• 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.

"Unreal Engine is Killing Games" - Vex

Zen2 consoles are simply not good enough
I understand where you're coming from, but I disagree that it's the consoles fault since UE5 games are a problem on PC too (although to a lesser extent). I'm not sure if it's developers not having the time to fully optimise their games, UE5 being a cunt to work with or a mixture of both.

Someone else mentioned above that UE5 has this generic look and I agree. I generally prefer it when developers use other engines, although I'm sure with the right artists, they could make the games looks incredible.

I don't get why Id don't license IDTech since it's a beast of an engine that looks awesome and performs really well. It's super scalable too from what we've seen with Doom Eternal.
 
Last edited:

Felessan

Member
The war crap? Fuck off
It's really a weak excuse, cause GSC relocated itself and most of their staff to Prague 2 years ago. They didn't do their dev in war environment

there's one thing Epic is really good at, making impressive tech demos. I can't believe the whole industry falls for their bait
Whole industry just calculated how much it'll cost in time and money to maintain their own engine and decide that it's easier to buy engine from a team that dedicate itself to it instead of adding 1-2-3 years and several tens of millions to project to try to do it themselves
As tech adds more and more features it becoming more and more expensive and time consuming to maintain in-house engine
 

ScHlAuChi

Member
UE5 games come with performance issues and the only commonality is the engine used. Whether or not you like using it is irrelevant and I do agree that it's nice to use. However, it provides subpar results especially on PC. UE4, shader compilation stutter, traversal stutter, poor results in general. UE4-5's renderer was single threaded up until 5.5. This was primarily responsible for the very poor performance. 5.5 improved things but there are still issues.
It isnt the fault of the engine that it provides subpar results on PC, go knock on Microsoft´s door and ask them why DirectX isnt up to scratch!
As for poor PC optimization, publishers simply dont want to spend the money to optimize for PC because it isnt the main platform for AAA games!
Unreal Engine cant fix the unwillingness of publishers to pay for that!

The promise of UE is to be easy to use, scalable across multiple platforms and it's performant. Unreal engine only gets a 2/3 score on it's promise. However, the one it fails horribly at happens to be the one that affects the end consumer..
It fails horribly according to you, and yet more and more companies (like CDProject Red) are switching over to it!
Why do you think that happens? Because they are all incompetent and dont listen to experts like you?
Have you even worked on any AAA game and know what it means to ship such a game? Clearly not....

The blame falls primarily on Epic and secondarily on developers. If many devs big/small are having issues delivering performant games with the engine, then it's the manufacturers fault. The secondary blame falls on devs for choosing to go with UE5 if they're not willing to partially rewrite the renderer and deal with other quirks that the engine has. Blaming the publisher is laughable and is really passing the buck on personal accountability / craftsmanship.
I already explained what the issue is, but if you need EPIC as boogeyman to feel better, fine!
Also, blaming the developers is laughable and shows you have no friggn clue what you are talking about!
In the 22 years I´ve worked in this industry, I can tell you one thing:
No developer wants to release a bad or broken game, they ALWAYS want to release optimal games - it is the money ppl that stop them.
But feel free to believe whatever you want :)
 
Last edited:
It isnt the fault of the engine that it provides subpar results on PC, go knock on Microsoft´s door and ask them why DirectX isnt up to scratch!
As for poor PC optimization, publishers simply dont want to spend the money to optimize for PC because it isnt the main platform for AAA games!
Unreal Engine cant fix!
It's Microsoft fault on pc yet some other engines deployed in games on pc don't have this issue... As if lmao. It's not Microsoft fault for Epic's poor engine performance just as it's not Sony's fault on consoles. UE3 was poor on consoles often having some performance issues. UE4 was poor on all platforms, UE5 was poor on all platforms. 3 engine iterations in a row? Yea that's on epic and their trash code. Although I do sympathize with them as their engine is used for a whole lot of things, from games, to movies to the professional industry. They need to cater to the needs of off their clients.

It fails horribly according to you, and yet more and more companies (like CDProject Red) are switching over to it?
Why do you think that happens? Becasue they are all incompetent and dont listen to random forum posters on NeoGAF?
Have you ever worked on any AAA game and know what it means to ship such a game? Clearly not....
According to consumers, the people who actually buy the product. As to why companies are switching over, it has absolutely nothing to do with the engine's prowess and everything to do with dollars and cents. By switching over to UE5, you have access to a large pool of potential hires. This means shorter lead times for projects, the ability to structure your hiring around contractors to save money. It gives you the ability to save money on training new hires. It's a well documented engine. You save developement costs as the engine is actively being developed by a 3rd party. You save on maintenence costs as that is managed by epic. The list of financial benefits goes on and on. Anyone who thinks the massive switch over has anything to do with technology is completely deluded.
I already explained what the issue is, but if you need EPIC as boogeyman to feel better, fine!
But in the 22 years I´ve worked in this industry, I can tell you one thing:
Developers ALWAYS want to release optimal games - it´s the money ppl stopping them.
That's a lie. Studios have horrible self control, scope creep, etc. They often blow through their alloted budget and then try to pass the blame to publishers. They're often inefficient in the use of time and resources often having to crunch at the last minute to meet deadlines. Outside of the construction industry, I can't think of another industry that is more irresponsible when it comes to project management.
 
Last edited:

Hugare

Member
I blame weak devs

0*X0IJurzM0jxaTmfP.jpg


Change "men" to "devs" and "times" to "engines"

UE 5 (and upscaling techniques) have simplified game development, creating lazy devs

People really needed to know how to code in order to make proprietary engines. Not anymore.
 

ScHlAuChi

Member
It's Microsoft fault on pc yet some other engines deployed in games on pc don't have this issue... As if lmao. It's not Microsoft fault for Epic's poor engine performance just as it's not Sony's fault on consoles. UE3 was poor on consoles often having some performance issues. UE4 was poor on all platforms, UE5 was poor on all platforms. 3 engine iterations in a row? Yea that's on epic and their trash code. Although I do sympathize with them as their engine is used for a whole lot of things, from games, to movies to the professional industry. They need to cater to the needs of off their clients.

According to consumers, the people who actually buy the product. As to why companies are switching over, it has absolutely nothing to do with the engine's prowess and everything to do with dollars and cents. By switching over to UE5, you have access to a large pool of potential hires. This means shorter lead times for projects, the ability to structure your hiring around contractors to save money. It gives you the ability to save money on training new hires. It's a well documented engine. You save developement costs as the engine is actively being developed by a 3rd party. You save on maintenence costs as that is managed by epic. The list of financial benefits goes on and on. Anyone who thinks the massive switch over has anything to do with technology is completely deluded.

That's a lie. Studios have horrible self control, scope creep, etc. They often blow through their alloted budget and then try to pass the blame to publishers. They're often inefficient in the use of time and resources often having to crunch at the last minute to meet deadlines. Outside of the construction industry, I can't think of another industry that is more irresponsible when it comes to project management.
Ok, so lets recap:
- Epic has a poor engine and does trash code
- Studios only switch because its cheaper and easier to find people
- Publishers are the victims of terrible developers

Ladies and gentleman, I present you:
The industry expert FalconPunch!
*applause*
 
CDPR adopting Unreal Engine 5 will be the best thing to happen to the engine. They're already solving a lot of issues with the engine regarding loading and asset streaming.. and merging the changes into the default engine.

It sucks that CDPR are dropping a clearly awesome engine, but I believe that them doing so will make UE better for many other studios and games in the future.
 

Lysandros

Member
For me the best game engine is the most efficient one. The one which enables as impressive results while using less hardware resources and developer time/effort than competing engines, or more impressive results while using the same amount of resources. The 'Jack of all trades' Unreal Engine, fails to meet this criteria. Besides, you would never get the most of a specific platform/application without a tailor made custom engine.
 
Last edited:

Wolzard

Member
how can the most well used engine in the industry have aspects of it undocumented still?

Engine too large, Epic probably makes its developers focus on releasing new features rather than documenting the engine. Furthermore, as there are many different use cases, it may not be possible to predict some situations.

On the other hand, studios that discover hacks on the engine probably do not share this information with other studios (unless they are from the same group, like Microsoft studios).
 
I wish I could find that enlightening keynote presentation about unreal engine. It really makes you think. There could possibly be NO ONE still at Epic that fully understands the engine code base. Traversal stutter will never be fixed because they simply do not have anyone who even knows where to begin.
 
CDPR adopting Unreal Engine 5 will be the best thing to happen to the engine. They're already solving a lot of issues with the engine regarding loading and asset streaming.. and merging the changes into the default engine.

It sucks that CDPR are dropping a clearly awesome engine, but I believe that them doing so will make UE better for many other studios and games in the future.
Is there actually a process by which a studio that has licensed UE can send updates back to Epic and have Epic integrate it into UE? Or am I misunderstanding what you're saying?
 
Is there actually a process by which a studio that has licensed UE can send updates back to Epic and have Epic integrate it into UE? Or am I misunderstanding what you're saying?
Yeah. You can give feedback and submit code to Epic and I suppose if it's worth their while someone will review it and see what issues it solves, and what new issues it could potentially create.. and then decide to merge it into the engine. However they're not going to just take any old code from some small or unknown devs and merge it. Large studios with well respected programmers and engineers will be in direct contact with the right people from Epic, and even then it's apparently a fairly involved process because of how many different things they have to consider. After all, it's not just game developers who use the engine, it's film and media production and VFX, professional visualization and simulation and so on.. the automotive industry as well.. so they have to consider how it could affect all these different aspects of the engine before anything becoming default in the engine.

But what I do know is that The Coalition has contributed lots over the years, and CDPR have been implementing improvements as well which have been merged. Obviously CDPR have some incredible engine programmers on their team and so they're going to do everything they can to tailor Unreal Engine to their huge open world games, and a lot of those changes will help make the engine better for everyone else.
 

Zathalus

Member
I'm guessing you haven't played the game, there are things missing from this game that are pretty basic. This particular Ukraine War stint hasn't been going for 15 years. Don't use war as the sole excuse it is this way. The game was rushed out the door no matter which way you slice it. Yes, I feel for them, what they had to go through, it is bad and has been difficult but don't give me that 'oh the war crap'
15 years? The game started development in 2018 with UE4. They then switched to UE5, COVID happened, then the largest war in Europe since WW2, and then they dealt with rather severe cyberattacks that manage to penetrate and steal information and likely cause other havoc. I’m surprised the game is not even more broken.
 

Justin9mm

Member
15 years? The game started development in 2018 with UE4. They then switched to UE5, COVID happened, then the largest war in Europe since WW2, and then they dealt with rather severe cyberattacks that manage to penetrate and steal information and likely cause other havoc. I’m surprised the game is not even more broken.
I'm not going to disagree with all the difficulty they have had which I cannot imagine as I've never experienced living in a war zone. But they could of held off on releasing for a little while longer to iron out some major issues. They are obviously still able to work on the game in whatever capacity. I'm sorry but I don't think the war magically stops them from fixing certain bugs and issues. If they are able to release patch updates, they are obviously still able to actively physically keep working on the game. I actually know someone who is Ukranian with family in Ukraine, I know it probably depends on where in Ukraine you are, but they are still going about their normal lives as best they can. Some people on here sound like the devs are sitting in a prisoner camp chained to a desk.
 
Last edited:

Men_in_Boxes

Snake Oil Salesman
Epic Games, the 20 billion dollar game / engine company, doesn’t know how to create a game engine.

Not only that, but all the other game companies who choose Unreal to create their games on, don’t know what they’re doing either.

Stop this nonsense immediately
 
What it's killing imo is game-making knowledge in general. I'm currently building a graphics engine from scratch and it is absolute hell but I'm am learning so much in the process and enjoying every minute of it and I'm going to be a higher skilled developer because of it.

When I was in undergrad all the best programmers I worked with were the ones who had heavy math/physics backgrounds. I imagine it's very similar working at a company where you've got people who know how to use UE and then you've got the people who've actually built their own engines and I know for damn sure who I would want on my squad.
 
Yeah. You can give feedback and submit code to Epic and I suppose if it's worth their while someone will review it and see what issues it solves, and what new issues it could potentially create.. and then decide to merge it into the engine. However they're not going to just take any old code from some small or unknown devs and merge it. Large studios with well respected programmers and engineers will be in direct contact with the right people from Epic, and even then it's apparently a fairly involved process because of how many different things they have to consider. After all, it's not just game developers who use the engine, it's film and media production and VFX, professional visualization and simulation and so on.. the automotive industry as well.. so they have to consider how it could affect all these different aspects of the engine before anything becoming default in the engine.

But what I do know is that The Coalition has contributed lots over the years, and CDPR have been implementing improvements as well which have been merged. Obviously CDPR have some incredible engine programmers on their team and so they're going to do everything they can to tailor Unreal Engine to their huge open world games, and a lot of those changes will help make the engine better for everyone else.
CDPR better get a discount on UE! Haha
 

ap_puff

Member
Isn't the problem with UE not the fact that the engine is bad, but it's like the engine is an entire toolbox and game devs are bringing the entire box with them when their game only needs a hammer and nails?
 
CDPR adopting Unreal Engine 5 will be the best thing to happen to the engine. They're already solving a lot of issues with the engine regarding loading and asset streaming.. and merging the changes into the default engine.

It sucks that CDPR are dropping a clearly awesome engine, but I believe that them doing so will make UE better for many other studios and games in the future.
Wasn't there some article/leak about the reason why they dropped the CDprojkt engine because all of the developers who knows how to use it left th company?
 

00_Zer0

Member
I understand that this is a thread about UE5, but if you look back at UE4 there are plenty of games that were a stutter fest even in that engine. I think it comes down to game optimization. If you look at Pikmin 4, it amazes me that the game runs at a smooth non-stuttery 30 FPS. For a mobile platform like Switch that is amazing considering that engine is known for stuttering.

I think that many devs just don't care to take the time to optimize their games anymore. It's a shame that this happens, but when even AAA games sell in the millions with shit performance people still buy the games giving the developer no incentive to rework their games in a more playable state. I'm still holding off on buying Jedi Survivor because of the hitches and uneven frame pacing on PC even with the recent patch.

Epic is an easy target to hate because of their store front, their near infinite dedication to Fortnite, and a popular engine like Unreal, but not even Valve, or Unity can match what Epic has accomplished with the support of their engine through the years. You can't say that about Valves Source 1 or 2 Engine or the Unity engine. It's no wonder Unreal has the lions share of developers using their modern engine, because the support and updates they get are unprecedented in the industry. and until a competitor comes in and changes that it won't change the status quo.
 

Husky

THE Prey 2 fanatic
Performance aside, I hate how games made in Unreal almost always feel lifeless compared to comparable games built in other engines. Obsidian's games are a huge example of that, as they went from Fallout: New Vegas, where countless small objects had physics and every corpse could be dragged, to The Outer Worlds, where ammo packs would float if you grabbed an item it was on top of, and corpses can't even respond to explosives. Avowed's gameplay trailers look equally lifeless. I wish Bethesda would let them use their engine.

The biggest counter I can think of is Batman: Arkham Knight, which used a heavily modified version of Unreal 3 (and which I think still visually surpasses many Unreal 5 games). There are some real nice destruction physics when driving the Batmobile.
 

BlownUpRich

Neo Member
Unreal Engine 5 was designed for running 30 fps games 1440p on this console gen, just like Matrix Demo shows.

And just like it was with Unreal Engine 4, Unreal Engine 3, Unreal Engine 2, etc

Was Gears of War running at 60 fps in Xbox 360 when it was released? No.

Was Unreal Championship on Xbox running at 60 fps? No.

Batman Arkham Knight at 60 fps on Ps4? No.

But now people wants 60 fps so Unreal Engine 5 is shit.

I think Unreal Engine 5 is fine... the current videogame users are who dont understand 60 fps needs a lot more power tan 30 fps.
Truer words were never spoken. All these UE5 demos that ran on PS5, the 2020 one & the Matrix one were running at 30 FPS, this engine is meant for 30 FPS on current-gen consoles as the baseline, not freaking 60 FPS, UE5 can literally create near-photorealistic graphics & animations (see Matrix, 2020 PS5 demo & the upcoming Marvel 1943 game demo they showed), and people expect those visuals to run at 60 FPS on PS5 & Series X? It's impossible.

Where did this newborn demand for 60 FPS started? Like holy crap I remember when 30 FPS was acceptable all the time so long as games are well-optimized with good frame-pacing, people don't realize that if those demos were running at 60 FPS on these consoles, they'd never be as impressive as what we have now, never.

Games like RDR 2, TLOU 2, Spider Man 2018, Shadow of the Tomb Raider, Uncharted 4 & many other 30 FPS games from last-gen would've never be like what they are had they run at 60 FPS. People don't realize that & it drives me nuts.

Thank you for this comment, Seider.
 

BlownUpRich

Neo Member
The issue isn't with the Unreal Engine itself but with the current priorities in game development. The excessive focus on achieving 60 FPS and smooth performance often comes at the cost of creative vision and technical innovation. This obsession with frame rates leads to games that feel uninspired, visually bland, and lacking in impact. Ultimately, it is not the Unreal Engine that is ruining games but the misguided priorities of developers and the expectations of the community.

In short, a big thank you to all the 60 FPS fanatics for turning this console generation into a soulless, uninspired mess.
So true, and then they ask & wonder "where is the generational leap?"

My guy, it's all in there, I'm like it's all being used in the frames per second that you obsess over so much, why do I care if GOW Ragnarok is running at a blistering 100 FPS on a console? I'd get a PC for that, consoles are the baseline for a gaming generation, but since devs are willing to run their games at 60 FPS, meaning much less development cost in creating new innovative tech that push the medium forward, more asset/texture reuse of previous generation games but a bit better, also since those games run at 60 or higher, they can, in some way, be ported down to the next-generation Switch, so now the console baseline is being skewed a bit, it's it's deviating & I don't like it.

We only have GTA VI to look forward to & maybe some others...
 
Ok, so lets recap:
- Epic has a poor engine and does trash code
Here's what Epic themselves had to say about this topic in the announcement of 5.5:
For 5.5 there are improvements to parallel translation, which issues RHI (Render Hardware Interface) tasks to translate RHI command lists into platform command lists. The impact of this change is a dramatic performance increase of up to 2x (dropping by 7ms on some platforms), reducing the number of stalls, and offering a minor reduction in drawcalls as well as small improvements to platform dynamic res and render thread time.
I don't think you actually use UE5 at all because if you did, you'd know how trash the code is.... Even the touted 2x improvement in parallel translation is not great. Theyt still haven't dramatically improved draw calls at all. Again, if you want to use UE5 for a mass market project, you must be render to rewrite part of the renderer. It's truly that bad performance wise.
- Studios only switch because its cheaper and easier to find people
Here's what the cto of CDPR had to say about this topic:
we spent a lot of resources and energy to evolve and adapt REDengine with every subsequent game release. This cooperation is so exciting, because it will elevate development predictability and efficiency
In a separate interview, he said this:
it's "refreshing" to be coming to Unreal with what they've learned while developing 2077 without having to start the process of "rebuilding" their own engine for the next game.

"Likewise there's some things that REDengine does better than Unreal [that] we're working with Epic to basically bring to that engine as well. So it goes both ways… It's about economies of scale: you can obviously do all these amazing things in both.
Multiple developers making the switch have echoed this sentiment. It's simply about financials, support, availability of workers, etc. It's not really due to technical reasons. If most studios could afford to keep funding their engine development and research, they would gladly rather own their own tech.

I know you really tried hard to sound smart and it's a nice attempt. However, you're just talking out of your ass. If I had more time, I'd link you to other developer commentary on the matter. Perhaps try using UE5 and posting in the ue forums sometimes. You may actually educate yourself so that the next time this topic comes up, you don't expose your lack of knowledge.
 
Last edited:

ZehDon

Member
Epic Games, the 20 billion dollar game / engine company, doesn’t know how to create a game engine.

Not only that, but all the other game companies who choose Unreal to create their games on, don’t know what they’re doing either.

Stop this nonsense immediately
No one is questioning that Epic can create a game engine. Or even a good engine. The question is about how well optimised that engine's feature set is for the ubiquitous role Epic is selling it for.

Unreal Engine is marketed as a "plug and play" engine that can deliver, basically, everything out of the box. Movie productions, AAA or indie game developments, even solo game devs - Unreal Engine is marketed as a complete solution for everyone at any skill level. As we're seeing, that's not quite true, and developers are finding themselves in a position where they need to do bespoke engine extension for the "all-in-one" engine they licensed to get it run acceptably once they start using those fancy features. To this point, a lot of the developers who use Unreal simply do not have engine-level programmers capable of optimising Unreal engine for their unique scenarios. And that's not their fault. At this level of complexity, there likely aren't enough software engineers in the games industry capable of producing the necessary low-level re-overrides and extensions to Unreal for all the specific games to get the level of performance out of it that we should be seeing. And that makes perfect sense: if you had the that kind of talent on staff, you probably wouldn't be licensing Unreal to begin with.

Within the major version 5 of the engine, Epic have already delivered some pretty good optimisations. However, they package these with new features using large re-writes that are simply too difficult to implement late in a games development. As a result, we see games like STALKER 2 releasing on older versions of the engine and running accordingly. Epic is selling an engine that's difficult to use well for the AAA purposes it's marketed for, and it's licensed by devs who took Epic at their word and are now struggling to get it running without issues because it's beyond their technical abilities.

High end game development in 2024 is insanely complicated, and we're simply seeing the results of that. Given more iterations, I suspect UE will deliver a dramatically better OOBE. However, it won't arrive for years, and the games released on the older versions will likely never be properly fixed.
 
Last edited:

Katajx

Member
I think CDPR is making a huge mistake moving to Unreal. Vex is right.
Depends on the lens you are looking at this from. I still don’t think that Cyberpunk was the game they set out to make, nor was their engine probably a great fit to do everything they wanted.

On top of that they have been trying to grow and they now have offices in Poland, Canada, and the US. I imagine it is difficult to help people get the most of out their proprietary engine opposed to middleware like Unreal.

I know there are plenty of people that will come out and disagree, but Unreal has been incredibly versatile compared to engines like Luminous, Creation, and Frostbite.

That’s not saying there aren’t great proprietary engines, but they are usually made and tooled for a specific type of game and they do those really well.

Look at Naughty Dog, Insomniac, Guerilla with Decima, and the types of games those teams continue to make.

RE Engine has also been getting its uses outside of Resident Evil, and while it works it hasn’t been without difficulties or drawbacks similar to EA with Frostbite.
 
Last edited:

Felessan

Member
Within the major version 5 of the engine, Epic have already delivered some pretty good optimisations. However, they package these with new features using large re-writes that are simply too difficult to implement late in a games development. As a result, we see games like STALKER 2 releasing on older versions of the engine and running accordingly. Epic is selling an engine that's difficult to use well for the AAA purposes it's marketed for, and it's licensed by devs who took Epic at their word and are now struggling to get it running without issues because it's beyond their technical abilities.

High end game development in 2024 is insanely complicated, and we're simply seeing the results of that. Given more iterations, I suspect UE will deliver a dramatically better OOBE. However, it won't arrive for years, and the games released on the older versions will likely never be properly fixed.
It's a really naive look at things. Most people who invest money and time don't go with "trusting a word", they do a due diligence and cost of alternative analysis.
And biggest customers (those AAA games) have the most say in the direction where product go. It's just in almost all big off-the-shelf products customers priority are as follow - Bugs > Features > Refactoring (improving existing code).
Refactoring is a long arduous process that might give rather limited improvements but inevitably delays new features, so it's done regularly especially on most critical parts, but done in moderation as to not pull too many resources.
 
Last edited:

blue velvet

Member
It's really a weak excuse, cause GSC relocated itself and most of their staff to Prague 2 years ago. They didn't do their dev in war environment


Whole industry just calculated how much it'll cost in time and money to maintain their own engine and decide that it's easier to buy engine from a team that dedicate itself to it instead of adding 1-2-3 years and several tens of millions to project to try to do it themselves
As tech adds more and more features it becoming more and more expensive and time consuming to maintain in-house engine
And yet you have games like Plague Tales from a relatively small developer with their amazing in house tech. Unreal slop kills innovation. They should've calculated how much stutter fest their engine has caused to the average gamers.


8mLgu0R.jpeg
 

FireFly

Member
And yet you have games like Plague Tales from a relatively small developer with their amazing in house tech. Unreal slop kills innovation. They should've calculated how much stutter fest their engine has caused to the average gamers.
Realistically, if you don't have the technical resource to optimize UE5, you're not going to be building and maintaining your own engine over the same time period. Asobo has 250+ people so is basically Remedy sized and like Remedy seems to have a strong tech focus.
 

ZehDon

Member
It's a really naive look at things. Most people who invest money and time don't go with "trusting a word", they do a due diligence and cost of alternative analysis.
And biggest customers (those AAA games) have the most say in the direction where product go. It's just in almost all big off-the-shelf products customers priority are as follow - Bugs > Features > Refactoring (improving existing code).
Refactoring is a long arduous process that might give rather limited improvements but inevitably delays new features, so it's done regularly especially on most critical parts, but done in moderation as to not pull too many resources.
Not really. You're misattributing application software design development principles to performant software design. For most application software, the bold will be true - largely because application performance is secondary to application feature set, which creates a more competitive application capable of generating new revenue streams through those new features. For performant software, software like game engines like Unreal, that's not going to be true because performance is a feature. Things like "mega lights" aren't a "new feature" - the engine has supported a variety of lighting techniques for literally decades - however, "mega lights" is a technique that provides additional performance for existing features through new code. In this regard, capturing new business stems from allowing the engine to run better under more circumstances for more games on more hardware.

The due diligence in this scenario is a feasibility study of the engine feature set, quantification of the cost of development of equivalent feature set and the expenses involved with maintaining that proprietary tech stack, measured against the licencing costs of Unreal combined with in-house development of any necessary customisation. Given Unreal provides bleeding edge features, like virtualised geometry, as well as access to the engine underlying code so full customisation is available, it'll be hard for most companies to compete. Hence companies like CDPR dropping their in-house engines for Unreal. It'll be more cost effective for them to develop and maintain their own extensions to Unreal rather than maintain the entire engine, given the sheer breadth and complexity of modern game engines. As 343i recently said, it stops every development house also needing to be a tech company.

Simply put: if every developer could produce industry leading engines and save the licensing costs of Unreal, they absolutely would. But the technical complexity required to do that has grown such that it's virtually impossible for all but the biggest, most successful, and most prestigious developers. The same is true for digital distribution platforms, to save Steam's cut of digital sales, which is why EA, Ubisoft, Activision, Microsoft, and Rockstar have all tried their hands at their own PC platforms. One by one, they've gone back to Steam because it's hard to beat someone who's focused on exactly one thing. EA doubled down on Frostbite, Ubisoft uses its AC and Snowdrop engines, Sony created its ICE Team years ago, while Microsoft has basically gone fully Unreal. We're seeing tech consolidation in the industry because, frankly, you all but need a fully funded and staffed dedicated software company doing the R&D and development to deliver AAA engines at this point. How many developers can afford that, on top of developing a AAA game?
 

Felessan

Member
And yet you have games like Plague Tales from a relatively small developer with their amazing in house tech. Unreal slop kills innovation. They should've calculated how much stutter fest their engine has caused to the average gamers.
It's not average players who pays them.
Refactoring is already a 3rd priority, and as it's mostly PC issue - it's priority is even further down the priority list.
There are always a lot of things to work on and priority usually chosen by a) impact (return on efforts invested), b) scale (how many people affected), c) weight of issue and it's tolerance

If stutter is annoying but still playable, affect only limited amount of players (and not the most paying ones) and requires significant efforts to fix it in-engine - it will be put on hold almost permanently and the problem will be left to particular devs to organize proper asset flow themselves.

And no, unreal does not stop innovation. You just might not see it. I would rather see good tools for simple and seamless ECS implementation and optimization, because all those "games are CPU bound" is a bullshit coming from arhaic design of CPU parallelization in game engines.
 

hlm666

Member
CDPR adopting Unreal Engine 5 will be the best thing to happen to the engine. They're already solving a lot of issues with the engine regarding loading and asset streaming.. and merging the changes into the default engine.

It sucks that CDPR are dropping a clearly awesome engine, but I believe that them doing so will make UE better for many other studios and games in the future.
Hadn't heard about that, any idea if they are going to be using the nvidia branch so they can go all in on ray tracing again? That makes me think it's just going to be the base engine though without any bells and whistles. Unless nvidia manage to push their stuff into the main branch.
 

Felessan

Member
Not really. You're misattributing application software design development principles to performant software design. For most application software, the bold will be true - largely because application performance is secondary to application feature set, which creates a more competitive application capable of generating new revenue streams through those new features. For performant software, software like game engines like Unreal, that's not going to be true because performance is a feature. Things like "mega lights" aren't a "new feature" - the engine has supported a variety of lighting techniques for literally decades - however, "mega lights" is a technique that provides additional performance for existing features through new code. In this regard, capturing new business stems from allowing the engine to run better under more circumstances for more games on more hardware.
There is no difference between application software and performance software unless we talking about fixed-feature performance software and game engine clearly not that.
Performance is an issue in almost any software - if a query runs for 2 hours it's as bad as freezes in games. And it's always a choice - whether optimize or add a new feature. Sometimes 2 hours is just unacceptable, sometime it's better to leave time at 10 mins and concentrating on process optimization to do 5 steps in one, instead of running 5 queries each for 3 mins. And sometimes you do things that eliminate need for an old and non-performing query at all
 

draliko

Member
Game development is killing games, maintaining an engine cost a shitton of money, especially developing good toolsets... An engine is way more than simply a renderer...
We always had Middleware to help not reinvent the wheel every time and modern engines are the natural evolution of that as most Devs don't have the resources to invest in proprietary engines especially considering the time it takes to make a single game.
We had a lot of proprietary engines back then because the complexity was way lower... And every single engine was made for a single game or a single genre, now only biggest studios have the money to spend in general game engines, the others pay royalty to use a premade one. Also finding someone that already knows the engine instead of teaching your proprietary toolset to every new employee is way easier and efficient seeing the people turnaround in studios. Add the fact that good programmers get payed more in other fields... And here's the perfect storm.
I'm waiting for ai to really be able to improve codebases, on the other hand this could be the death of low level programmers...
 

ZehDon

Member
There is no difference between application software and performance software unless we talking about fixed-feature performance software and game engine clearly not that.
Performance is an issue in almost any software - if a query runs for 2 hours it's as bad as freezes in games. And it's always a choice - whether optimize or add a new feature. Sometimes 2 hours is just unacceptable, sometime it's better to leave time at 10 mins and concentrating on process optimization to do 5 steps in one, instead of running 5 queries each for 3 mins. And sometimes you do things that eliminate need for an old and non-performing query at all
To be honest, you seem to be disagreeing for the sake of it; I clearly wasn't talking about long-running SQL Server queries or nHibernate hydration timings. The simply fact is: the time and cost investments of modern AAA engines has reached a point where it's not feasible for most. This has consequences, such as development houses employing less low-level technically proficient software developers. This isn't a controversial statement. The history of engine licencing is littered with countless of examples of technically ambitious developers unable to wield cutting edge tech.
 
Last edited:

Felessan

Member
To be honest, you seem to be disagreeing for the sake of it; I clearly wasn't talking about long-running SQL Server queries or nHibernate hydration timings. The simply fact is: the time and cost investments of modern AAA engines has reached a point where it's not feasible for most. This has consequences, such as development houses employing less low-level technically proficient software developers. This isn't a controversial statement. The history of engine licencing is littered with countless of examples of technically ambitious developers unable to wield cutting edge tech.
I agree with you on this.
On the side note - one of the reason why even big houses move to UE is that developing engine requires time and things already bloated as they are to further stretch time-to-market
There are a real risk now that game will become obsolete during long development time because market preferences shifted way to much during this period
 

Herr Edgy

Member
Performance aside, I hate how games made in Unreal almost always feel lifeless compared to comparable games built in other engines. Obsidian's games are a huge example of that, as they went from Fallout: New Vegas, where countless small objects had physics and every corpse could be dragged, to The Outer Worlds, where ammo packs would float if you grabbed an item it was on top of, and corpses can't even respond to explosives. Avowed's gameplay trailers look equally lifeless. I wish Bethesda would let them use their engine.
You can enable this sort of functionality with some checkboxes.
Just comes down to what the dev wants to prioritize, has the skill to do, or perfmem concerns.

Reality of the situation is that game development is hard. UE drops the bar for entry considerably, but the ceiling never comes down along with it.
Since the bar for entry is low, you will have a lot of people try their hands; some of them only semi-competent or lacking in manpower. You won't be able to compete with a AAA company that has both manpower and very competent people. Shocking, I know.
 

SHAKEZ

Member
I believe that were at a point where pushing graphics os meaningless since devs can't optimize for shit, so people can't play those games anyway...
Unreal and unity should push for better optimization tools instead of more graphical impressive tools, since those never get utilized to their full potential anyway.
 

ScHlAuChi

Member
Here's what Epic themselves had to say about this topic in the announcement of 5.5:
So what? Them continuing to improve the engine does prove that their code is trash?
Can you actually point to concrete examples of that "trash code"?

I don't think you actually use UE5 at all because if you did, you'd know how trash the code is.... Even the touted 2x improvement in parallel translation is not great. Theyt still haven't dramatically improved draw calls at all. Again, if you want to use UE5 for a mass market project, you must be render to rewrite part of the renderer. It's truly that bad performance wise.
I´m a designer, not a coder, but I trust the one I work with, who has shipped multiple games.
Can you name any of the UE mass market projects you worked on where you replaced the renderer and demonstrated it to have much better performance than that of Epic?

Here's what the cto of CDPR had to say about this topic:
In a separate interview, he said this:
Multiple developers making the switch have echoed this sentiment. It's simply about financials, support, availability of workers, etc. It's not really due to technical reasons. If most studios could afford to keep funding their engine development and research, they would gladly rather own their own tech.
Yes, of course those are reasons, but the biggest reason is the speed-up in production due to superior tools and workflows!
That is a technical reason - tech does not just mean that your engine can do pretty pictures!

I know you really tried hard to sound smart and it's a nice attempt. However, you're just talking out of your ass. If I had more time, I'd link you to other developer commentary on the matter. Perhaps try using UE5 and posting in the ue forums sometimes. You may actually educate yourself so that the next time this topic comes up, you don't expose your lack of knowledge.
You know, people who work on projects dont have the time to post on UE forums, they are usually too busy getting a game done!
But as you claim I´m talking out of my ass, please educate us what games you have shipped in what capacity to have such superior knowledge :)
 

Men_in_Boxes

Snake Oil Salesman
No one is questioning that Epic can create a game engine. Or even a good engine. The question is about how well optimised that engine's feature set is for the ubiquitous role Epic is selling it for.

Unreal Engine is marketed as a "plug and play" engine that can deliver, basically, everything out of the box. Movie productions, AAA or indie game developments, even solo game devs - Unreal Engine is marketed as a complete solution for everyone at any skill level. As we're seeing, that's not quite true, and developers are finding themselves in a position where they need to do bespoke engine extension for the "all-in-one" engine they licensed to get it run acceptably once they start using those fancy features. To this point, a lot of the developers who use Unreal simply do not have engine-level programmers capable of optimising Unreal engine for their unique scenarios. And that's not their fault. At this level of complexity, there likely aren't enough software engineers in the games industry capable of producing the necessary low-level re-overrides and extensions to Unreal for all the specific games to get the level of performance out of it that we should be seeing. And that makes perfect sense: if you had the that kind of talent on staff, you probably wouldn't be licensing Unreal to begin with.

Within the major version 5 of the engine, Epic have already delivered some pretty good optimisations. However, they package these with new features using large re-writes that are simply too difficult to implement late in a games development. As a result, we see games like STALKER 2 releasing on older versions of the engine and running accordingly. Epic is selling an engine that's difficult to use well for the AAA purposes it's marketed for, and it's licensed by devs who took Epic at their word and are now struggling to get it running without issues because it's beyond their technical abilities.

High end game development in 2024 is insanely complicated, and we're simply seeing the results of that. Given more iterations, I suspect UE will deliver a dramatically better OOBE. However, it won't arrive for years, and the games released on the older versions will likely never be properly fixed.
The question then becomes… Why do you know this, but the giant successful game companies don’t?
 
Where did this newborn demand for 60 FPS started?
For me, it was using an OLED TV. Due to OLED panels having essentially zero pixel blur, low framerate games often look extremely choppy. It requires games with well-implemented motion blur and such to counter this.

That said, I still choose quality 30fps in games that handle it well (Final Fantasy XVI, Control).

I think that 40fps in a 120hz container is the best option overall.
 
Top Bottom