Amy Hennig Talks More About Struggling With EA's Frostbite Engine

In house development teams being required to use a tool versus third party tools that have to view developers as their customers was never going to be a fair comparison though. Granted there are exceptions, but an engine that has to win your business, with multiple releases to iterate is going be approached ease of use, documentation and so forth very differently. It's like comparing customer service at DMV to a restaurant...

 

I follow quite a few industry people and the common perception is that executives love the idea of Unreal shaving time off development, but in practice it never happens and UE and middleware inevitably lead to larger teams than in-house development. So damned if you do, damned if you don't.

Frostbite is a known thing. It's an internal engine. I saw something yesterday from a Frostbite dev saying studios are allowed to branch and then submit back into main. If the company has decided Frostbite is the solution going forward, then it's on the studios to understand where their risk points are working with the engine and fix them. That's just the reality of the situation. If it's not good then it's on them to make the case that they need something else and either develop it internally or purchase one externally. Anthem was in pre-production for 5.5 years. I'm sure Frostbite may be less perfect for the game they wanted to build than Unreal is, but they had 5.5 years to convince EA that they needed something else, or improve Frostbite.

 

Game development is a huge field. It includes writers, artists, community managers and anyone else that shows up in the credits. But as a game studio, having domain expertise in the sub-systems a game engine would need, and having a tools team is pretty much mandatory because every game requires new tools and new features. I don't fully understand how multiple studios can make stabs at developing games, fall short and then the engine does not improve. Worst case, studios should be submitting improvements back into the engine so someone else can benefit from their failures.

So the Star Wars team missed the mark. They had issues with Frostbite. Amy Hennig said they took it for granted that it would have features, tools that were common at Naughty Dog. But that's not an engine failure, that's a human failure. Someone made assumptions that weren't true, that added high risk to the game they were trying to make. They can say, it couldn't do X, Y and Z and that made our development go off the rails, but it was their job to identify X, Y, Z and fix it before they went into production. If someone is saying, making Star Wars: Drake's Fortune in 3 years, but the engine requires a full year of upgrades, then that timeline might be a problem. But you have to identify the problem in the first place. The studio is accountable for that.

 

Yeagh. The real issue here is game development is huge and complicated. Any off the shelf solution will pose problems. Any in-house solution will require huge amounts of work (bear in mind something like Unreal has been a WIP for twenty years!!). And no matter what you have, the bigger it gets and more diverse it gets, the more unwieldy it becomes.

It's probably true that every project will run into one of a number of different "Oh my god this is sooooo crap" situations no matter what engine is used.

 

Not to mention the mods referred earlier in this thread, that Amy Hennig's team did for adventure games with large levels.

There's certainly not a lack of work done towards Frostbite to make it more versatile. Maybe DICE isn't being very effective at aggregating these modifications in recent builds.


One thing I can say, though. I just played the single-player campaign of Battlefront 2 because I was interested in the story aspect of it, and effectively the single-player elements (interaction with environment, transition to cutscenes, NPCs, etc.) are cringe-level bad. Sure the graphics and sound are great, but the thing is practically a multiplayer game with bots and a linear corridor map.
In some levels you play as a hero and the viewpoint forces itself to 3rd person. Playing as Luke Skywalker or Kylo Ren in BF2's single-player campaign (3rd-person lightsaber combat) is several light-years behind the 15 years-old Jedi Academy.

 

Because that's literally what it actually is. (and that obviously has nothing to do with Frostbite). The sole purpose of the SP portion of Battlefront 2 is to stop people from bitching about the lack of SP as they did for BF1.

 

How do you know that the the different EA teams working with Frostbite does not do this?

 

So you think the clients' request for a single-player experience means the developer does well to release a mediocre single-player experience?

Not sure I follow this line of reasoning. Maybe DICE and EA did, but maybe that's part of the reason why the game tanked.

 

Yes? Dice has been releasing mediocre SP campaigns for their MP focused games since Battlefield 3. Battlefront 1 was actually an oddity with it not having a SP campaign.
The game "tanked" because of the micro transaction mess. It had nothing to do with SP at all.

 

Yeah, I could see one could be thinking "frostbite would have the tools we usually use to make uncharted" for a person who for years exclusively use one engine.

I only dabble in game development very rarely but years ago I almost exclusively use CryEngine.

When I want to expand my horizon to "unreal" and "unity" due to tons of praises of how awesome they are, how easy to use they are...

I kind of shocked when I couldn't even generate/create/modify a terrain/level with ease like in CryEngine. I thought something like that would be a standard feature / tool. Dunno if they know have those tools... Nowadays I just use Construct 2.

---

Anyhow, the weird thing is that.. Those uncharted star wars team should already realize that early in preproduction. Heck, anthem team should already know and well understand the deficit of Frostbite due to their past experiences....

 

https://wccftech.com/ea-confirms-st...ine-4-is-confident-the-wait-will-be-worth-it/

Respawn is using UE4 for their Star Wars game. Interesting. Maybe the technical debt in Frostbite is too great at this point.

 

Nope. The reason is simple: The project started end of 2015/beginning of 2016 a full year & a half before EA acquired Respawn (Nov 2017).

 

Makes sense.

IIRC, the Titanfall games pretty much only used Source because it was what they used to prototype early on, but then indecision, delays & other studio issues meant that they had to stick through with what they already had. Unreal Engine was one of the other engines in consideration at the time.

I suppose if they were working to pitch the SW title to EA, they needed something to prototype again, and would not have had access to Frostbite pre-studio-buyout.

 

Days Gone, releasing in just under 2 weeks on PS4, aslo uses a "modified version of the Unreal 4 Engine". In the thread for that game I linked to a PlayStation Access six minute '21 thing you didn't know' video with one of the 21 things was "the engine does not have inbuilt support for two-wheeled vehicles. [...] Those extra wheels are still there, they are just invisible, turned horizontal and with all physics and collision detection turned off".

It's also a game with crafting and inventory management - some of the same challenges facing EA's teams. In seems some external teams are coping with UE4's deficiencies better than EA's internal teams, even when core engine doesn't support the physics for the type of vehicle that's your primary means of getting about in a massive openworld.

 

I've been reading up about networking to try and get peer-2-peer across the internet working. IPv4 was developed by DARPA for a test system, without it ever being intended to become a standard. But, become a standard it did, leading to a need for NAT and then hacks to talk through those undocumented NATs, leading to all my woes.

People need to be wary of prototypes. They have a tendency to persist and become final products without the 'scrap it and implement a better system based on what we learnt' step manifesting.

 

It's still an important mark a totally unremarkable event that EA is letting an internal studio release a game without frostbite. Perhaps it's a It shouldn't be any sign of change.

From Jason Schreier's post mortem on Anthem, it's clear that when EA decided to unify all their games on frostbite, they didn't scale up the engine's support team accordingly.
Internal dev teams need to wait for weeks before they get a response from support tickets, because the support team basically works full time for the games that make more money (i. e. FIFA & friends).
That's probably why so many devs end up making their own tools (which apparently go unused after the projects are axed), because they get tired of twiddling their thumbs waiting for support.
UE4 and Unity definitely have a much stronger and larger support team because that's their core business (unlike DICE's).

 

What else are they supposed to do? Cancel a nearly finished game on a super popular brand because they are discriminating against other game engines?

 

What?

I don't understand your point. No one is suggesting EA should cancel anything. The decision to greenlight the game with non-frostbite happened when they accepted Respawn's pitch years ago, and the game definitely wasn't nearly finished then.

 

??
The game was never ever going to use Frostbite anyway. Respawn was contracted as an external studio to make this game. Frostbite is exclusive to EA's own studios and they sure as hell weren't going to make them switch engine 1 year & half later when they acquired them.

 

Did I suggest that?