Assuming that FO76 is a testbed for the multiplayer part of (a more FO4-like) FO5 (that will use the same game engine), do you still think frame locking can/should be absent with no additional problems?
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Fallout 76 No Longer Has Its Frame Rate Tied to the Physics Engine
- Thread starter cageymaru
- Start date
Nafensoriel
Limp Gawd
- Joined
- Nov 23, 2015
- Messages
- 382
It shouldn't exist in any game produced after 2015 based on the performance now available to at least 50% of the US/EU markets that these companies target. Steamstats isn't an assured metric by any means but over half its users have at least 4 threads now so a good eyeball measurement of "minimum spec".
The only reason to ever use it was performance related and bolt on. Physics wasn't an internal part of CE or gamebryo. This was mostly true for engines of the day which is why havok was so popular. The downside with anything added on its performance cost. A gamer never knows how razor thin a performance bottleneck exists until modding is truly unleashed. Restricting everything to frames allows you to do a few animation steps far more simply and removes the need for any additional sudden script overhead during "battles". At the time and with the hardware open to them(remember most people back then when production started had 2 cores tops) trying to keep certain things under control was very important. Most games of this era won't even stress a modern CPU beyond those 2 cores.
So no, No title needs frame locked physics anymore. We have POWER to spare in modern computers. Even when Kyle shows low FPS for a 4k system that doesn't equate to CPU restriction. We can now EASILY reduce cpu load by dumping calculations onto a GPU and settling for reduced graphical fidelity especially considering even old school tricks like bump maps can make up for a ton of low-end texture issues.
Will this mean future Bethesda titles are bug free? Probably not. All it means is that unless they truly do something studio killingly stupid this time they have the opportunity to correct the two major bottlenecks in the engine. If they choose to go the stupid route then they will deserve what they get. No one is immortal in the game development world.
RPGWiZaRD
[H]ard|Gawd
- Joined
- Jan 24, 2009
- Messages
- 1,217
About TIME, now at least Fallout 76 was good for something. I played Skyrim and Oblivion and Fallout games at 120Hz even with the physics bugs (I got used to hearing flying around vegetables and plates when entering buildings), less of an annoyance than being stuck with 60 fps.
Bethesda needs a new engine rather than doubling down on a two decades old fossil of an engine for Elder Scrolls VI and whatever their upcoming Science Fiction RPG is called.
Nafensoriel
Limp Gawd
- Joined
- Nov 23, 2015
- Messages
- 382
And I repeat you dont understand what an engine is or does. Replacing it solves zero issues.
Want a quick primer on engine vs every other error for FO4/Skyrim? If the problem can be corrected by:
1] The console
2] The Creation Engine Editor
3] Any BSA, ESM, ESP, or ESL file
4] Anything that can be corrected by a papyrus script
It does not constitute an engine error.
Engine errors are typically arithmetic. A wrongly calculated collision sphere because the equation to "Draw" such was written .02 instead of .2 somewhere.
Since I just sat down and tried Starctizen during their demo I'll pick at them a bit. To the average gamer, they'll see that first station standing on the platform and go "Woah this engine is awesome"... To anyone who has actually touched a game engine, they'll notice the absolute lack of geometry and the extremely high resolution textures and post processing. In other words, well-done tricks to make a scene as efficient as possible for the most graphical punch. To gamers from the early 80s and 90s these types of tricks were common and practically an art form. They fell out of favour in the late 90s early 2000s because hardware finally got fast enough to really punch true 3D graphics. If you notice it was also the time period people said games "looked terrible" because we went from well-crafted textures to simple ones to accommodate VRAM issues as well. In short, the engine is not its graphics. Its what allows the graphics to exist.
In Skyrims/FO4/FO76s case there is exactly zero reason every actor cant be 5 million polygons with 8k textures. Compared to UE4 this might cost slightly more performance to display the same thing but then again UE4 can't handle large worlds and cell loading anywhere near as efficiently as CE. It's a trade-off.
Your argument is basically saying every single game needs a new engine. Let's look at the original engines release dates:
Unreal/Epic: May 1998 > 2002 > 2007 > 2014
RedEngine: May 2011 > 2012 > 2015 > Cyberpunk
NetImmerse(gamebryo): 1997 > 2003 > 2005 > 2009 > 2012
Creation Engine(a gamebryo fork): 2011 > 2015 > 2018
Feature-wise you'd be surprised how close UE4 and CE are.
So why does everyone love UE and hate gamebryo? It's honestly simple.. Scope. Most of UE target audience has been FPS games. Games that don't really have a ton going on in the background and are extremely easy to optimize compared to open world models. When a gamer plays a UE FPS they see well-crafted art assets and a screen absolutely crammed with polygons. Then the same gamer loads up something like Fallout 3 and screams that it sucks because they, as uneducated gamers, have no reference point to understand why the art isn't as crisp and the polygons aren't as high. They can't rationalize how much space they are looking at and don't understand why typical occlusion tricks to reduce on-screen polygons don't work in open worlds as well. This lack of education means people automatically assume UE4 can do everything CE can and vice versa. In truth neither can do what the other can as well.. that's why they both exist. UE4 is great for small space rapid development of relatively simple games and concepts. CE is built entirely around a seamless world-building experience. You also can't just slam them both together and expect to get the best of both worlds. Entitlement driven demands can't suddenly change the laws of physics and computers just can't do everything yet.
In reality, nearly every time a video game is released it's on an updated engine. Period. Dot. End. Updating it before the tech debt reaches a certain point it's meaningless.. and CE isn't remotely at crushing tech debt levels especially since they gutted it and modularized the whole darn thing.
This is simply untrue. All game engines have limitations. Like it or not, those limitations limit the scope and capabilities of the game and what the developers can do. The Hero engine used for Star Wars: The Old Republic is 32bit only. It suffers from a number of problems that being on a new engine would solve right away. Their work around was to run the game in two 32bit processes that are tied together. This introduces inherent latency problems and a number of other issues. That engine can't be fixed without massive changes so those limits are baked into the game now. This is an extreme example, but engines that allow for things that previous engines didn't allows us to move forward with graphics, game play, and the scope of the game itself. Being tied to a 20 year old engine has to have its drawbacks. Sure, engines get modified all the time, but some limits can't be overcome without massive rewrites or a new engine. Batman: Arkham Knight was done on a heavily modified UE3 and it suffered for it. It would have benefited from being done on UE4, but that would have delayed the game considerably. Going with a modified UE3 was probably a business decision above all else. Still, you can't tell me that modifying an old ass engine and trying to circumvent inherent limitations is just as good as going with a new engine that doesn't have those limits in the first place. I think Batman: Arkham Knight was a great game but it would have been even better on UE4.
Of course, there are tools and all of that which come into play, but FO76 isn't like FO4 in that it isn't a single player game where a ton of modding will be supported. A game as a service which discourages modding shouldn't be bound to that. I can see why a developer wouldn't want to throw their tools away and retrain all its staff, or have to hire developers with experience in a new engine or deal with a brand new engine from a company that may not have great support, etc. So there are advantages to going with what you know vs. reinventing the wheel. I understand and acknowledge this.
I was going to respond to a whole bunch of shit individually, but I'll change tact and come at this from another angle. Since you know the Creation Engine better than I do, how much of the shitty graphics, texture pop-ins, enemies giving up on animations and other things I've seen with FO76 a function of laziness or ineptitude on the developers part vs. engine limitations? I have done some modding with UE, and I can tell you UE4 doesn't have certain limitations UE3 does. So its a better engine, period. I know CE is modified and improved every time a new game comes out with it, but I imagine that there are some limitations there which could probably be eliminated were the engine to get a massive overhaul or outright replaced. I'm not saying they should go UE4 or anything like that but saying that replacing the 20 year old engine for something newer and purpose built for Bethesda's needs would solve zero problems seems like a bit of a stretch to me.
Nafensoriel
Limp Gawd
- Joined
- Nov 23, 2015
- Messages
- 382
Really? Hero is 32 bit only? So Elder Scrolls online using a 64bit version they internally modified must be a glitch huh? Especially considering the ESO engine is a hero fork just like CE is for gamebryo.
The "limitations" you talk about do not exist. That is because an engine is a tool. When the tool is not in use it is entirely modifiable. Just like a hammer or screwdriver. When the tool is in operation it is not easy to modify. Game engine updates happen early in a project for a reason and are generally feature complete long before anyone actually starts making a game out of that tool. If you want a good story of how easily this part of the process can kill a project or studio.. I direct you to look at ME:A. TechDebt can make it unprofitable to upgrade parts of an engine however though this is very rare with how engines are built now.
Before I answer your question directly though since you worked with the UE3/4 transition let me state that the gamebryo to CE transition was more significant. CE mimics UE4 by being a modular engine at its core. Any similarity it has to gamebryo was intentional. It is also true to state CE did not completely update everything though it did give the option to update it in the future. So as far as tech debt associated with the engine there are no significant barriers to quality beyond the script engine(in FO4 it was limited to 1 script /frame). Technologically... The engine can do just about anything graphically UE4 can do with the appropriate addons(UE4 comes with more bolt-ons obviously).
FO76 though? Cockup. Straight Cockup. I don't know if they underestimated the budget or didn't throw enough people at it but while trying to make atlas work with multiple people was a daunting task almost everything was broken explicitly because of that netcode.
I want to silently applaud the ID team and Bethesda for even attempting it first though. It doesn't matter that they screwed it up they at least attempted it. I cannot express in words how monumental a task it is to get atlas(cell loading world of FO4 etc) to work with multiple people. Gotta remember for all its bugs it still mostly works and this was their first shot at it. Tip your hat to those that took up the challenge before you roast em for their bugs is all I'm saying.
1] shitty graphics
Can't and won't argue this. It's an upgrade yes. Models are more detailed and arts been updated. Lighting was overhauled too it looks like(performance wise massively so). Nothing here is a limitation of the engine though. It is entirely on the heads of the art department. That can either be because they did not have much of one for this project(which is entirely possible) or a limitation of the budget just causing less time per asset. The engine can handle more though so there is no bottleneck here or age. In fact with the lighting change its probably an order of magnitude more efficient than FO4 was.
2]texture pop-ins
Almost invariably this will happen in any game with large enough textures. In this case, it's probably a cross between netcode and just bad QC. I can say that it WILL become a problem in the future since we are starting to seriously consider 4k textures for games now... and that is for any engine.
There is nothing in CE that prevents them from tightening this up and reducing its effect, however. Heck windows 10 even has some built-in tricks to this effect they might not be using. Bethesda has been known to miss some simple optimizations before that they later implement.
3]enemies giving up on animations
This is entirely 100% netcode/Atlas related. It IS the engine. Since it is a brand spanking new part of the said engine I can't really say its limiting since previous iterations couldn't even do the same thing. I can say its a point blank example of why this game needed another year of QA/QC. The engine is, unfortunately, doing what it is being told to do. The problem is it's being told to do it late(sudden animation resets or loading of animations from the T) or it's being told to display the wrong animation. I would need multiple people to see if this was localized to one client or if it was server related(or both). This one goes back to the original problem of making atlas work for multiple people though. Remember in FO4 all cells and most scripts load based on the player as the absolute centre of that universe. Conceptually its easy to think of doing this for multiple people.. practice wise getting it all to sync perfectly is a monolithic challenge. Every single instance you see of these kinds of bugs have the same root cause.. this syncing interaction is not working properly because someone on the chain isn't getting synced right away or in a reasonable time frame. Since it works I feel I can safely state that it is entirely fixable in a reasonable time frame. The fact that it wasn't suggests something else caused them to release what they had to have known would be a buggy mess.
All in all that's my entire impression of FO76. It was an ambitious project that was underestimated budgetarily and time and forced to release in an unfinished state due to said underestimation. I also think Bethesda did not really explain the scope of the project. While upgrades were made it doesn't seem anything beyond the netcode was the main focus. This ENTIRE project was getting CE online. In that light, it might have actually been better to do what many suggest and release this as an expensive FO4 upgrade. I know why they didn't go that route though. FO4 couldn't do what 76 does. To much changed internally to do that netcode. In the end, they failed to explain the project was NOT ever going to be a significant upgrade over FO4 and it was NEVER supposed to be on the same depth as a singleplayer fallout game.
Engine wise? Nothing holds back CE except papyrus and Bethesda themselves.
TLDR:
CE is robust and updated minus papyrus.
FO76s netcode was an awesome undertaking that was removed early from the oven.
Bethesda needs more/better artists.
Really? Hero is 32 bit only? So Elder Scrolls online using a 64bit version they internally modified must be a glitch huh? Especially considering the ESO engine is a hero fork just like CE is for gamebryo.
No, I wasn't clear enough. Only the pre-alpha version used for SW:TOR relies on a 32-bit executable. It actually runs the game in two 32-bit executables that depend on each other to get around the limits of a 32bit program. This was a choice on the developers part and has nothing to do with the later versions of the Hero engine, nor what any other company did with it. What Bethesda did with the Hero engine is immaterial as they modified the shit out of it so all bets are off.
BTW, Zenimax has always denied ESO uses the Hero Engine. https://news.softpedia.com/news/The...e-the-Hero-Engine-Developer-Says-272344.shtml
Yes, engine limitations are a real thing. I don't know why you can't seem to grasp that. UE3 can only handle worlds that are so large and doesn't support textures over a specific size etc. I've done lots of texture modding for UE and you can't make the game use textures past a certain size. You end up with problems when you do. UE4 raises these limits, but it has them none the less. Maybe CE is different, I don't know. I don't know that engine. I just know what the final products look like after release and they are less than impressive technologically speaking. The games may be fun, but they wouldn't be nearly as popular without the modding community backing them. That's another reason why FO76 is such a dumpster fire.
Last edited:
viivo
[H]ard|Gawd
- Joined
- Sep 7, 2005
- Messages
- 1,695
I guess if you're going to like and defend one turd sandwich you may as well embrace the entire crap buffet.
Nafensoriel
Limp Gawd
- Joined
- Nov 23, 2015
- Messages
- 382
Sounds like a creative solution on their part then. I can imagine the latency issue would have been... interesting though. Zenimax Online has always tried to evade the hero engine argument but its hero. Sure they gutted it and "built" a new engine "purpose-built for their game" that just happens to work exactly like a commercial engine with new bells and whistles? Sounds like gamebryo to CE conversion talk to me
. Yes, technically they do use their own in-house engine heavily influenced by Hero though.I'm not dismissing engine limitations. They just don't exist until production starts. Ok, to put it into a workflow before the first idea hits the whiteboard you have to decide what direction you are going to go. You discuss your previous project, if possible, and any major limitations or issues attached to it. Then you have a bit of brainstorming about how to work around those and what kind of features you can expect to get with those improvements. Along the way, you also figure out what budget you have for engine upgrades OR replacement. There aren't really engine limitations at this point since everything is open for improvement. The only limit really is hardware and budget.
If you mean rendering limits? I don't think Bethesda has ever actually pushed the engine that direction. I don't even think the engine can be used to its full capacity that way until they update that damned script engine BOAT ANCHOR holding it down. For FO4 it was stretched to its limit hence why I've point blank said it will have to be replaced for any future projects. I could be having rose coloured goggles on about that though. Fixing that one issue unleashes a metric ton of bottlenecks though.
/viivo
Define turd? Unity isn't made for anything other than fun stupid projects. Some people are insane enough to make full fleshed games out of it but it's one step shy of "game creator Plus!". If you have the talent its one of the ways you turn into a super uncle by giving your nephew a custom-made video game for Christmas without spending all damn year on it.