatest techniques so I stay relevant as well as a desire to showcase my skills. Most will likely do their best and let the engine scale it down if need be.

I think this is the way to go to be honest.

And I think the relative differences in platform installed base should factor into the decision also. If as at the moment twice the number of players will be playing on the PS4 than the XB1 then I would focus more of my development targeting the PS4's HW spec, and then just drop resolution and switch off/move to less impressive graphical effects on the XB1. After all the GPU archetecture is the same, so there's little reason why your engine shouldn't simply scale easily with the relative differences in platform performance.

It makes me wonder, since in the Xbox/PS2 gen, games were developed for the Xbox utilising higher fidelity graphics and effects, despite it being a far far less popular platform in terms of sales. And even some PS360 shipped on the 360 with better lighting, AA, texture filtering, particle transparency effects, resolution etc. So why should any developer this generation think that targeting technical parity between platforms is a good idea, when the platform that is more capable is also more successful in the marketplace?

Surely you should just target the most popular console and design your game around that, then just port up/down to any others? Surely this makes the best business sense, since gamers that buy the most performant console have their expectations met, and those who choose the weaker box, can't complain because they MUST have chosen the weaker box for reasons outside its relative graphics performance on multi-platform games.

 

[Deleted member 11852]

Surely you should just target the most popular console and design your game around that, then just port up/down to any others?

Crazy talk

The advantage of targeting the weakest/most-complicated platform first is that anything you get to work will be sure to work on the other platforms(s). Where you don't want to be is finding that actually, you can't port down without breaking something fundamental.

 

[Prophecy2k]

Crazy talk

The advantage of targeting the weakest/most-complicated platform first is that anything you get to work will be sure to work on the other platforms(s). Where you don't want to be is finding that actually, you can't port down without breaking something fundamental.

But the differences in HW archetecture aren't so different that getting a game to work on the XB1 will be so much of a problem. Yes the memory system is different, but not so much that it would take an excessive amount of time/energy/money to port down to. It's nothing like the PS3-XB360 situation.

If anything, since Sony's boasts about PS4 and its "time to triangle", you could have your game running nicely on PS4 in relative no time. Then spend the majority of the rest of your dev cycle tweaking and tuning the XB1 performance to try to get it as close as you can. I'm sure this is what we have seen on the majority of MP games thus far.

 

[Deleted member 11852]

But the differences in HW archetecture aren't so different that getting a game to work on the XB1 will be so much of a problem. Yes the memory system is different, but not so much that it would take an excessive amount of time/energy/money to port down to. It's nothing like the PS3-XB360 situation.

I agree with the sentiment but bitter experience of trying to moving code X from platform Y to platform Z (often with great confidence and bravado!) then finding it just couldn't work without changing something fundamental, has taught me never to assume this.

Graphics can always be scaled back but building a complete game system - engine, logic, AI, mechanics - often built in isolation by particular coding teams who may think they can port it then find they have less resource on the target platform than they hoped - can be a tricky bugger. I imagine. I've never done games but have written a lot of software solutions where there are as many, often more, moving parts.

 

[temesgen]

I agree with the sentiment but bitter experience of trying to moving code X from platform Y to platform Z (often with great confidence and bravado!) then finding it just couldn't work without changing something fundamental, has taught me never to assume this.

Graphics can always be scaled back but building a complete game system - engine, logic, AI, mechanics - often built in isolation by particular coding teams who may think they can port it then find they have less resource on the target platform than they hoped - can be a tricky bugger. I imagine. I've never done games but have written a lot of software solutions where there are as many, often more, moving parts.

I don't think its realistic to expect a scripted event with 10 enemies on the XB1 to become 14 on the PC or PS4. What is reasonable however is for the visuals to scale with the GPU, we'll likely see more shaders, better lighting, more effects and higher resolution on the PC then less fidelity on the PS4 and XB1.

IMO PC=PS4=XB1 isn't as likely as PC>>>>>PS4>XB1, wrt graphics of course, core gameplay will be same.

 

[Deleted member 11852]

I don't think its realistic to expect a scripted event with 10 enemies on the XB1 to become 14 on the PC or PS4. What is reasonable however is for the visuals to scale with the GPU, we'll likely see more shaders, better lighting, more effects and higher resolution on the PC then less fidelity on the PS4 and XB1.

I wouldn't expect any scripted event to materially be written form platform to platform, but much much the PlayStation 4 and Xbox One versions of GTA V will figure higher density pedestrians and vehicles, that may be where overhead GPU budget is spent.

But frankly, as I posted waaaay above, I think it's waaaay too early to tell. Developers are still coding for the new consoles as they did for the old consoles with little exploitation of the new hardware like GPGPU, even where it deviates from PC norms. When developers start writing code specifically for those these new consoles, then we'll have a better idea of the real differences in performance and can then better estimate how things will play out in the years to come.

 

[temesgen]

I wouldn't expect any scripted event to materially be written form platform to platform, but much much the PlayStation 4 and Xbox One versions of GTA V will figure higher density pedestrians and vehicles, that may be where overhead GPU budget is spent.

But frankly, as I posted waaaay above, I think it's waaaay too early to tell. Developers are still coding for the new consoles as they did for the old consoles with little exploitation of the new hardware like GPGPU, even where it deviates from PC norms. When developers start writing code specifically for those these new consoles, then we'll have a better idea of the real differences in performance and can then better estimate how things will play out in the years to come.

With the exception of XB1's memory the consoles are essentially identical sans a few CUs and perhaps CPU speed. Most modern games are being developed on commercially produced middleware, art assets have to be handled by Direct X or Open GL but once those are brought over developement shouldn't be a mystery.

I fully expect effects to be tailored to the strength of the platform but this isn't like previous generations. The development for XB1/PS4 and PC should be more akin to coding across AMD and Nvidia cards which is to say they should have some idea early in the process what needs to be cut back or where they can add effects. To put it another way GTA 6 on the PC won't play different on a high end card versus a mid-level or low end one, its the visuals that scale.

 

[Deleted member 11852]

With the exception of XB1's memory the consoles are essentially identical sans a few CUs and perhaps CPU speed. Most modern games are being developed on commercially produced middleware, art assets have to be handled by Direct X or Open GL but once those are brought over developement shouldn't be a mystery.

DirectX and OpenGL are only an issue on desktop, where multi platform games sell a small fraction of what they do on console.

I fully expect effects to be tailored to the strength of the platform but this isn't like previous generations. The development for XB1/PS4 and PC should be more akin to coding across AMD and Nvidia cards which is to say they should have some idea early in the process what needs to be cut back or where they can add effects. To put it another way GTA 6 on the PC won't play different on a high end card versus a mid-level or low end one, its the visuals that scale.

Have you played Minecraft on PlayStation 3 or Xbox 360? Because the worlds are tiny representations of what they are on PC. This is an example of what I'm talking about. If you start with a grand scope of game on a powerful machine, you may find you have to make all sorts of compromises when moving it to a different platform.

 

[temesgen]

DirectX and OpenGL are only an issue on desktop, where multi platform games sell a small fraction of what they do on console.



Have you played Minecraft on PlayStation 3 or Xbox 360? Because the worlds are tiny representations of what they are on PC. This is an example of what I'm talking about. If you start with a grand scope of game on a powerful machine, you may find you have to make all sorts of compromises when moving it to a different platform.

I've not played Minecraft but unless I'm wrong the core gameplay is essentially the same across all platforms. Granted you could make something on a highend PC that wouldn't model on the consoles or tablets or even phones but that is kind of an extreme example. Most games aren't fundamentaly different today from a core gameplay mechanic than they were last generation or arguably the generation before. Its the visuals which are advancing. That was one of the core arguments about the importance of Kinect coming out of E3 a year ago. Some Xbox enthusiast said unless the core gaming mechanics were built upon with technology like Kinect console gaming wouldn't really progress beyond visuals. Perhaps they were right but in many ways stories haven't fundamentally changed in a long time and we still spend quite a bit of money on entertainment with much of that $ being thrown at the visuals/special effects.

 

[Deleted member 11852]

I've not played Minecraft but unless I'm wrong the core gameplay is essentially the same across all platforms.

If by "core gameplay" toy mean you can remove and place blocks, then yes. If you "core gameplay" you mean it's a limitless procedurally generated world spanning any number of biomes, each with their own unique attributes, then no. Minecraft on last gen consoles is a heavily paired back affair.

And that's basic out-of-the-box Minecraft.

 

[Shifty Geezer]

But the differences in HW archetecture aren't so different that getting a game to work on the XB1 will be so much of a problem.

Let's say your game struggles on XB1 when you try it. What do you change? Remove number of objects? Can't simplify geometry without redoing all your assets. Tone down shaders? Finding the bottleneck and a suitable solution could be very difficult, time consuming, and leave you with a game somewhat different to your original design. Now consider starting with XB1. Everything is exactly as you want it (design model complexity, shader complexity, lighting model, to fir the machine) and ports directly to PS4 with no worries. You can then just add stuff with no worries about breaking anything if you go for the easy wins (better AA, post ambient occlusion effect, higher res buffers, etc.).

The latter option is so much easier, it's definitely the better economy.

 

[Prophecy2k]

Let's say your game struggles on XB1 when you try it. What do you change? Remove number of objects? Can't simplify geometry without redoing all your assets. Tone down shaders? Finding the bottleneck and a suitable solution could be very difficult, time consuming, and leave you with a game somewhat different to your original design. Now consider starting with XB1. Everything is exactly as you want it (design model complexity, shader complexity, lighting model, to fir the machine) and ports directly to PS4 with no worries. You can then just add stuff with no worries about breaking anything if you go for the easy wins (better AA, post ambient occlusion effect, higher res buffers, etc.).

The latter option is so much easier, it's definitely the better economy.

You're probably right Shifty, afterall, many devs that have worked on HD remasters have openly expressed the sentiment that its much easier to up-port a game to more capable HW than to downport.

Still, quite how badly a game would scale to the XB1 if dev'd exclusively on PS4, is still up in the air. Given the scalability of modern game engines, I doubt you'd have to do any more than lower resolution, texture filtering, AO and/or switch to a cheaper less impressive AA method. It won't compromise your vision anymore than it would do having to make a PS360 version, which is a very real reality with most MP games so far this gen.

I make the assumption that a reasonable dev will begin a multiplatform project and set geometry/texture budgets for all in-game assets with the weakest target platform in mind. The rest is all runtime graphics/gameplay systems, built into the engine, which should be scalable across HW. I'm not sure that between PS4 and XB1 it's all that much of an issue.

 

[Deleted member 13524]

Let's say your game struggles on XB1 when you try it. What do you change? Remove number of objects? Can't simplify geometry without redoing all your assets. Tone down shaders? Finding the bottleneck and a suitable solution could be very difficult, time consuming, and leave you with a game somewhat different to your original design. Now consider starting with XB1. Everything is exactly as you want it (design model complexity, shader complexity, lighting model, to fir the machine) and ports directly to PS4 with no worries. You can then just add stuff with no worries about breaking anything if you go for the easy wins (better AA, post ambient occlusion effect, higher res buffers, etc.).

The latter option is so much easier, it's definitely the better economy.

Then again, if a 3rd party developer does that, they're bound to not take as much advantage of the more powerful system as they could.
For games that aren't trying to reach any pinnacle of graphics fidelity, that's perfectly fine.

However, if they're working on an AAA production competing with first-party titles on IQ in both platforms, then they're in for a bad time.
I can't imagine Infinity Ward or DICE using that approach in the long term, for example. Not after the audience sees how Uncharted 4 and The Order look like.