because it's a shit engine, most likely. Think horses, not zebras. If making a usable, extensible engine with modern bells and whistles was easy, unity and unreal would have way more legitimate competition and godot wouldn't permanently struggle for third place feature-wise
@@sacwingedbatsatadbitsad4346 No it's not. It's the only engine capable of running path tracing in real time at 60fps, something that Unreal Engine 5 won't hope to do anytime soon in any real-world examples.
I don't by into their bold statement that their engine is better than unreal. I think it's probably one of their tech guys who was trying to explain to exec that their engine is better for what they do, and the exec ran with it like our shit is better overall :)). In reality if their engine was better, they would pivot immediately into releasing it rather than making a game like cyberpunk with it.
UE Dev here - the issues you mention mainly stems from developers unfamiliar with the engine, or devs taking shortcuts. This means they use generic functionality instead of optimized tools, that requires engineers to tweak. Also, the engine's source code is open to anyone. You don't need Epic to fix bugs, anyone can do modifications.
Optimized tools? Like not using UE5? Why bother fixing UE5 code? You have to both know how to make it work from scratch and which parts need fixing. That's just plainly more work, especially since the engine has extreme feature bloat by trying to be an all-in-1 toolkit. Sounds like you're as much of a "dev" as all the whiners on twitter are "artists".
We’re seeing a hideous trend where these game developers and companies are using UE5 as a bandaid fix for issues that a new engine can’t fix. UE5 isn’t going to fix poor writing, art direction, narrative, gameplay, etc. and companies leap onto the UE5 fix when gamers are mad at their games thinking that showing their IP’s within a UE5 benchmark test is somehow going to fix deep seeded internal issues that have nothing to do with the engine. The only thing this mass migration to UE5 is going to do is continue to homogenize the games and companies within the industry making them even more bland and unappealing, thus making it easier for larger companies like Microsoft to buy the studios up and shut them down.
Right now, I get tired of seeing the gradual noise transparency effect, which is signature of the UE. But seriously, what I've heard, Unreal makes stylization remarkably difficult (albeit still possible, of course). Especially new developers are much likely to take the path of least effort and leave stuff at default settings, leading to indie games looking samey like they were made in Roblox using the basic internal stock models. Later in time, if indie devs start to migrate to other engines, Unreal will end up as solely AAA game engine with only large studios' eployees being knowledgeable enough to use it in production.
You realize what that means.. right? It means that developers are using it wrong, cutting corners or not bothering to optimise enough. That doesn't mean there's an underlying issue with the engine.
As others stated, these issues aren't specific to the engine, but it's a) indie or AAA devs not allocating resources towards optimization of the game, b) they're not familiar with all the different ways to optimize the engine (a need to upskill) and c) the engine versions themselves have been rapidly evolving, so we're not seeing all the latest features and built-in optimizations in a lot of these new releases. Those wont be coming around till the next couple of game release waves. A lot of the problems mentioned also apply to other game engines out there, and the issue with proprietary engines are limited documentation, knowledge, and in the case of EA's Frostbite engine, these engines are typically not built for certain game genres, so they have to take years completely retooling them. I don't see the performance hits being much different than a lot of AAA games out there.
That is all a problem, though. Most AAA games run poorly and look bad; that is not an excuse to accept more poor running and badly optimised games. Also, the threading issues have been present since UE3, though, it's kind of a given since it's a general purpose engine. The problem is that no studio is going to take years refactoring an engine they don't own, to optimise it for a game that may or may not sell well. So they will never allocate resources to optimising or refactoring Epic's engine; and Epic is not going to refactor their own engine for someone else's game. So it's a catch 22 with no resolution. We are basically stuck with the current cycle of poorly optimised games on two year dev cycles that play horribly on top-end hardware.
@@Billy-bc8pk Other AAA studios don't have much of an excuse, but Epic is constantly improving and optimizing the engine. UE already supports multithreading to an extent, the threading issue is also improved in the latest versions and there are ways to mitigate it. The problem is that there are a million ways to do different things, for different game types. They actually do work with other major studios (including CDPR now), and those lessons in return are added back into the engine. Open world for instance is an area they're focusing more on. It's all a case by case scenario. The biggest case study they have is Fortnite. Are there FPS performance issues there? It's fricken running on mobile. Again, studios need to actually apply the lessons that UE puts out in its documentation, Unreal talks, and forums.
Most people learn from other people that also don't really know what there doing. 99% of tutorials for unreal will tell you to cast to any blueprint at all that you need info from. That alone is a massive performance hit. Not to mention the amount of things people run on ticks. Most people have no idea that spawning something then destroying it does not completely remove it. Thus, things build up very quickly. Literally even a bullet in a fps game. If you just spawn a bullet then do delete it or do nothing to it, it will still add to the objects in the world and just stack up until your on 10 fps
Monopoly is always bad for the market. No matter how you look it, from which perspective you look it, even if it looks good for the industry, it's end result is always bad.
Yes it's a monopoly, and so many people refuse to accept this and waste thousands of dev hours and funding on Godot or Unity modifications instead of investing in a 3rd party fork of UE that fills the gap of what's missing in the UE dev space(fast optimization workflows). 5:12 --This is not theoretical, we already proved it's NOT effortless and we have more proof on the way that destroys the "fixed cost of nanite" lies going around the dev space. 7:27 --The take here is a little too simplified. The game is mostly static, there is no reason to have a dynamic GI solution. And while we don't advocate for lightmaps, we need systems that are precompute/stream local lighting information and allow moderate dynamic interaction which UE doesn't really provide. Something like a DDGI+Distance Field Indirect Shadow method would be better. Lumen reflections where not the problem in SH2R and anyone who watches the analysis on it we did will know that. 6:11 --More like UE designs these demanding features to only work properly when combined. And again, reflections are not really that much of a problem. That example you used for instance is mostly traced from screenspace (that's been done since at least 8th gen consoles) and then fallback on HW accelerated reflections. Sad to see you not mention that half the pipeline in UE5 requires Temporal smear from algo's like DLSS or TAA but I guess that's something only we cover. UE5 does not help developers or gamers, they only help the highest level of employers. Many gamers and developers are sick of these issues but like you and I are saying: It's already a monopoly. So instead, lets try fixing the problem that's been here all along(UE via fork, plugins, usf/source patches etc).
Thank you for commenting! Unfortunately I am not a developer so a lot of this video was based on research, so I am sorry for using more simple examples or missing topics like Temporal smear. Regardless comments like this are very helpful for me and others to better understand certain aspects of UE5 and appreciate you typing all this info out.
@@major_trenton You're all good, just wanted to clear some things so the wrong things aren't blamed but don't worry about missing these topics since we are pretty much the only people that cover this. But I'm glad to see other youtuber's define this as a monopoly!
@@arisumego I suggest caching up with our views. We are very specific about using UE5 and will not be supporting ANY other engines. Too many resources have been wasted on such unrealistic goals.
As much as I adore it, I think a good example from this is a little demo called Stonemachina. It's an indie sauce like where you play as little Stone chess pieces and you fight Angels based on Italian mythology. It also has DMC style switching and it's absolutely phenomenal. HOWEVER, it is also in unreal engine 5, with photorealistic textures and sky lighting, meaning it _chugs_ sometimes. Not to mention it's a demo so it's not fully optimized from the get-go. So this fantastic game with fantastic gameplay is being brought down by graphical choices and how demanding the engine is. I'm having dips on a ryzen 7700x for an _indie game._ Now don't let it deter you from trying out the demo it's actually fantastic. There are graphical options that won't melt your computer but my point still stands
There's a lot of wrong in this. I'm a UE dev with a game released with all of Unreal's features, so here is my view on the subject: Primarily, the engine is not the issue, aside from stuttering (which is slowly getting fixed probably thanks to the help of CDPR) the issue comes from the developers misinterpreting Epic's statements on their technologies. The most recent GDC and Unreal Fest talks clarify a lot better what the technologies are, how to use them properly and what are the limits. AAA games today are released half-baked and poorly optimized leading to bad promo for Unreal Engine while other lesser known studios are releasing much better games with the same engine. So the problem is not the tool but the user. Stuttering issues are real but can be mostly avoided by implementing some sort of shader pre-compilation on game launch, this doesn't avoid everything but it helps. Using demanding features such as Nanite and Lumen is a developer choice, and there are no other techniques that cost less for "Performance Mode" as it would basically be reworking the entire game for just a "Performance Mode", these modes don't promise 60fps, just "performance". For example using Lumen, you can't use anything but Lumen since you'd need to relight your entire game for it to not look horrible when turning it off. Nanite has a base cost and in my experience doesn't cost a lot more when using a lot of Nanite asset, it is simply a better LOD solution in my opinion and I use it heavily, including the Experimental Nanite Tesselation in a packaged project. For the upcoming Mega-Lights, initial testing helps getting a lot more performance out of lights, which are a huge bottleneck in my project and probably many other projects, the feature also fits in perfectly with all of the engine's features including Nanite and Nanite Tesselation with no additional cost. For the monopoly concern, there are other engines competing, but they don't have the same promotional power that Epic has, there's obviously Unity with Unity 6 and soon 7 that will probably be better than the previous Unity versions and might just reach a level closer to UE and there are a lot of games especially indie ones made in Unity. There's also the Flax Engine that is improving over time, Godot for simpler games, and probably others that I don't know. Even if Epic had a full monopoly, they would probably still develop the engine, because I feel like Epic is more passionate about their product than some other engines, which is a great thing and is very promising for the future. To conclude, yes, Unreal Engine is all about promises, promises that they actually hold, they continuously improve, fix and add new features to the engine which makes it better every year, and I'm sure they'll figure out the rest of the issues in due time, but for now AAA studios have to stop being lazy and actually make good games with the engine instead of producing bad cash grab after bad cash grab
Dead island 2 is a prime example of how to use UE5. Regardless of how good(not good) the story is , that game is visually stunning asf and has some of the best graphics the ps5 has seen. And it runs beautifully.
Thank you for this video. It's good to see more people waking up to these issues. The stuttering is caused by a lack of optimisation for multithreaded processes. Assigning cycles to processes on sub-threads could alleviate this problem, but the main issue is that with a general purpose engine, Epic can't know WHERE to assign parallel processes since everyone's game is going to be different and require different computational modals. In order to properly resolve this it requires a complete backend refactor of how the multithreaded processes are subdivded via their task priorities. There is no easy way around this, though; most of the logic and upfront computational thoroughput is handled on the main thread, and so there is constant stop-and-go stutter due processing issues. Attempting to resolve this could end up being years worth of refactoring. And it could end up having knock-on effects on other recently implemented systems, such as the new mega-lights, as well as existing systems such as Nanite and Lumen. And rewriting all of those could have serious impacts on in-development projects relying on Epic to provide support for their UE5 projects. This puts Epic in a very precarious position: do you rewrite the backend multithreaded handlers to improve thread performance and hopefully reduce micro-stuttering, causing potential problems with existing plugins/systems that may also require refactors? Or do you keep pumping out new features and let the studios handle the refactors on their own dime/time? In the interim, Epic doesn't lose much because everyone is switching over to using the UE5 since most top companies are losing their engineering talent who maintained their proprietary engines. Don't be surprised if the Battlefield after the next one also switches to UE5, since many of EA's Frostbite engineers left to form and join Embark Studios. In the long run, this could end up being severely damaging, as people will soon recognise that UE5 games have serious performance ceiling issues, and may not be worth it, resulting in people becoming less enthused to play games running on UE5. That could hurt Epic big time, as people may end up associating the engine with the negative stigma that games release in poor functioning states with frame stutter and playability issues, which may have severe long-term effects on sales.
It's an amazing engine which is why everybody's using it, and that's why in just about any industry a product that stands out from the rest becomes the standard. That said, if Epic Games stagnates due to their success, then there's only one way to go from there... and that's down, which is the the pattern these powerful, influential corporations tend to go through. But in the meantime, I don't see the problem in enjoying a fairly solid product that can put out some high-end, professional grade visuals for both game and film industries. If their throne is overtaken by another engine in the future, that means it's got to be faster, easier to learn, and all around better if it can dethrone UE... so I don't see a problem with the way things are going with unreal
UE5 is better for movies and commerical rendering than it is for game development. It might be the worst thing that happened to gaming since loot boxes.
@@Native_Creation No it is built for general purpose creation; Godot and Unigine are built more-so directly for games, along with Unity, which is why all three have better running games despite not having as many features as Unreal.
@@Native_Creation UE5 has fundamental issues that even Epic themselves couldn't fix for their own games. And Nanite is more often than not performing worse than traditional LOD systems. It's great for movies and commercials but flawed for gaming.
@@mariano4068 That is a meaningless platitude when *on practice,* the real world result is to have UE games all have this uncanny similarity between them in lighting and texture work.
@@gorofujita5767You'd be surprised how many games are developed in UE. Just to name a few: Abzû, No More Heroes III, Black Myth: Wukong, Hi-Fi Rush, It Takes Two, Code Vein, Sifu and Pumpkin Jack. Now would you say all these games look the same? I won't deny that UE has a somewhat "default" look to it, but talented developers can make it look really unique. It's an incredibly versatile engine.
9:93 in the scene I tested Megalights in which had Nanite and Lumen enabled with hardware ray tracing, it doubled my FPS from 50 to 100 on my 4090 in 1440p using the cinematic preset from the editor
Nobody cared about crysis having 3 fps, all just got shooked by what the game was able to do. Now how many years later? unreal is doing stuff that we haven't seen before just like back then with crysis and ppl are just complaining even tho we haven't even seen a game yet so it is all just speculations... I hate humanity...
Every engine is just a tool with it's own pros and cons. You choose an engine to save the time building one yourself. I have faced issues with optimization with UE, but I don't think it's the engine's fault. There's no real monopoly on the market, IMO. However, unreal could improve on it's learning material so that devs understand how to optimize their games more. At the end of the day, if you rely 100% on visual scripting, you're going to run into issues later on.
The main stutter issues people seem to complain about is shader-compilation and traversal stutter. But visual scripting could be in the mix, warping the impression for the worse, who knows.
@@SnakeEngine I'm relatively new to game development, but I know from the web dev world, visual page builders just came with so much optimization issues that needed external 3rd party tools to clean up. Game engines have so many more factors that come into play, I think it's just a matter of developing an optimized workflow.
Hard to talk about monopoly with UE5 since the engines the games use are generally proprietary engines, so no other studies use them, and there is no impact on the general market
I think that's their policy, presenting features so the new guy can go: omg, wow How does it work? 🤔 We'll improve it in the future nvm... hmmm guys I have idea, let's hire ALS guy for the next update
I'm very interested in the way graphical fidelity seems to have hit a point of diminishing returns while performance and optimization continues to get worse. The overhead on modern engines just doesn't make sense to me. Time was you could do driver-level supersampling to get 4x as much resolution and still hit the 60FPS mark, and meanwhile games today need to upscale to run halfway decent. Yeah, raw graphics look better compared to games 5 or 10 years ago, but how much better really? I'd rather have something like Skyrim running at native resolution (or higher) without terrible TAA that makes everything a blurry, ghosting mess. These days if a modern game even deigns to include MSAA it will tank the performance significantly, because the headroom is so small to begin with.
Game devs back in the days had to come up with tricks to fake fidelity. Now we are at a point where finding tricks takes to long and the old tricks get pushed beyond their boundries. I am also really missing the effort old games put into gameplay details. Just watched a comparison between dead rising and it remake. For some reason the original looks better and has more gimmicks. The new Doom games where amazing with the new gore effects. Not the first to do it but they did a great job. New dead island is also AMAZING in terms of just killing zombies. We need more games that focus on fun and details and less on ultra realistic graphics
And even though there have almost monopoly of game engines, they are barely profitable. This area is cocked, a competitor, at the same level, will not be able to sustain itself without a Fortnite. It also proves it's a bad idea to maintain a proprietary engine, you're multiplying the risk of bankruptcy.
A lot of these issues aren't unique to Unreal. In fact, if developers take the time to properly optimize their games UE5 can look and run extremely well.
I don't like the way they push unreal to movie and films in same package of the game engine. Game engine should compact and have good performance. Bloating game engine is never be good. I wish they make unreal lite version where only barebone game engine who's running fast and less performance issue
monopolizing? lol no quoting a few games made in an engine means nothing compared with over 1000 games created each month. proceeds to handpick some games that have some issues, ignoring the vast majority of other games that don't have that particular issue, then generalizing, concluding, wrongly, the engine itself has that issue, without offering any proof of that. how about that particular game was coded poorly? No? Then why other games don't have that issue? Crickets.
Unreal engine 5 is a go to for established lazy "AAA" devs for two reasons - fancy graphics are the main (and usually only) priority in modern games. And secondly they don't need to bother with optimisation at all, cuz nowadays they can tell people: "Stop being poor" and make them buy expensive new CPU and GPU's. This ofcourse is very convenient for them, and their investors from the monopolies of the PC components cartel - force you to scrap your old PC.
Add in the marketing bit. All the time I catch game journalists just having the urge to point out that a certain game "will be running on Unreal Engine 5", especially for game announcements.
It is when the games run poorly, don't look great, and the gameplay suffers for it. As showcased with Until Dawn, the last gen version looks and runs better.
UE3 which was all the OG mass effect didnt even allow any AA besides FXAA which is the worst. Injecting smaa also difn't work. So I was never a fan og untral gor those reasons.
I dont think UE5 is user friendly at all. I only went for UE 5 because I really like C++. If I knew of another engine I could use C++ with that was battle tested, I would walk away from UE5 and not look back.
@@brk29 Yeah, but he probably wants the performance advantage of C++ and dislikes dealing with the garbage collector. Yes, you are not dealing with garbage anymore in C#, but still with garbage collection since you want to avoid it in games.
@ then, he can use Unigine, Flax, or something similar, they are not battle tested in terms of published games. But, he can battle test them. There is also cryengine but I wouldn’t say user friendly to that either. Oh and O3DE. Not battle tested yet but seems promising. I still think C# is currently not that behind of C++ in terms of performance. If you are not after something really specific you can pretty much do the same stuff in C#.
We need mor UE3/4 games and using other engines that are capable of cool and good things, pointing or having as a goal having "MAX" graphics isn't the future of gaming, it's a part of it, and executives just don't understand and ask for "the best"
Some good points, but mostly you're talking garbage. The issues you attribute to UE are misplaced, as the devs don't know how to optimise the game, and take shortcuts. The UE is heavily customisable, and its source is accessible and can be built with custom changes. Your statement: Lumen and Ninite are a problem when used together, again is a dumb statement, as it's all about how they are used and not the sole fact of being used together. I wish people stopped making videos on subject they don't fully understand. You make bold statements against UE5, but there were 15000 games released in 2024 according to steam db alone, so pointing out 5 notable released in UE5 is just not good enough - many sources list only a minor percentage of them as UE5.
news flash perfect games don't exist perfect game engines also don't exist perfect consults or PCs don't exist nobody lives like mr. beast making millions by faking vids and having scams
RUclips gamer logic. valve can push EDI in his games and company can ask for tones of money from game devs and gamers can delate accounts and other stuff and his custom game engine who is in private mode what are most gamers respond aw we just ignore it when epic allows for everyone to use their engine for free + the source code it's also available + free games from the store every month and free assets and now FAP who is all in one store and asset place who gives 80% on their sells so how did the modern gamer respond to it aw i hate epic because there are a greedy Corop boo Hoo how tone deaf are you all.
@@user-be5kj1bw3d also go read my other comments the once where i am talking about other game engines and completion to epic like aw i don't know valve and unity and you say its a shitty game engine yet you didn't saw any issues when wukong made over 10k this year aw how convenient. you guys are just as much as grifters as sweet baby ink and EDI care about repetition for the modern audience.
@@user-be5kj1bw3d " hates monopoly's" forces everyone to go to valve greedy ass because every game dev has 100 dollars to spend + the 30% tax cut from valve. and valves own over 30% of the gaming publisher you guys are so tone deaf it blows my mind
Honestly if Epic Games make their game store as better and optimized as steam and also gives you an option for offline installer like GOG launcher and optimized Unreal Engine with a plugin to downgrade assets or projects + fix game stutters + their own c++ editor without requiring visual studio, they would really succeed and loved by gamers. I really like Epic but they barely listen to users reviews or criticisms. Hopefully with CDPR, they manage to get their crap together, they would fix a lot of stuff like Kojima Productions did with Decima Engine. Honestly their UI designers suck big time and that is a fact.
you do realize that there plenty of other game engines out there right some of them are built even from the source codes of cry engine and unreal need proof well did you look up wicked engine yet how about upbge how about you guys promote some other engines then freaking unity if you want unreal to have " competition" aw what happened to cry engine cry engine was justness as good as unreal and the cry games kick ass on realism even back in 2009
WHEN WILL YOU MAKE THIS VIDEO????? Why DONT YOU MAKE A VIDEO titleD: BAD ACTOR YOUTJUBERS/CONTENT CREATORS ARE RUINING GAMING BCS OF THEIR FALSE NARATIVE CREATIONS OF 0 KNOWLEDGE ABOUT TOPIC REPORTING,COPY PASTING MISSINFORMED,MISREPRESENTED TAKES FROM OTHER SO CALLED KNOWLEGABLE GAMER CONTENT CREATORS???? WHEN WILL YOU MAKE THIS VIDEO?????
aw boo hoo no completion against epic yeah, let's just ignore game engines like unity 6 , flax engine , falcon , upbge , cry engine , wicked engine and so on no , no push SLOB CONSUME SLOB WAIT FOR NEXT SLOB freaking source engine 2 epic not having competition my ass you guys are just lazy and grifters
Nanite isnt demanding tho. It increases performance. That all its there for. Poor Performance isnt entirely on the engine. The game devs need to optimize their games
Actually nanite can tank performance a lot. It's mainly there to make your life easier and not go through the painful LOD creation process. But it comes with a big base performance cost, and this cost doesn't grow much if you add a lot of objects to process. While traditional LODs have a per-object cost, meaning the more objects you have, the more it's gonna cost on your performance. So if you do a Nanite vs LODs with 10 objects, Nanite is gonna loose. If you do the same test with 10k objects, Nanite wins. So it's a question of compromise as always. The problem is, as you stated, that devs generally see this kind of tools as easy ways to get good results and optimization isn't properly dealt with.
RE: Nanite isnt demanding tho. It does NOT increase performance unless developers didn't optimized their there. You're comment makes ZERO sense. It's literally 100%+ slower than optimized topology and we proved it several times on the UE forums and in our Nanite Video. Nanite also demands an incredibly slow shadow system. Stop spreading garbage information. UE5 lacks workflows to optimize models efficiently with a high enough quality that 9th gen should be showing.
@@ThreatInteractiveIt does, when used in the right circumstances. You can’t just use nanite in any projects for any reason and wish to gain more performances. It’s mainly for films and areas of high fidelity with multiple geometry that needs to be rendered. People don’t understand that since they just took epics word for word of « optimization » without questioning themselves where Nanite is actually useful.
@@ThreatInteractiveAnd for the slow shadow system? Yes and no, it’s just « beta ». Which will most likely be fixed in ue5.5 from what I’ve tested. Better fidelity on VSM, good performance. So yes optimization is a trade off for what you need. But some things are mostly unachievable without those features that are yet being developed in the « BETA » Unreal Engine 5. Which is the beta for the next engine… One day hopefully, they just need time.
@@CodeWithTristan "Beta" is not an excuse. Epic's definition of "production ready" putting that system in fortnite regardless of it's poor quality. The point of geometric detail to "gain" with Nanite hurts performance so much, developers coat the image in Vaseline upscalers making all the detail a joke. And VSM's look like crap thanks to relying on temporal smear for uneeded denoisment. RE: So yes optimization is a trade off for what you need- Optimization is getting incredible results with smart/faster systems.
I feel like you're throwing the baby out with the bath water. UE5 is a great engine but alot of these issues stem from inexperience with C++ (a notoriously hard language), unfamiliar with the engine and western work culture. Firing your entire staff after a game releases so you don't have to give them raises and then hiring inexperienced fresh graduates for cheap to replace them, and then having them fall feet first into code they didn't write isn't good, not matter what engine you use. We see company's that foster talent like From software being able to make high quality games very quickly.
I don't understand why CDPR doesn't license out the red engine. It would fix the teaching issue and pay for its own upkeep.
Because the OG devs who made it, left after Blood and wine in 2016, and made their own studio Rebel Wolves.
because it's a shit engine, most likely. Think horses, not zebras. If making a usable, extensible engine with modern bells and whistles was easy, unity and unreal would have way more legitimate competition and godot wouldn't permanently struggle for third place feature-wise
Because it was a patch work of a gazillion fixes specific to CP. It was a miracle they actually fixed it.
@@sacwingedbatsatadbitsad4346 No it's not. It's the only engine capable of running path tracing in real time at 60fps, something that Unreal Engine 5 won't hope to do anytime soon in any real-world examples.
I don't by into their bold statement that their engine is better than unreal. I think it's probably one of their tech guys who was trying to explain to exec that their engine is better for what they do, and the exec ran with it like our shit is better overall :)). In reality if their engine was better, they would pivot immediately into releasing it rather than making a game like cyberpunk with it.
UE Dev here - the issues you mention mainly stems from developers unfamiliar with the engine, or devs taking shortcuts. This means they use generic functionality instead of optimized tools, that requires engineers to tweak. Also, the engine's source code is open to anyone. You don't need Epic to fix bugs, anyone can do modifications.
This could be knowledge they had prior to them going Open Source(Partially) - in UE4 and UE3 dayts, they were completely closed source.
Optimized tools? Like not using UE5?
Why bother fixing UE5 code? You have to both know how to make it work from scratch and which parts need fixing. That's just plainly more work, especially since the engine has extreme feature bloat by trying to be an all-in-1 toolkit.
Sounds like you're as much of a "dev" as all the whiners on twitter are "artists".
@@AdvancedScummerReally? I for some reason remember UE4 having it's source code available.
What about stuttering in Fornite?
@@crackny4n Well technically it was "Source Available" But it definitely wasnt open source by any means.
We’re seeing a hideous trend where these game developers and companies are using UE5 as a bandaid fix for issues that a new engine can’t fix. UE5 isn’t going to fix poor writing, art direction, narrative, gameplay, etc. and companies leap onto the UE5 fix when gamers are mad at their games thinking that showing their IP’s within a UE5 benchmark test is somehow going to fix deep seeded internal issues that have nothing to do with the engine.
The only thing this mass migration to UE5 is going to do is continue to homogenize the games and companies within the industry making them even more bland and unappealing, thus making it easier for larger companies like Microsoft to buy the studios up and shut them down.
Right now, I get tired of seeing the gradual noise transparency effect, which is signature of the UE.
But seriously, what I've heard, Unreal makes stylization remarkably difficult (albeit still possible, of course). Especially new developers are much likely to take the path of least effort and leave stuff at default settings, leading to indie games looking samey like they were made in Roblox using the basic internal stock models. Later in time, if indie devs start to migrate to other engines, Unreal will end up as solely AAA game engine with only large studios' eployees being knowledgeable enough to use it in production.
It never looks as good as when Epic themselves use it. This is a huge mistake for Halo, and will enable them to chew through even more temp workers.
You realize what that means.. right? It means that developers are using it wrong, cutting corners or not bothering to optimise enough. That doesn't mean there's an underlying issue with the engine.
As others stated, these issues aren't specific to the engine, but it's a) indie or AAA devs not allocating resources towards optimization of the game, b) they're not familiar with all the different ways to optimize the engine (a need to upskill) and c) the engine versions themselves have been rapidly evolving, so we're not seeing all the latest features and built-in optimizations in a lot of these new releases. Those wont be coming around till the next couple of game release waves.
A lot of the problems mentioned also apply to other game engines out there, and the issue with proprietary engines are limited documentation, knowledge, and in the case of EA's Frostbite engine, these engines are typically not built for certain game genres, so they have to take years completely retooling them. I don't see the performance hits being much different than a lot of AAA games out there.
That is all a problem, though. Most AAA games run poorly and look bad; that is not an excuse to accept more poor running and badly optimised games. Also, the threading issues have been present since UE3, though, it's kind of a given since it's a general purpose engine. The problem is that no studio is going to take years refactoring an engine they don't own, to optimise it for a game that may or may not sell well. So they will never allocate resources to optimising or refactoring Epic's engine; and Epic is not going to refactor their own engine for someone else's game. So it's a catch 22 with no resolution. We are basically stuck with the current cycle of poorly optimised games on two year dev cycles that play horribly on top-end hardware.
@@Billy-bc8pk Other AAA studios don't have much of an excuse, but Epic is constantly improving and optimizing the engine. UE already supports multithreading to an extent, the threading issue is also improved in the latest versions and there are ways to mitigate it. The problem is that there are a million ways to do different things, for different game types. They actually do work with other major studios (including CDPR now), and those lessons in return are added back into the engine. Open world for instance is an area they're focusing more on. It's all a case by case scenario. The biggest case study they have is Fortnite. Are there FPS performance issues there? It's fricken running on mobile. Again, studios need to actually apply the lessons that UE puts out in its documentation, Unreal talks, and forums.
Most people learn from other people that also don't really know what there doing. 99% of tutorials for unreal will tell you to cast to any blueprint at all that you need info from. That alone is a massive performance hit. Not to mention the amount of things people run on ticks. Most people have no idea that spawning something then destroying it does not completely remove it. Thus, things build up very quickly. Literally even a bullet in a fps game. If you just spawn a bullet then do delete it or do nothing to it, it will still add to the objects in the world and just stack up until your on 10 fps
if you spawn and delete it will be marked for garbage collection
Monopoly is always bad for the market. No matter how you look it, from which perspective you look it, even if it looks good for the industry, it's end result is always bad.
Yes it's a monopoly, and so many people refuse to accept this and waste thousands of dev hours and funding on Godot or Unity modifications instead of investing in a 3rd party fork of UE that fills the gap of what's missing in the UE dev space(fast optimization workflows).
5:12 --This is not theoretical, we already proved it's NOT effortless and we have more proof on the way that destroys the "fixed cost of nanite" lies going around the dev space.
7:27 --The take here is a little too simplified. The game is mostly static, there is no reason to have a dynamic GI solution. And while we don't advocate for lightmaps, we need systems that are precompute/stream local lighting information and allow moderate dynamic interaction which UE doesn't really provide.
Something like a DDGI+Distance Field Indirect Shadow method would be better. Lumen reflections where not the problem in SH2R and anyone who watches the analysis on it we did will know that.
6:11 --More like UE designs these demanding features to only work properly when combined. And again, reflections are not really that much of a problem. That example you used for instance is mostly traced from screenspace (that's been done since at least 8th gen consoles) and then fallback on HW accelerated reflections.
Sad to see you not mention that half the pipeline in UE5 requires Temporal smear from algo's like DLSS or TAA but I guess that's something only we cover.
UE5 does not help developers or gamers, they only help the highest level of employers. Many gamers and developers are sick of these issues but like you and I are saying:
It's already a monopoly. So instead, lets try fixing the problem that's been here all along(UE via fork, plugins, usf/source patches etc).
Thank you for commenting! Unfortunately I am not a developer so a lot of this video was based on research, so I am sorry for using more simple examples or missing topics like Temporal smear. Regardless comments like this are very helpful for me and others to better understand certain aspects of UE5 and appreciate you typing all this info out.
@@major_trenton You're all good, just wanted to clear some things so the wrong things aren't blamed but don't worry about missing these topics since we are pretty much the only people that cover this. But I'm glad to see other youtuber's define this as a monopoly!
@@poleve5409 Research the NvRTX branch(a UE fork, you access it just like UE source code). We intend copy the management of that project.
forget forks, there a variety of great game engines out there for whatever specific kind of game you wanna make
@@arisumego I suggest caching up with our views. We are very specific about using UE5 and will not be supporting ANY other engines. Too many resources have been wasted on such unrealistic goals.
As much as I adore it, I think a good example from this is a little demo called Stonemachina. It's an indie sauce like where you play as little Stone chess pieces and you fight Angels based on Italian mythology. It also has DMC style switching and it's absolutely phenomenal.
HOWEVER, it is also in unreal engine 5, with photorealistic textures and sky lighting, meaning it _chugs_ sometimes. Not to mention it's a demo so it's not fully optimized from the get-go. So this fantastic game with fantastic gameplay is being brought down by graphical choices and how demanding the engine is. I'm having dips on a ryzen 7700x for an _indie game._
Now don't let it deter you from trying out the demo it's actually fantastic. There are graphical options that won't melt your computer but my point still stands
There's a lot of wrong in this.
I'm a UE dev with a game released with all of Unreal's features, so here is my view on the subject:
Primarily, the engine is not the issue, aside from stuttering (which is slowly getting fixed probably thanks to the help of CDPR) the issue comes from the developers misinterpreting Epic's statements on their technologies. The most recent GDC and Unreal Fest talks clarify a lot better what the technologies are, how to use them properly and what are the limits. AAA games today are released half-baked and poorly optimized leading to bad promo for Unreal Engine while other lesser known studios are releasing much better games with the same engine. So the problem is not the tool but the user.
Stuttering issues are real but can be mostly avoided by implementing some sort of shader pre-compilation on game launch, this doesn't avoid everything but it helps.
Using demanding features such as Nanite and Lumen is a developer choice, and there are no other techniques that cost less for "Performance Mode" as it would basically be reworking the entire game for just a "Performance Mode", these modes don't promise 60fps, just "performance". For example using Lumen, you can't use anything but Lumen since you'd need to relight your entire game for it to not look horrible when turning it off.
Nanite has a base cost and in my experience doesn't cost a lot more when using a lot of Nanite asset, it is simply a better LOD solution in my opinion and I use it heavily, including the Experimental Nanite Tesselation in a packaged project.
For the upcoming Mega-Lights, initial testing helps getting a lot more performance out of lights, which are a huge bottleneck in my project and probably many other projects, the feature also fits in perfectly with all of the engine's features including Nanite and Nanite Tesselation with no additional cost.
For the monopoly concern, there are other engines competing, but they don't have the same promotional power that Epic has, there's obviously Unity with Unity 6 and soon 7 that will probably be better than the previous Unity versions and might just reach a level closer to UE and there are a lot of games especially indie ones made in Unity. There's also the Flax Engine that is improving over time, Godot for simpler games, and probably others that I don't know.
Even if Epic had a full monopoly, they would probably still develop the engine, because I feel like Epic is more passionate about their product than some other engines, which is a great thing and is very promising for the future.
To conclude, yes, Unreal Engine is all about promises, promises that they actually hold, they continuously improve, fix and add new features to the engine which makes it better every year, and I'm sure they'll figure out the rest of the issues in due time, but for now AAA studios have to stop being lazy and actually make good games with the engine instead of producing bad cash grab after bad cash grab
Dead island 2 is a prime example of how to use UE5. Regardless of how good(not good) the story is , that game is visually stunning asf and has some of the best graphics the ps5 has seen. And it runs beautifully.
Thank you for this video. It's good to see more people waking up to these issues. The stuttering is caused by a lack of optimisation for multithreaded processes. Assigning cycles to processes on sub-threads could alleviate this problem, but the main issue is that with a general purpose engine, Epic can't know WHERE to assign parallel processes since everyone's game is going to be different and require different computational modals.
In order to properly resolve this it requires a complete backend refactor of how the multithreaded processes are subdivded via their task priorities. There is no easy way around this, though; most of the logic and upfront computational thoroughput is handled on the main thread, and so there is constant stop-and-go stutter due processing issues. Attempting to resolve this could end up being years worth of refactoring. And it could end up having knock-on effects on other recently implemented systems, such as the new mega-lights, as well as existing systems such as Nanite and Lumen. And rewriting all of those could have serious impacts on in-development projects relying on Epic to provide support for their UE5 projects.
This puts Epic in a very precarious position: do you rewrite the backend multithreaded handlers to improve thread performance and hopefully reduce micro-stuttering, causing potential problems with existing plugins/systems that may also require refactors? Or do you keep pumping out new features and let the studios handle the refactors on their own dime/time?
In the interim, Epic doesn't lose much because everyone is switching over to using the UE5 since most top companies are losing their engineering talent who maintained their proprietary engines. Don't be surprised if the Battlefield after the next one also switches to UE5, since many of EA's Frostbite engineers left to form and join Embark Studios.
In the long run, this could end up being severely damaging, as people will soon recognise that UE5 games have serious performance ceiling issues, and may not be worth it, resulting in people becoming less enthused to play games running on UE5. That could hurt Epic big time, as people may end up associating the engine with the negative stigma that games release in poor functioning states with frame stutter and playability issues, which may have severe long-term effects on sales.
It's an amazing engine which is why everybody's using it, and that's why in just about any industry a product that stands out from the rest becomes the standard. That said, if Epic Games stagnates due to their success, then there's only one way to go from there... and that's down, which is the the pattern these powerful, influential corporations tend to go through. But in the meantime, I don't see the problem in enjoying a fairly solid product that can put out some high-end, professional grade visuals for both game and film industries. If their throne is overtaken by another engine in the future, that means it's got to be faster, easier to learn, and all around better if it can dethrone UE... so I don't see a problem with the way things are going with unreal
UE5 is better for movies and commerical rendering than it is for game development. It might be the worst thing that happened to gaming since loot boxes.
Yep, 90% of UE5 games I've seen are a stuttering mess with blurry graphics. I really hope this trend dies quickly.
@@Zweite93Finally met people that have this same feeling
It is specifically built for game development, but it's up to the game devs, not Epic, to release an optimized game
@@Native_Creation No it is built for general purpose creation; Godot and Unigine are built more-so directly for games, along with Unity, which is why all three have better running games despite not having as many features as Unreal.
@@Native_Creation UE5 has fundamental issues that even Epic themselves couldn't fix for their own games. And Nanite is more often than not performing worse than traditional LOD systems. It's great for movies and commercials but flawed for gaming.
I like what the devs at unreal are doing with the engine, but tbf I also worry for a future where all games look the same.
The developer decide the look, not the engine.
@@mariano4068 sure
@@mariano4068 That is a meaningless platitude when *on practice,* the real world result is to have UE games all have this uncanny similarity between them in lighting and texture work.
@@gorofujita5767You'd be surprised how many games are developed in UE. Just to name a few: Abzû, No More Heroes III, Black Myth: Wukong, Hi-Fi Rush, It Takes Two, Code Vein, Sifu and Pumpkin Jack.
Now would you say all these games look the same? I won't deny that UE has a somewhat "default" look to it, but talented developers can make it look really unique. It's an incredibly versatile engine.
9:93 in the scene I tested Megalights in which had Nanite and Lumen enabled with hardware ray tracing, it doubled my FPS from 50 to 100 on my 4090 in 1440p using the cinematic preset from the editor
Nobody cared about crysis having 3 fps, all just got shooked by what the game was able to do. Now how many years later? unreal is doing stuff that we haven't seen before just like back then with crysis and ppl are just complaining even tho we haven't even seen a game yet so it is all just speculations...
I hate humanity...
Every engine is just a tool with it's own pros and cons. You choose an engine to save the time building one yourself. I have faced issues with optimization with UE, but I don't think it's the engine's fault. There's no real monopoly on the market, IMO. However, unreal could improve on it's learning material so that devs understand how to optimize their games more. At the end of the day, if you rely 100% on visual scripting, you're going to run into issues later on.
The main stutter issues people seem to complain about is shader-compilation and traversal stutter. But visual scripting could be in the mix, warping the impression for the worse, who knows.
@@SnakeEngine I'm relatively new to game development, but I know from the web dev world, visual page builders just came with so much optimization issues that needed external 3rd party tools to clean up. Game engines have so many more factors that come into play, I think it's just a matter of developing an optimized workflow.
Hard to talk about monopoly with UE5 since the engines the games use are generally proprietary engines, so no other studies use them, and there is no impact on the general market
I think that's their policy, presenting features so the new guy can go: omg, wow How does it work? 🤔 We'll improve it in the future nvm... hmmm guys I have idea, let's hire ALS guy for the next update
In short, UE has extreme feature bloat and tries to woo developers with ease of access and glossy pictures, instead of optimized tools.
I'm very interested in the way graphical fidelity seems to have hit a point of diminishing returns while performance and optimization continues to get worse. The overhead on modern engines just doesn't make sense to me. Time was you could do driver-level supersampling to get 4x as much resolution and still hit the 60FPS mark, and meanwhile games today need to upscale to run halfway decent.
Yeah, raw graphics look better compared to games 5 or 10 years ago, but how much better really? I'd rather have something like Skyrim running at native resolution (or higher) without terrible TAA that makes everything a blurry, ghosting mess. These days if a modern game even deigns to include MSAA it will tank the performance significantly, because the headroom is so small to begin with.
Game devs back in the days had to come up with tricks to fake fidelity. Now we are at a point where finding tricks takes to long and the old tricks get pushed beyond their boundries.
I am also really missing the effort old games put into gameplay details. Just watched a comparison between dead rising and it remake. For some reason the original looks better and has more gimmicks.
The new Doom games where amazing with the new gore effects. Not the first to do it but they did a great job. New dead island is also AMAZING in terms of just killing zombies. We need more games that focus on fun and details and less on ultra realistic graphics
And even though there have almost monopoly of game engines, they are barely profitable. This area is cocked, a competitor, at the same level, will not be able to sustain itself without a Fortnite. It also proves it's a bad idea to maintain a proprietary engine, you're multiplying the risk of bankruptcy.
Nice video and audio. I have no comment on unreal as i respect them for all their work ❤
Too bad they all kind of just look the same. Fancy looking asset flips.
Bro go back to the late 2000 Unreal Engine 3 was the same, but I agree. Everything starts to feel samey
A lot of these issues aren't unique to Unreal. In fact, if developers take the time to properly optimize their games UE5 can look and run extremely well.
Fear not! There is still the Unity Engine. So not all hope is lost.
looks like i now wanna fix all of these problems in my game
I don't like the way they push unreal to movie and films in same package of the game engine. Game engine should compact and have good performance. Bloating game engine is never be good. I wish they make unreal lite version where only barebone game engine who's running fast and less performance issue
This. A lite version without all the plugins and add-ons would be a huge boon for litework development.
The new features will need time to be improved, that's the true.
Good video!
Thank you!
Running these UE5 games with DX11 seems a temporary solution for now. SH2 and UD both run better with DX11 commandline added to their executable
Yeah, stuttering is also a skill issue. Overload a truck or use a blowtorch wrong, and things are going to happen.
monopolizing? lol no
quoting a few games made in an engine means nothing compared with over 1000 games created each month.
proceeds to handpick some games that have some issues, ignoring the vast majority of other games that don't have that particular issue, then generalizing, concluding, wrongly, the engine itself has that issue, without offering any proof of that.
how about that particular game was coded poorly? No? Then why other games don't have that issue? Crickets.
Unreal engine 5 is a go to for established lazy "AAA" devs for two reasons - fancy graphics are the main (and usually only) priority in modern games. And secondly they don't need to bother with optimisation at all, cuz nowadays they can tell people: "Stop being poor" and make them buy expensive new CPU and GPU's. This ofcourse is very convenient for them, and their investors from the monopolies of the PC components cartel - force you to scrap your old PC.
Add in the marketing bit. All the time I catch game journalists just having the urge to point out that a certain game "will be running on Unreal Engine 5", especially for game announcements.
Game engines built with future hardware in mind have always existed, so I don’t see why this would be a problem.
It is when the games run poorly, don't look great, and the gameplay suffers for it. As showcased with Until Dawn, the last gen version looks and runs better.
UE3 which was all the OG mass effect didnt even allow any AA besides FXAA which is the worst. Injecting smaa also difn't work. So I was never a fan og untral gor those reasons.
Why every game today is blurry and ugly? Im just playing old games today
Unity needs to step up
Unreal engine works on the liner gameplay but will not work on open work.
I dont think UE5 is user friendly at all. I only went for UE 5 because I really like C++. If I knew of another engine I could use C++ with that was battle tested, I would walk away from UE5 and not look back.
if you learnt c++ you can learn anything else pretty quick. Like C#.
@@brk29 Yeah, but he probably wants the performance advantage of C++ and dislikes dealing with the garbage collector. Yes, you are not dealing with garbage anymore in C#, but still with garbage collection since you want to avoid it in games.
@ then, he can use Unigine, Flax, or something similar, they are not battle tested in terms of published games. But, he can battle test them. There is also cryengine but I wouldn’t say user friendly to that either. Oh and O3DE. Not battle tested yet but seems promising. I still think C# is currently not that behind of C++ in terms of performance. If you are not after something really specific you can pretty much do the same stuff in C#.
Nother great video
Thank you!
We need mor UE3/4 games and using other engines that are capable of cool and good things, pointing or having as a goal having "MAX" graphics isn't the future of gaming, it's a part of it, and executives just don't understand and ask for "the best"
Current Day's Cryengine basicly (but can it run unreal XD)
Some good points, but mostly you're talking garbage. The issues you attribute to UE are misplaced, as the devs don't know how to optimise the game, and take shortcuts. The UE is heavily customisable, and its source is accessible and can be built with custom changes. Your statement: Lumen and Ninite are a problem when used together, again is a dumb statement, as it's all about how they are used and not the sole fact of being used together. I wish people stopped making videos on subject they don't fully understand. You make bold statements against UE5, but there were 15000 games released in 2024 according to steam db alone, so pointing out 5 notable released in UE5 is just not good enough - many sources list only a minor percentage of them as UE5.
news flash perfect games don't exist
perfect game engines also don't exist
perfect consults or PCs don't exist
nobody lives like mr. beast making millions by faking vids and having scams
RUclips gamer logic.
valve can push EDI in his games and company can ask for tones of money from game devs and gamers
can delate accounts and other stuff and his custom game engine who is in private mode
what are most gamers respond aw we just ignore it
when epic allows for everyone to use their engine for free + the source code it's also available + free games from the store every month and free assets
and now FAP who is all in one store and asset place who gives 80% on their sells
so how did the modern gamer respond to it
aw i hate epic because there are a greedy Corop boo Hoo
how tone deaf are you all.
you can stop simping for a shitty company and shitty engine now
@@user-be5kj1bw3d go look in the mirror
you tone deaf its mind blowing
almost on SJW levels
@@user-be5kj1bw3d also go read my other comments
the once where i am talking about other game engines and completion to epic
like aw i don't know
valve and unity
and you say its a shitty game engine
yet you didn't saw any issues when wukong made over 10k this year
aw how convenient.
you guys are just as much as grifters as sweet baby ink and EDI care about repetition for the modern audience.
@@user-be5kj1bw3d " hates monopoly's"
forces everyone to go to valve greedy ass
because every game dev has 100 dollars to spend + the 30% tax cut from valve.
and valves own over 30% of the gaming publisher
you guys are so tone deaf it blows my mind
Honestly if Epic Games make their game store as better and optimized as steam and also gives you an option for offline installer like GOG launcher and optimized Unreal Engine with a plugin to downgrade assets or projects + fix game stutters + their own c++ editor without requiring visual studio, they would really succeed and loved by gamers. I really like Epic but they barely listen to users reviews or criticisms. Hopefully with CDPR, they manage to get their crap together, they would fix a lot of stuff like Kojima Productions did with Decima Engine. Honestly their UI designers suck big time and that is a fact.
you do realize that there plenty of other game engines out there right
some of them are built even from the source codes of cry engine and unreal
need proof
well did you look up wicked engine yet how about upbge
how about you guys promote some other engines then freaking unity
if you want unreal to have " competition"
aw what happened to cry engine
cry engine was justness as good as unreal
and the cry games kick ass on realism even back in 2009
WHEN WILL YOU MAKE THIS VIDEO?????
Why DONT YOU MAKE A VIDEO titleD: BAD ACTOR YOUTJUBERS/CONTENT CREATORS ARE RUINING GAMING BCS OF THEIR FALSE NARATIVE CREATIONS OF 0 KNOWLEDGE ABOUT TOPIC REPORTING,COPY PASTING MISSINFORMED,MISREPRESENTED TAKES FROM OTHER SO CALLED KNOWLEGABLE GAMER CONTENT CREATORS????
WHEN WILL YOU MAKE THIS VIDEO?????
aw boo hoo no completion against epic
yeah, let's just ignore game engines like unity 6 , flax engine , falcon , upbge , cry engine , wicked engine and so on
no , no push SLOB CONSUME SLOB WAIT FOR NEXT SLOB
freaking source engine 2
epic not having competition my ass
you guys are just lazy and grifters
None of the big engines can run on low end hw.
Yeah, unity seems to suck too with simple games on a laptop with integrated graphics.
Such poignant content
that AI voice
Still not AI but thanks for watching. Check out my livestream for me playing games if you want proof lol.
This channel has nooo idea about Unreal Engine lol
Bump
Nanite isnt demanding tho. It increases performance. That all its there for. Poor Performance isnt entirely on the engine. The game devs need to optimize their games
Actually nanite can tank performance a lot. It's mainly there to make your life easier and not go through the painful LOD creation process.
But it comes with a big base performance cost, and this cost doesn't grow much if you add a lot of objects to process.
While traditional LODs have a per-object cost, meaning the more objects you have, the more it's gonna cost on your performance.
So if you do a Nanite vs LODs with 10 objects, Nanite is gonna loose. If you do the same test with 10k objects, Nanite wins.
So it's a question of compromise as always. The problem is, as you stated, that devs generally see this kind of tools as easy ways to get good results and optimization isn't properly dealt with.
RE: Nanite isnt demanding tho.
It does NOT increase performance unless developers didn't optimized their there. You're comment makes ZERO sense.
It's literally 100%+ slower than optimized topology and we proved it several times on the UE forums and in our Nanite Video.
Nanite also demands an incredibly slow shadow system. Stop spreading garbage information.
UE5 lacks workflows to optimize models efficiently with a high enough quality that 9th gen should be showing.
@@ThreatInteractiveIt does, when used in the right circumstances. You can’t just use nanite in any projects for any reason and wish to gain more performances. It’s mainly for films and areas of high fidelity with multiple geometry that needs to be rendered. People don’t understand that since they just took epics word for word of « optimization » without questioning themselves where Nanite is actually useful.
@@ThreatInteractiveAnd for the slow shadow system? Yes and no, it’s just « beta ». Which will most likely be fixed in ue5.5 from what I’ve tested. Better fidelity on VSM, good performance. So yes optimization is a trade off for what you need. But some things are mostly unachievable without those features that are yet being developed in the « BETA » Unreal Engine 5. Which is the beta for the next engine… One day hopefully, they just need time.
@@CodeWithTristan "Beta" is not an excuse. Epic's definition of "production ready" putting that system in fortnite regardless of it's poor quality.
The point of geometric detail to "gain" with Nanite hurts performance so much, developers coat the image in Vaseline upscalers making all the detail a joke. And VSM's look like crap thanks to relying on temporal smear for uneeded denoisment.
RE: So yes optimization is a trade off for what you need-
Optimization is getting incredible results with smart/faster systems.
I feel like you're throwing the baby out with the bath water. UE5 is a great engine but alot of these issues stem from inexperience with C++ (a notoriously hard language), unfamiliar with the engine and western work culture.
Firing your entire staff after a game releases so you don't have to give them raises and then hiring inexperienced fresh graduates for cheap to replace them, and then having them fall feet first into code they didn't write isn't good, not matter what engine you use.
We see company's that foster talent like From software being able to make high quality games very quickly.