Do exclusive developers push visuals more than AAA developers?

Status
Not open for further replies.
Rendering techniques are tools to be used in order to create a world for the player to be immersed in, the hard part is balancing all that to achieve something visually coherent while at the same time keep the performance in check. Even if you are using the most impressive rendering techniques, if you get one thing wrong then your world collapses. I will bring up AC:Unity once again, one of the most impressive looking games at its high points, there were times when i would just stop and wander at the little details, the combination of every element to make the game look right , examples:

15650566687_9eb7fb8027_o.png
15650554297_36bf2f0a39_o.png
15649771369_c4af6be92d_o.png
15650159728_3fd9348f53_o.png
15215758264_f97f07efda_o.png

It really looks impressive doesn't it? But then, i had to move, sometimes run, and there, my immersion is gone. NPCs clipping into each other, some decided to turn into ghosts and walk through buildings, others decided they were spider-man and they walked in ways they shouldn't be able to, when i was walking past a huge crowd i could see people changing hairstyles and body types in ways i couldn't explain. And then comes the performance issues, which thankfully i never experienced first hand as i only played the PC version on a high-end PC, which is another layer on top of all the NPC shenanigans. In the end, i can't possibly bring myself to call AC:Unity one of the best games visually, no matter how impressed i was at times, because it's not coherent, the highs and the lows are worlds apart.
 
Another example of what i am talking about is Ori and the Blind Forest. A visually stunning game imo. Or Cuphead.

Are you talking about visually/artistically impressive or exclusive developers getting more out of the hardware than multiplatform developers? Because I thought the topic was the second of those options and if that's the case then you absolutely have to factor in the overall world size and openness into any assessment of a games technical merits.
 
Are you talking about visually/artistically impressive or exclusive developers getting more out of the hardware than multiplatform developers? Because I thought the topic was the second of those options and if that's the case then you absolutely have to factor in the overall world size and openness into any assessment of a games technical merits.

Sure, you can factor world size but not any more than you factor correct body and facial animation. I think the problem in this thread is that many people value different things in different ways, like we talked about few pages ago. For example, I value animation, VFX_Veteran likes correct use of AO more than i do. If he prefers a game that is using AO in the best possible way to date and i prefer a game with the best animations, can we really find a middle ground? How many times have we seen people downplay the visuals of a game because it's not open world? Should we start doing the same to open world games because they are not as huge as No Man's Sky too?
 
Another example of what i am talking about is Ori and the Blind Forest. A visually stunning game imo. Or Cuphead.
That there proves the futility of the current line of argument. "I think FC4 looks better/worse than UC4." When a 2D game that's not technically demanding can look gorgeous, how can better/worse visuals really be compared? The visuals are capped by what the engine is doing.

I suppose you're interpreting the opening question as specifically the results on screen regardless of game type. If we don't take game style and context into account (open world won't look as good as linear, but can still push the hardware every bit as much), how do we compare different games and what the devs of those games are accomplishing?
 
This thread is a stinkhole of apple-orange comparison and 'my favourite visuals' flag-waving.

You should be comparing games in as close genres as possible, multiplatform to exclusive (on the same platform) with roughly comparable team/budgets to decide if exclusive developers are getting more out of the hardware than multiplatform developers.
 
This thread is a stinkhole of apple-orange comparison and 'my favourite visuals' flag-waving.

You should be comparing games in as close genres as possible, multiplatform to exclusive (on the same platform) with roughly comparable team/budgets to decide if exclusive developers are getting more out of the hardware than multiplatform developers.

But it's very hard to do that, even if we take Tomb Raider for example which was multiplatform and now will be, at least for a while, exclusive on X1. If we were to compare what CD achieved with Tomb Raider Remastered on X1 with what they'll ship in a few months on the same platform on a modified engine, we would also have to consider the funding and technical help they received from Microsoft and all the improvements on their engine. It's really hard to take specific situations like these and say compare teams with the same funding and technical knowledge in two separate games. So again it comes down to personal opinion on the matter.
 
It looks visually in the same ball park as UC4 to me

Yeah, I understand, I even met people who still think that Gears of War and Order 1886 look "in the same ball park".
But I don't think I want to discuss that, as I've already said. :)

open world game

All Uncharteds are "open world games" for the sake of graphics, because they heavily use streaming and all their assets are dynamic.
Nobody renders "the whole world" with the same LOD ever. Therefore there is almost no difference between "open world", where your movement is constrained by streets, and "linear" game along the same streets.
Even more so in Drake 4, where you can drive in multiple paths to multiple places.
 
Yeah, I understand, I even met people who still think that Gears of War and Order 1886 look "in the same ball park".
But I don't think I want to discuss that, as I've already said. :)

I know how you feel. I've met people who think Just Cause 3 looks like a PS3 game. What a world we live in eh?

Sure, you can factor world size but not any more than you factor correct body and facial animation. I think the problem in this thread is that many people value different things in different ways, like we talked about few pages ago. For example, I value animation, VFX_Veteran likes correct use of AO more than i do. If he prefers a game that is using AO in the best possible way to date and i prefer a game with the best animations, can we really find a middle ground? How many times have we seen people downplay the visuals of a game because it's not open world? Should we start doing the same to open world games because they are not as huge as No Man's Sky too?

I agree it's largely open to personal preference/bias. So that basically concludes the thread.

Question: Do exclusive developers push visuals more than aaa developers?

Answer: Not in an objectively demonstrable way (technical stats/evidence) or in a subjectively demonstrable way (majority consensus).

There, done.
 
Nobody renders "the whole world" with the same LOD ever. Therefore there is almost no difference between "open world", where your movement is constrained by streets, and "linear" game along the same streets.
Even more so in Drake 4, where you can drive in multiple paths to multiple places.

To some extent, that is true, especially considering how a few gens ago many games simply used a system where a single big mesh described the whole level, with an oclusion system culling individual polys.

But still, there is a large increase in complexity between a linear game with streaming and a open world one. The order, speed and types of assets to be streamed is WAY more predictable in a linear game, along with all the hacks and micro-optimization that can be done on the content creation side when you know a certain location will be used for that one mission only and nothing else.
 
Question: Do exclusive developers push visuals more than aaa developers?

Answer: Not in an objectively demonstrable way (technical stats/evidence) or in a subjectively demonstrable way (majority consensus).

That is in direct contradiction with what CDPR said though

http://www.eurogamer.net/articles/2...he-witcher-3-graphics-downgrade-issue-head-on
Developing only for the PC: yes, probably we could get more [in terms of graphics] as there would be nothing else - they would be so focused, like if we would develop only on Xbox One or PlayStation 4. But then we cannot afford such a game.

I think it's safe to say that developers working only on one system can squeeze more out of it, the same can be achieved in multi-platform games too but requires more time, which means inflating the budget. I think that's why we see so many released games being 1080p on Ps4 and 900p on X1 with little to no differences, time and money don't allow for two completely separate builds for the two platforms. So the question should not be whether FP devs can push graphics more on the same system but why do they.
 
is WAY more predictable

It's always predictable, the only difference is what is the configuration of cells that you will load.
I would say: if you have something "unpredictable" in your render pipeline - you're doing it wrong.
On the other hand HDDs nowdays are so slow, compared to RAM, that these micro-optimizations do not matter much, you can load probably 1.5mb per frame in the best case (using in-place decompression and other black magic).
It essentially means: some mesh data and maybe couple of maps, not more than that. Most of the maps and meshes will stay resident all the time.
P.S. side note for PC: streaming on PC is a total PITA, you need to:
a) async read from HDD into block-aligned buffer
b) create DX-aligned buffer/texture and copy from block-aligned
c) pass it to driver and it will copy it to the GPU local memory
I.e. three memory copies for one asset.
 
The thread was pointless because the question was not the right one it would have been better to ask do an exclusive game push further a hardware than a multiplatform one?
That's also a complicated question. I can write a shit loop that'll hammer any CPU and a shit shader that'll make any GPU beg for mercy. Really the questions needs relative efficiencies to be weighed by these are either not measurable, or measurable only in subjective terms.

I agree the thread is pointless, though! ;)
 
That is in direct contradiction with what CDPR said though

http://www.eurogamer.net/articles/2...he-witcher-3-graphics-downgrade-issue-head-on


I think it's safe to say that developers working only on one system can squeeze more out of it, the same can be achieved in multi-platform games too but requires more time, which means inflating the budget.

I believe this is the crux of it. Sure if you have a fixed budget and team size then you're naturally going to get a better result focusing on one platform rather than 3, that's just common sense.

But that in no way equates to a rule (that some people suggest exists) that the highest tier graphics will always come from exclusive devs. As you say, simply increasing budget (team size) and/or time on a multi platform game can produce the same results. And since multi platform games have larger markets and thus potentially larger revenue they can also potentially afford the larger budget without negatively impacting profit.

I think that's why we see so many released games being 1080p on Ps4 and 900p on X1 with little to no differences, time and money don't allow for two completely separate builds for the two platforms.

Or they are basically the same machine with one having a little more horsepower to the tune of around the different in those two resolutions. It's not like you can do some particularly clever tricks with the Jaguar or GCN architecture with one console that you can't do with the other (okay perhaps the esram does mix things up a little there).
 
That's also a complicated question. I can write a shit loop that'll hammer any CPU and a shit shader that'll make any GPU beg for mercy. Really the questions needs relative efficiencies to be weighed by these are either not measurable, or measurable only in subjective terms.

I agree the thread is pointless, though! ;)

Yeah but you'd have to write that shit loop twice whereas if you only had to write a shit loop for one system you'd save half the time. Then you'd also have Microsoft or Sony helping you on exactly how you can write that loop in order for it to completely hammer the CPU and GPU on your system of choice. I think that's the point some of us are trying to make. It should be obvious that writing code for multiple systems also brings more difficulties when optimizing for each platform specifically and all the platforms put together. If that wasn't the case each game could have a Ps3/360/Wii U/Android/iOS/Ps4/X1/PC release.

But that in no way equates to a rule (that some people suggest exists) that the highest tier graphics will always come from exclusive devs. As you say, simply increasing budget (team size) and/or time on a multi platform game can produce the same results. And since multi platform games have larger markets and thus potentially larger revenue they can also potentially afford the larger budget without negatively impacting profit.

Yes that is possible. That all i was saying in this thread actually.
 
Or they are basically the same machine with one having a little more horsepower to the tune of around the different in those two resolutions. It's not like you can do some particularly clever tricks with the Jaguar or GCN architecture with one console that you can't do with the other (okay perhaps the esram does mix things up a little there).

The difference in OS, API and RAM/CPU core allocation should be enough to make them different enough to be treated as different platforms, the thing is, we haven't really seen that this gen! Most multi platform games look almost the same with the only difference being res and maybe performance. Someone brought up for example Second Son, which has a 80mb frame buffer, this game needs to be changed quite a bit to fit that in the eSRAM of the X1. Plus when you develop in multiple platforms you also have to think about the lowest common denominator; why are people so happy that games are starting to move into the current gen, leaving last gen systems behind? As the guys from Ubisoft said

http://www.dsogaming.com/news/ubiso...-from-cross-generation-development-decisions/
Our mandate as a graphics team was to lead on current-gen, keeping the last-gen engine the same as what shipped Far Cry 3. That gave us a lot of constraints as we had to keep the last-gen working, which affected a lot of our decisions throughout the project. By the end of the project, as was probably inevitable, we had to go back to the PS3 and Xbox 360 and polish those up, but it meant that we had a better product than FC3 on all platforms.

Would a multi-platform developer design a game without thinking about X1s eSRAM constraints like I:SS? I don't think so, as that would require two completely different builds to be made for the same game, or one of the builds to be considerably inferior to the other one (see Ground Zeroes). And to be honest, i am completely fine with multiplatform games being 1080p on Ps4 and 900p on X1 if that means that the extra budget goes to actually making the game better. So far a very small amount of multi-platform games, if any, are significantly different between the X1 and Ps4 and most of them run on FOX engine.
 
Status
Not open for further replies.
Back
Top