Why More Games Are Switching to Unreal Engine, & Star Citizen Doesn't | Halo Studios
HTML-код
- Опубликовано: 12 окт 2024
- Halo Studios is moving forward with Unreal Engine over the Slipspace engine, and this brings up the topic of why games use their own engines and what makes it more difficult. Here is a look at why some games switch, and some like Star Citizen don't
If you'd like to watch these streams live you can follow me here:
/ spacetomatogaming
/ @spacetomato
Audio Podcast: anchor.fm/laun...
Support me:
Patreon: / spacetomato
Paypal: paypal.me/Spac...
Giveaway: www.spacetomat...
Gaming Community: / discord
Star Citizen Organization/Guild: gii.spacetomat...
If you'd like to join Star Citizen, choose a referral code from our randomizer to get extra credits in-game!
jointomato.com
#starcitizen #StarEngine #UnrealEngine
when talking about engines its better to keep it realistic instead of theoretical.. When Star Citizen started development unreal was a complete Dump, and that's mainly the reason why cry engine was chosen. in hindsight it probably would have been better to wait a couple of years to develop on unreal 4 instead but the end result would still pretty much be the same, the thing holding the game back is not related to tools or graphics its game logic becasue no other game has been made that has such a huge playable area. Moving engine now would almost definitly mean the death of CIG becasue the Unreal Engine is so far behind on the specific stuff Star Citizen needs to actually work. The Graphics are only a part of a game engine, game logic is much inferior on the Unreal Engine in a clean slate state.
Even if they had waited, the UE4 was lacking in a ton of stuff that came out of the box with CryEngine 3 that saved CIG a ton of time in the long run, such as procedural IK animations, which was baked into the engine, so they didn't have to do a bunch of different IK-specific animations for every type of undulation. CryEngine 3 also had global illumination baked in, something that didn't work well in UE4 for years. So that helped with creating planetary scale environments with proper day/night cycles without having to "cheat" it with pre-baked shader cycles.
Given how much refactoring was required to get the CryEngine to do what it now does as the Star Engine, I don't know how well that would have worked out with the UE4. On the upside they could have licenced a bunch of plugins from the asset store to help speed up some things, but art asset development has never been their short coming, and that'sm ostly what the plugin store helps facilitate. All of their biggest hurdles came from -- as you mentioned -- game logic and network integration, two things the UE4 does not do well at all in any capacity.
While Blueprints are a brilliant design interface to make workflows super easy, they also come with a lot of unnecessary processing bloat, and this oftentimes shows in the end-product with a lot of slowdown and over-reliance on heavy brute forcing from the processor. There's also the issue with AI, which CIG would have had to gut and redo from the ground up using their own tools, since UE4's built-in AI tools were not (and still aren't) very good.
So yeah, probably would have ended up in the exact same spot, having to refactor a ton of stuff and redoing lots of the backend to bring the engine up to par for Chris' vision. Strangely enough, all of the things that transpired with Crytek and Amazon actually helped with CIG brandishing the Star Engine as their own.
Unreal engine is great, but it doesn't do what Star Citizen needs an engine to do.
@@Nemoticon I disagree with this sentiment. I see people say stuff like this pretty often and I think it just comes from not understanding how the engine actually works. Just like they altered CryEngine, Unreal can also be altered for their purposes and unreal 5.4 and now 5.5 has so much you can do with it. It currently is the best and there is no reason they couldn't make the engine do what star citizen needs to do. The main issue is that it would be a nightmare to switch engines now and they should not switch. The worst part would be having to train everyone on unreal workflows. Though it uses C# it mostly uses blueprints which is unique to unreal. They are committed and will need to finish the game in their own engine despite how good unreal is today.
@@TheLoneLlama I know engines an be altered to do the job... but CIG is making their engine, based on the CryEngine from the ground up, completely custom built for the game. If they were to do the same with Unral, they would need to have A LOT of input from a company outside of their own team. It's just not realistic to do so. It wouldn't just be 'altering' the Unreal 5 engnie as we know it, it would be re-writing so much of it that it would barely be recognisable any more. But that's no shade on Unreal 5, it's just the Star Citizen is a weird and unique project run by a weird and unique developer. Therefore, it isn't the right engine for the job.
@@TheLoneLlama If you were to ask Epic Games... they'd probably say, "yes we 'could' do Star Citizen on our engine, but we wouldn't want to, lol. CIG have done whats best for them and we're happy where we are".
Sean Tracy as lead developer of Maelstrom, coming from Crytek which we all know how insane the physics was in Crysis that gives me a lot of hope for Starcitizen and destruction on a scale we've never seen before.
And yet we still have ships that supposedly weigh tens of tons being moved by relatively moderate winds...
I loved Crysis
@@bclt64 I know, but as you see even the flightmodel is not finished yet, so we will have to wait and see when the Maelstrom gets in to see if that changes things too.
@@TruthIsKey369there is a reason no one used the Cry-engine.
That was 2007. Cryengine has been pretty shit since then. Nobody wants to use it
The reason why games choose to use existing engines is to not reinvent the wheel when there's no need to. If a game engine can provide all the features required for your game, or at least the vast majority of them and it's fairly easy to implement the rest, then it's a good idea to use an existing game engine. For example, if you want to make a CoD clone you should not make your own game engine, UE is a good bet although I would stay away from lumen if you want a competitive multiplayer game.
But if your game has very large and specific features that have never been done before, if there is no engine that is specifically designed for those kinds of features, and the scope is too large to just adapt an engine, you make your own engine. Then making games with similar features will be much easier and most importantly much easier to optimize.
Switching to a different engine NOW would undo basically ALL of the work CIG has done over the past decade. Not only have they put an insane amount of work overhauling CryEngine into the StarEngine we have today, but literally everything about the game would have to be redone from scratch: all the game mechanics, planet tech, locations, ship flight, the ships themselves, EVERYTHING would all have to be remade in Unreal from basically scratch.
I'm a bit sad to hear CDPR is using UE5 for Witcher. I know they had a lot of problems with the red engine pushing CP2077 out, but holy shit, does that game look and feel amazing in 2024. I know a lot of it comes down to art direction and creativity, but so many UE5 games look like other UE5 games. Games dont look like CP2077. The closest for me is RDR2, which is also built on a custom engine. Maybe the devs at CDPR and Halo will make it work, though. If you have better art direction and are less lazy in making assets, we should see more games coming that have a unique look and feel.
The lead engineer for the Slipspace engine(Halo infinite) lied to the higher ups about it's issues and limits. Design decisions were made to make halo more "modern" with scope glint and other things that no one actually wanted, while ignoring what we did want, in addition to making everything we had Day 1 in CE not present at launch... I'm talking the bare minimum of gamemodes. Meanwhile all criticism was quickly silenced by moderation and the dedicated Halo youtubers were toxically positive about the whole thing because their livelihood depended on it.
Truth is, the fragmention grenades from the OG CE look and sound better than any explosion in Halo Infinite. The artistic design was that of a cartoon and the flood was never looked at again because they wanted a "T" rating ....even in Halo Wars 2 they redesigned the flood entirely to get that "T" like it stopped any of us when Halo 1 came out?
Meanwhile Microsoft has this garbage contractor policy that only allows 6 months for a contractor to work for them....
This means a complete stranger has 6 months to learn an entire toolset...make something useable....and then all that knowledge goes into the void every 6 months.
Microsoft policy is a problem, 343s leadership was a problem and may still be a problem, and the focus is a problem.
Star Citizen may have some problems but they aren't to this level and what Star Citizen is trying to achieve would probably not work with UE5 due to how custom everything is. To CIG's credit...they hire people instead of relying mostly on contractors. Keeping that knowledge in house is what you need to run a custom engine.
EDIT: DON'T GET ME STARTED ON THE MICROTRANSACTIONS!!
This, this, this. You said everything that I don't have the patience to type out. Also, what a lot of people don't consider is that while Unity and UE5 are good general purpose engines, if you want to do anything specific you have to refactor core aspects of those engines. It's easier with Unity because there is less bloat. With UE5 there is A LOT of bloat -- but the bloat is necessary for its general purpose use. In order to get it to do what you want you have to tear out a lot and start from scratch like what Embark Studios did. But most studios -- and certainly not 343 with its contractors -- is going to refactor the UE5 for Halo, and there is no way in the world anything you want the engine to do systematically will be accomplished within a six month refactor.
From what I know of the history of the Star engine was based on CryEngine (the first game I saw on that engine was Far Cry which was revealed fairly quietly and was a revelation in terms of gameplay) the engine was chosen for its ability handle large areas, and scale the detail well, I think it was using voxels or something... Unreal etc were at the time geared towards rendering the inside of boxes. StarEngine has built upon the idea of detail and scalability- with the scale getting pretty damned huge... to the point that they had to double to number of digits used to locate things in the game maps.
Add into this keeping track of thousands of unique entities with the persistence in real time, it's suddenly looking at a project that requires much money, time and ambition that for most game engine producers is too niche to sell, or certainly sell and make a profit.
Unfortunately the vocal minority doesn't understand this simple fact and they regurgitate hatred onto the project on a daily basis on Spectrum and Reddit, spewing misconceptions and sometimes outright lies about the project. But, I guess they'll never stop...even when SC becomes more of a real game from an alpha to beta and then release, there's still going to be a contingent (perhaps smaller) of the same people whining and moaning despite them being proved wrong
There are pros and cons to the industry centering on unreal, the pros are that hiring new devs is easier as experience is more likely to carry over and you get consistent upgrades to the engine without the decelopment costs
The cons, one company has a scary level of control over the industry and centralization will consolidate bith success and failure
Remember how much impact Unity had on the indy game landscape when they went off the deep end, it'd be pretty devastating for many years if Epic somehow ruined everything
Yeah, lack of competition is an issue everyone is overlooking
I think CIG is committed to Star Engine now. Not only because switching to a new engine would trash a decade of work, but because I think one of their monetization plans going forward is to license out the tech they've developed.
The engine video should have been a hint to this really, they would be silly if they got all this tech working and didn't license it out, I'm surprised more people aren't under this assumption.
about the 45min quantum travel. expectations haven't changed for people who were backers back then. but we have people now who came later and can't fucking read what SC is about
I'm fine with 45min or longer jumps if we have gameplay on the ship itself, maintenance, tuning, switching/preparing for FPS or prep. snubs/small fighters or killing the time lest playing pool/snooker with the buddies.
@@Phoenixstorm36
Agreed.
As an example, lots of Star Citizen players talk about how they want Euro truck but in space, without actually knowing what that means.
In Euro truck, there's plenty of gameplay in the traversal. You have to follow the road for one, you have to avoid cars, navigate traffic laws, etc. Not to mention you get to see many different vistas throughout the travel.
Quantum travel in star citizen has none of that. You just press a button and do nothing. You don't even see any unique vistas; you're just sitting there doing nothing, staring at empty space for a while. Maybe in the future you'll have to time quantum hops or something with the new quantum boost, but that still doesn't add enough gameplay. It's still nowhere near an actual trucking sim.
I'm not saying they should add roads to space or anything silly like that. But a little actual gameplay when quantum traveling (and something physical, not some arbitrary minigame) would help a lot.
Even the most basic of game engines from scratch is a none trivial task. Basic IO, rendering, shading, window management, UI, single threaded, multi threaded, sound, networking, 2D or worse 3D rendering, physics, lighting, shadows, reflection, occlusion, dynamic vs static, animation, plugins with external tools and data formats fir assets (audio, fx, vfx, music, fmv, mo-cap, models, sprites, images. Software Engineers (Competent ones) are not cheap and prod ready game engines require a lot of them including lots of domain expertise which is often the premium engineers with years of experience at the top of their game from multiple complex fields.Especially if you want to run a small specialised team to push the boundaries.
Honestly, the list is endless and scope creep does not even begin to describe modern AAA/AA commercial, proprietary in-house and even open source/community developed engines like Unreal, Unity, Godot/Redot, Stride, RPG Maker, Frostbite and on and on.
To make something user friendly, relatively bug free, performant, modular and affordable let alone with an ecosystem if assets, plugins, integrations, pipelines and support? Mammoth task for even large dedicated well funded teams.
I helped make a basic 3D rendering engine with physics and a simple editor with support for importing assets and viewing them in my final year of Uni and it took me and 5 other students with far too.much time, not financial constraints and nerd level focus 8 months of pretty much full time work with late nights plus borrowing art and sound engineers to make assets and models/prefabs to test out the editor, pipelines and such,
I only worked on the renderer and physics and only dabbled in everything else as needed. It was hard, I learnt a lot and made a lot of mistakes along the way but damn, taking that to production and commercial viability for making indie games would take years more work. plus deploying games to multiple platforms, making the entire process unified. Nightmare fuel. I now specialise in databases and networking low level system software with huge data sets and simulation/machine learning and I owe a lot to these sort of hobby/thesis projects fir building my fundamentals and getting deep in the weeds if OS, kernel, hardware, GPU, CPU and memory and IO/networking. It forces you to learn a bit of everything and deep enough to be useful if you have the time.
Devs that do this in game dev do not get paid enough. Let alone AI, cutting edge ray tracing, DLSS, frame generation, driver compatibility and so in. its a real commitment that could easily last decades. Plus we built it using Rust and I worked on a port using Zig recently. Really fun but its difficult.
Three sane reasons to make one? Curiosity/learning/fun challenge, lots of money and engineers on staff, push tech forward with a new unique edge on your engine. Anything else is insane and usually not financially viable, it'a a huge money, time and talent sink.
My understanding is that Unreal 4 won’t be able to handle the enormous space of a star system. Unreal engine 5.1 just got the 64 bit coordinate system, while star engine has had it since PU was released in it’s current incarnation
This is true. Unreal5 started on that a couple years ago, but it wasnt mature until recently. A lot of games are just now starting to leverage UE5 tech. Only a few were in on UE5 out of the gate.
The problem with all these developers switching over to the same engine, games are going start looking and feeling like more of the same. Im not saying UE isnt an amazing engine or that they shouldnt use it, I'm just saying that the devs that stick with using their own engine will stand out from the rest. The Star Engine is perfectly fine and has a lot of potential, no reason to switch from it, especially when CIG is this far invested into it
Star Engine is a heavily refactored version of CryEngine 3. So all the updates are done inhouse by CIG. I hardly doubt that Crytek will provide any updates for the cryengine 3 since the latest version they are building is already 5.6
There's a reason why UE is called the "stutter engine" and its because of the streaming of assets through their nanite system, and right now it is more of a crux than a good solution for fidelity.
None of these engines would work for SC for one simple reason.
None of them had the precision required to render on the scale that SC requires.
Remember that they had to upgrade to 64 bit coordinates system in order to acquire the ability to pinpoint specific points in space, without 64bit coordinates, you couldn't jump accurately, you'd end up thousands of kilometers away from where you wanted to be, it was random where you ended up after a major jump. It was hell to travel.
Correct. However, UE5 does now, but it is only just maturing (large world support). And devs still would need to craft that uniquely for what SC has done. People dont understand that no engine really supports it all out of the box, unless you are creating a generic game.
@@ephalanx1 I did use the word "had" past tense, UE5 does have it now, 10 years after, simply because RSI showed everyone it can be done.
People simply don't understand engine development. it takes decades to craft an engine with new ground braking capabilities. That's the main reason many companies keep using the same old engine for years and years, there's also the cost in $ and in time.
@@UtahSuka Agreed. I was really just echoing you for the most part. We are on the same page.
Wish more knowledgeable people like you would post on Spectrum... nowadays it seems filled with toxic detractors whining and moaning about everything and making misconceptions and even spreading lies about the project to suit their doomer narrative. But, I guess interacting with fools is a zero sum game...in the end, you can't really debate with trolls and malcontents, they'll just use discourse as an excuse to keep on spewing out their rubbish
The further the Unreal Engine goes, the less physics is involved. They brag about their CHAOS plugin, but it is nowhere to be seen and I think its their other tech like lumen and nanite that destroys the physics aspect. They're mainly focusing on fidelity and for me, I like fidelity but having physics and destruction makes a game a lot more fun i most cases. CIG knows this since they are working on Maelstrom that looks good so far.
You argue for star engine but mention physics 😂😂😂😂. Start citizen physics are exceptionally poor.
This. If anyone doesn't understand what Trusthiskey is talking about, look at the physics comparison that Keen Software did comparing the different engine frameworks that they were experimenting with for Space Engineers 2. The Chaos Engine has the WORST scaling out of all modern physics engines. It's good for small-scale physics in limited rendering runtimes, but it does not scale well at all for large scale battles and open-world environments.
Unreal has been battle testing Chaos physics for a while in the Lego Fortnite mode, and it works well
The hardest thing about physics is network replication and I hope they release some tools to make physics gameplay easier to make
@@dumpylumpo5525 Additionally the issue with physics is the volume of interactions and entities involved. With smaller interactions, collisions, animation, and network auth, it will tend to work fine. Scale it up and that is when you begin to see the cracks.
All the old BF titles had good destrution.
2042 destruction was a joke.
Im willing to bet, the new BF title will havr shitty to no destruction
Halo sound track still gives me chills. I am also sporting a Halo tattoo for the last 15 years now.
The license fee for UE is 5% after the first million dollars earned. Which is significantly less than most engines today. The issue for CIG is that they could have had to be updating their systems every time UE put out an update, ad when SC was first being developed UE4 was just being announced, and wasn't even out yet, and going from UE3 to UE4 wasn't directly possible due to it going from U Script to C++ and Blueprints.
I think they would have had been better off in UE over CryEngine simply because CryEngine gave the company I was working for at the time a lot of headaches, to the point when my project got under way, they switched to Unreal. So now I am a big proponent of UE, but I haven't looked back and still use UE to this day.
Unreal, while not perfect, is massively easier to develop and modify to what you need than CryEngine was, and probably what Lumberyard was, and now what StarEngine is, though it has likely gone through a lot of changes since then, and the devs working on that engine are going to be the most expert at that, so I would defer to them on that aspect.
I have always thought that CIG will build different games, license the star engine or both. Just a thought that I have had. I mean when SQ42 launches it will either be great or suck. If it is bad, the revenue stream will fall. Then what. The development of SQ42 and Star Citizen are crowd funded. 42 and Citizen can only draw so much. They will have to build a SQ42 sequel as well as maybe something totally different.
So essentially game engine development is too hard and they are unwilling to keep enough info on hand to bring others up to speed, that means they slack off and use someone else's software instead of putting in the required work. I hope this means cig succeeds like unreal tournament did and then some.
Its not only hard, but expensive to pay engine developers. They are not cheap. Some publishers/studios do not have that sort of cash on hand. Its not always about laziness, but practicality and fiscal responsibility.
@@ephalanx1
Yeah but Microsoft has enough money. There are just plain lazy and greedy.
@@ephalanx1 translation of that is that is too hard, it is not only that job area that is corrupted by greed but part of being responsible with money is keeping production in house and not 'renting' from another company, sorry but your response is a cop out.
@@patryn36 You are making the assumption that every studio is full of money and has tons of people, so many studios are made up of less than 20 people. I think you are misunderstanding or misrepresenting the games industry. The people who are making the engine are not the people who are developing the games. If you have people who are already hired and ready to make a game but they have to wait around while you hire an entirely different group of people to design an engine then you are burning money for no reason and you will likely run out of funding before the game is ever made. Just because you hired people to build the engine doesn't mean "you" built it. The devs will still be using an engine made by someone else. Its no different than using unreal. The only important thing is that we get good games made by people who care, the engine is irrelevant.
@@TheLoneLlama i never said anything about what part of the game companies were doing which so the one making assumptions is you. I am fully aware of the general structure of these outfits and they are lazy and corrupted by greed. A telltale sign a person is corrupted by greed is the belief they are deserving of a raise like i see and hear of so many and in the case of the crew i work amongst i do far more than any of them simply because i choose to and i am not damaged in some way and yet a few of them i have heard moaning they are not paid enough. What many do not think about, and i will bet you are one of these people, is that the more a business has to pay its employees the more that gets passed onto the customers, factor that across enough other businesses and that raise they champion for gets wiped out. As for the engine issues, with people prioritizing the easy path like they do in most things and the inherent greed you lot value like i outlined above, it is exactly as i put it in my original comment here that you responded to, not my problem you and others do not want to admit to your own failings.
The problem with Halo.....is not the engine.. the game was absolutely beautiful once it came out. The problem is they just dont have the heart that made halo great. And they are focused on the wrong things. While i am always happy to play a new halo game. It just feels lifeless ever since 343 took over. Before it had suspense. And masterchief was allowed to be a badass. Now its just not halo. Its just another fps game.
I enjoyed the campaign for a little bit but once it opened up it felt like a fieet person just cause mixed with far cry. Not halo
A game engine handles more than just a game's art. It handles EVERYTHING. It's the ENGINE of the game - it's how it runs at all.
And it's probably the reason 343 abandoned coach co-op.
@@LucidStrike in fully aware of what a game engine does as I do work in unreal now.... Changing to a new engine is not gonna fix 343 main issue...which is they lack that spark that Bungie has with the first 3 halos....we have seen it 4 times now that 343 just cannot produce a proper halo experience. They focus on the wrong things when doing so.. there's no reason to of ditched couch co-op. Literally every halo game had it before 343 got their hands on the franchise. And the new engine could absolutely handle it. They ran out of development time and the higher ups told them to only focus on pvp...this was all in the update posts they put out.
I follow everything that goes on with halo in general.
Halo infinite was dying 2 years ago and when they switched focus to pvp that's what's kept it alive. But from experience every single person I know that's not into competitive pvp biggest complaints were the lack of replayability on the story ...and the complete lack of co-op even online co-op
Yeah okay so after 10+ years they're not going to change engines... shocker. I mean, do y'all want to be alive when this comes out right? lol They change game engines and we won't see a final game until 2030. o7
If they change engines we won't see a game until 2050. It will already be 2030 by the time Star Citizen has all of its Kickstarter features using the current engine. Engine swaps are EXTREMELY time consuming and can take anywhere between half a decade to a decade to convert, depending on the engine and the project.
I've been following the project since 2020 and after nearly 5 years of constantly checking in on development progress they've barely made much progress to the game, it will not be finished by 2030.
@@DeepStone-6 Had no idea that PES and upping the player count from 100 players per server shard to 1,000 players per shard was "barely much progress".
I've been on this ride since 2012. I accepted the nature of the development cycle in 2016.
From my perspective its progressed greatly.
However yall on crack if you feel cig should change the engine now 😂. I am not going to live forever man.
Man that Halo music was so good. Can't believe the TV show thought they should change it =D
Unity has shot themselves in the head as far as any devs wanting to use it going forward, and AFAIK no open universe space games use Unreal.
There are no open-world games in general that use Unreal. The large world coordinates system just isn't well optimised for it. But it's a great engine for cinematic visuals and more compact combat experiences. It basically makes you wonder: will they go with smaller instanced areas for Halo to get the best performance out of the UE5 or attempt to go open-world and try to design around the engine's limitations?
Large coordinates were only native to ue5 , I’m making an open world space game but it’s too late for me to change. I’m sure more will be coming once ue5 get more performant.
1. Unreal is still booty if the devs are booty. Get ready for a lot of slop powered by unreal as many dev studios use it to cut corners.
2. Unreal Chaos Physics is actually just ass to work with. It is also way too resource intensive for what you get. Walkable vehicle interiors with large scale destruction just doesn't work in Unreal as it exists right now. The effort to make that work would be orders of magnitude more than whatever you save by switching to Unreal.
3. Unreal handles asset streaming poorly to say the least.
Its so much more complicated than just switching engines. Each "engine" functions very differently from the next. A vetern designer in one would produce at a novice (or perpahps a bit higher) level in any other. Most use different coding languages, most handle animation, lighting, scripting, game logic all differently than thier counter parts. To think it's an easy switch is to simply clarify that you have no background in game production or its complexities.
“and you know what’s special about me? I couldn’t f***ing do it” LOL 😂😂
I would really appreciate a deep dive on why seamless meshing matters. Technically, what new gameplay is enabled by server meshing?
Grabby Hands was integrated into the cargo and loot handling system.
The way they should have handled a new engine is to have a dedicated team of people for the tech side. Not siloed into a single team. So for example the Xbox Core tech group is responsible for developing the core engines while the teams focus on using the tools. As you explained, engine development is not a simple task anymore so the resources required need dedicated people/teams to keep improving it. 343 invested the time and money (like $500 million) into the engine revamp but without the large team still working on it it becomes stagnant.
Undead Labs learned this lesson too. First they were on CryEngine like the first game, then used something else, then started over again on Unreal Engine where they are today.
This is a very good topic to cover especially with respect to Star Citizen. We can similarly look at Starfield and Creation Engine. There is a lot of misunderstanding around engines and why particular studios take the directions they take. Many times its about practicality, timing, fiscal responsibility, and on hand or acquired talent. Excellent coverage as usual ST.
The story, dialogue, and plot of Halo after #3 continually went downhill.
What I think would be awesome is if they integrated AI into the engine. The idea of having in game characters that spoke using AI, rather than from built in scripts. Think how far they could take the game if they implemented a full blown AI into every possible aspect of the game. Think of developing actual personal relationships with AI characters! The universe would really be limitless!
I do hope Sean ended the interview with "Thanks Steve". I wonder if that was already a thing back then.
Jealous you and your dad had a game together. Mine never discouraged my hobbies - aside from gaming in general - but he also never encouraged any of them but reading. I try to encourage my son in everything he shows interest in, particularly when he shows actual aptitude, like with flying in SC.
This post is the best compliment that u ever made to star citizen guys
32:05 I really dont like that definition but I see where you folks are coming from. I know that it is cool to be able to look into another server and therefore we are seemingly "skipping" the loading screen.
I just like to have the 'seamless background loading tech' separate from the 'distributed computation tech'. For a while plenty of games have the background loading/streaming/Interest Management tech already, even in single player games. But only a few of those games, mostly MMOs, have some sort of distributed computation. And that is a whole separate capability from loading.
They are analogous in that the server running the simulation is also streaming the state data in and out of memory as needed. It doesn't just stop simulating a ship; it unloads it entirely. The culling boundaries here are the object containers, which is where things correlate with the client side asset loading. They are still two separate systems, though.
@@Aethid Thank you for your response. I understand what you mean, and yet, the devil may be in the details.
The detail is that the server doesn't directly load or unload a game object right when that game object crosses the boundary.
Game objects are always pre-loaded way ahead of time already. The same goes for syncing them up, way before it is handed off from one server to another. "Loading" before "syncing" before "simulating/authority". These are steps which dont happen all at the same time, they happen sequentially over a long period of time. So they should be seen entirely separate from each other and saying that Server Meshing is just a fancy next gen loading screen isnt really correct. Server Meshing is operating on the level of simulation, not loading.
See it this way: OCS and PES were already about loading. CIG isnt working on loading anymore. They moved on from that to distribute the simulation across multiple game servers. Therefore, the loading system and the authority management system should be seen as separate tech features.
However, I do understand that Server Meshing can only work if each server (and clients too) only load what is currently (or potentially soon-to-be) relevant for that computer. So OCS/PES are essential and an integral part of the whole Server Meshing feature set.
And still: I dont like that definition because it misses the actual functionality of what Server Meshing is about: Distributed simulation/computation. That is the big feat. Making it all about loading again just doesnt do it justice nor does it seem accurate or correct on a technical level. We had that already with OCS/PES.
@unobtanium I understand that. What I'm saying is that server meshing in many ways is the server side analogue to client side content streaming, which is the modern alternative to a loading screen. The server is now "content streaming", even though "content" to the server is very different to what it is to a client. Most online games also force a "loading screen" on the client for a server change (even though the client is just waiting on the server), so from the player's perspective, server meshing avoids a lot of loading screens.
@@Aethid Wouldn't Server OCS for the game server be the better anologue to Client OCS for the player client?
I dont think the "content" loaded is much different between client and server. Both have to load the static geometry and dynamic game objects (entities). (They utilize Entity Snapshots, Entity Aggregates and Object Containers for that.) They are doing this since 3.8 when Server OCS went live. The only thing they differ is that the client has to load geometry models and textures into VRAM of the GPU. Something which the server doesnt need to do since its not rendering images like the clients do. But that should be even more irrelevant in the Server Meshing discussion, as again SM is specifically about the simulation aspect of the engine, not loading or rendering.
The fact that other games hide server transfers between loading screens doesnt make Server Meshing a fancy loading screen. Again, its a good description for the layman to make them understand why it is a big deal and what they can expect. But from a technical perspective it is absolutely missing the point. Besides, even WoW had smooth, uninterrupted server transfers in the hub world back in the day, where everything fades out behind you and fades in from the new server without an actual loading screen. So just boiling this topic down to loading screens is not a nuanced enough perspective to go about this anyways. It is much more complicated than that. So, yeah, that is more or less why I still dont like that definition.
@unobtanium Yes, and server OCS is the core tech of server meshing. Server meshing is not a single feature, but a collection of many capabilities coming together. The ability to compartmentalise the simulation, sync that state to a shared data backplane, and load/unload subsets of state into servers for simulation is the majority of what is server meshing. That is server-side OCS.
WoW's ability to seamlessly move players between shards and phases is similarly sophisticated, and not really representative of what most games are capable of. Even other modern MMOs can't do what WoW was doing a decade ago. Also, again from the player's perspective, all that tech boils down to a lack of what would otherwise have been a loading screen.
Why would CIG switch to Unreal from StarEngine? Unreal can't handle the scale of Star Citizen.
Using Cryengine in 2012 could be understood. But then they went full "remake everything" mode in 2016, it would have been the right moment to switch over to UE.
Developing your own engine takes 25% of your dev time... so why not use somebodies else engine, pay 25%, and get an engine that is already up and running, instead of building the basement of your own house while you already try to move in in the 10th floor?
Maybe, but then they acquired a bunch of Crytek engineers who actually built the engine. There was no way CIG would have had an opportunity to get a bunch of VERY experienced devs if they went with UE, especially devs who actually have built an engine. UE would have had to heavily modified too and with most likely less coders and less experienced ones.
lets hope the consolidation into unreal doesnt lead us to a stagnation in technology like IE did to web for so long.
Great video, buddy as always
unless unreal can support server meshing, no reason to.
The Main benefit of Unreal Engine is the ability to Find Programmers who know how to use it. Its NOW Industry standard meanwhile its One time Competitor Cryengine got left behind and Only a handful of games use it Star Citizen with its Own version of it being probably the Biggest one. The Hunt and New World being the other Two I know of. oh and Kingdom Come Deliverance 1
The real issue is the scope of both official & community dev support for it compared to UE. Just less of a headache to use UE if it fits your criteria.
And KCD 2 now, also the engine that Far Cry uses is based on Cry Engine.
Unreal engine is the goat of game engines :D But a game engine is only good for the things it was designed for, if you want something that is more specific... eventually those specific things are included in unreal or unity.... I think SC requirements are not fully answered by any of those game engines, so that's why they are making their own game engine even thought is a monumental heavy work, lots of lower level coding (speaking about C / C++) and gfx specific coding (shaders and renderings)
Halo!? ...same team, same problems. Engine won't help sadly... Miss games when those were good, Halo also...
Unreal engine is good, but it has it's flaws... Biggest problems are with their nanite, even though, for game large like SC is, nanite would probably help a bit over LODs...
I read they scrapped the whole upper management from 343
@@therecklesswarlock6439Even if they get new management, I'll judge their company after 1 or 2 games. Even 343 when first launched was for creating games for "fans".
@anshukandulna1844 im with you on that one. We need to see results before we start believing. We've been hurt too many times.
game engine monopoly
Gearbox at Xbox has used UE since forever, and are one of the studios that has the most knowledge using the engine n the industry. Why did Xbox not consider UE for Halo long before? THAT reeks incompetence.
Ha you nailed it! Finally someone can see! Love you Tomato!
If people really ask this question, than i think they are really stupid. The work that has to be done to port this all over to UE5 is another 5- 10 years work. I guess people are just insane, they think hit button and its done or something.
Resume @1:10:00
Because gamers do not know what a game engine is, besides the name and "its looks".
It is a redundant argument.
Wait, you failed at the Tower of Hanoi solver? :D
No, making your own tools is very common in game design even if you work in UE, Unity or whatever you picked.
I always build different tools with some simple programmer art UI, but most general purpose engines by now have multiple import tools available by default or the external editors have export tools available to export in a supported format.
Halo CE best game ever
Unreal engine has limits with large-scale projects like SC from what I hear. Is this true?
Without HDR and Pathtracing the graphics will look outdated. It has to compare with Cyberpunk 2077
Funnily enough, Red Engine does both superbly and is the only engine that enables those features in an open-world setting that can run native 60fps at 16.6 m/s frame buffering without major hitching. So far no other engine, including the UE5, is capable of that right now, which is kind of funny, because CDPR moved away from Red Engine to UE5 as well. Technically speaking, if Halo was chasing graphics, they should have licenced the Red Engine.
Both will be shipped with SC. Confirmed.
Pathtracing? I dunno dude... Games dont even use actual Raytracing to render its scene yet and we want pathtracing?
I rather go for good graphics and high fps, than great graphics and low fps. That is why I still have an issue with Nanite. It is running great but you have to pay half your fps upfront.
@@Billy-bc8pkor maybe id tech?
Unreal Engine is legit garbage, it was never a good engine but did what the industry needs not the players, every game i played in UE engine the anti-aliasting sucks, always blurry as hell and the performance is like meh even in high end pc's, you can tell by new UE5 engine how heavy that thing is...
Living in the building while you are building it. That's CIG and SC/SQ42.
And that's many other game companies too...
Readily available engines didn't just exist back in the days. There are so many companies who made there own engine. Most of the successful games were made on their inhouse engines. And that is the best way to make any big game projects even to this day. Because that's how you make sure that the engine will support the game that you are building.
The only big difference is that CIG lets us playtest their game at this early stage, something that almost no other company allows with a project of this magnitude. But CIG had to do this because they were a crowd funded kickstarter project, starting with a team of 5 or so. They were literally an indie company for many years, entirely depending on crowd funding.
As Aztaable said, it's also many other successful studios. Hello Games did the same thing, but funnily enough people praise them for it instead of deriding them for it like they do CIG.
"StAr EnGiNe" or "unprecedented tech". RSI simply just yells scope creep. A scope creep that went haywire. When they will release something tangible and production ready for other companies, probably others will move on to greened pasture or develop their own more efficient solutions.
It’s very simple, and the same reason backers can’t walk away from Star Citizen easily: sunk cost fallacy.
halo studios dei detected no buy
Considering the tech debt and insurmountable compounded bugs they have created, it would still be faster to finish the game if they restarted with UE5 right now, than if they continue to stumble along with a broken Star Engine.
thanks for ruining empire of the ants for me
Because they wasted a decade on outdated tech, trying to reinvent the wheel when they did not have to.
They wanted to save money, now it's going to cost an entire game redeux to stay up to date.
Halo latest game was garbage
Who the hell would ever use star engine
RED engine sucks balls, so does the Creation Engine 2. So dated and bad, and you can see it in their last games. Cyberpunk 2077 looks good if you mod it, but it has so poor performance and bugs allround, so it is still bad.
The nost likely hobest answer why they are using their own engine, sunken vost fallacy
bcs switching will be insanely complicated and game will be never done(which still might happenú, but the game already looks pretty bad compared to UE..by the time the game is done (or in early very early beta) , the game will look very, very old... wel i doubt they will ever switch bcs they are bad with this engine, and ue might be too complicated for their little brains.. all they do is sell jpgs and gifs....
What's halo? Pfft. Never was a fan of the franchise.
using the "sales engine" chris has made the finest crysis mod ever. shame it will never be a game.
CIG CryEngine or as they call it StarEngine, has been outdated, like 2-3 years ago. Whatever CIG does to keep it alive; it will not be enough. I guess that is also why they implement all BS time sinks in order to make game play slower so the Engine can follow. Yeah, SC still has allot of eye candy fluff but in relation to game play and mechanics the engine is a bottle neck!