Amy Hennig on Frostbite challenges in Visceral's Star Wars production

[NXGamer]

Any chance you can put in a link to the original interview, please? https://www.usgamer.net/articles/amy-hennig-interview-uncharted-4-leaving-ea-ragtag-star-wars

Click to shrink...

Good interview, thanks.

I do not think Amy is saying what everyone is reading here. An engine, or Repository is only a collection of code that is as good as the tools it comes with or the programmers to deliver what you need. Frostbite is designed to deliver a specific set of requirements, mechanics etc and 3rd person, high level animation systems, rigging, collision points, horizontal and vertical traversal etc etc is not it. This means months maybe a year or so of work has to happen before you have a Vslice that you can show to any execs or public. A full 12 months of cost and essentially R&D means that when a milestone is up and you, in essence, have a Small arena, animation rig, combat and maybe some cinematics this makes them nervous to invest a further 3 years and 3x the cost. Bigger team size, ramp up into full productions stage, contractors, marketing etc etc) then the plug gets pulled as the RoI does not fit into their financial requirements/predictions.

The point she is making (I think) is that this prolonged prototyping phase put the workstream at risk as return was slower than If they had an engine that already had some of this functionally within it. EA are one of the worst examples of VC's in games, hence the cutthroat style.

All engines will have issues, quirks, strengths and weaknesses and I am sure Frostbite is no different.

 

[Falchion]

Yeah I definitely wish we could've seen what this was looking like.

 

[sandboxgod]

I just recently finished SW:BF2 Single player story (just the base). It's not bad but not the greatest story ever told either.

But I can only imagine Amy Herning's Uncharted SW game would've been amazing and it really sucks we'll never get to see it. I can definitely see how her game was completely different from Frostbite's normal type of game. Just feels like only DICE can make this engine behave proper is what I am gathering from these articles.

 

[Forsaken82]

So what was it about a 3rd person SP game not fitting with EA's business plan? I mean they green-lit it, so it shouldn't have been a surprise that there was limited opportunity for microtransactions.

Click to shrink...

There are a ton of factors that come into play here. It could have had nothing to do with "well we don't want a TP Single Player game" and everything to do with mismanagement. Failure to deliver on milestones. If Frostbite is that big of a struggle to develop for, the milestone failures are a huge plausibility and EA just said fuck it.

 

[Aztechnology]

I get that not licensing an engine saves them money. But from a holistic view does it really? With all the issues and drawbacks that we see? It's hard to quantify things that aren't directly sales related like developer satisfaction and maintenance time etc. (I mean you can, but I doubt they do).

It just seems like it has to be worth paying the fee in these cases. I mean engines like Unity and Unreal solve a lot of headaches for you and are quite adapatable comparatively to what these engines sound like. I just don't get it.

 

[Spartancarver]

Unsurprising.

DICE themselves can't even work with the engine properly, as evidenced by how insanely buggy and unstable BF V and Star Wars Battlefront 2 are. Not to mention the fiasco of Battlefield 4 launch.

 

[XaviConcept]

I think the number 1 thing I would love for fans and even the non-developer side of the industry would be to really understand the purpose of engines and what they do, its probably where the most misconceptions lay.

I am not surprised at all that Frostbite would really hamper their progress, there would be so much shit they would have to start from scratch.

 

[Aztechnology]

I think the number 1 thing I would love for fans and even the non-developer side of the industry would be to really understand the purpose of engines and what they do, its probably where the most misconceptions lay.

I am not surprised at all that Frostbite would really hamper their progress, there would be so much shit they would have to start from scratch.

Click to shrink...

Sure but. The point being, there's plenty of very good licenseable engines. And they still are designing plenty of elements from relative "Scratch". Sure scripts are object associable I'd assume etc. But the fundamental capabilities are fairly limited I'd imagine (from what we hear) comparative to other developed engines. So the question is really, is it just licensing fees that it comes down to?

I've personally only worked with Unity and Unreal. As obviously it's not like EA's engines are available to hobbyists or Indie devs.

 

[Kinn]

So there IS a precedent on EA games using UE then. Fucking hell, why didn't they let Visceral use UE4 for Ragtag, dammit!?

Click to shrink...

I dont think Bioware was owned by EA when ME 1 was released.

 

[Asbsand]

the barrier numerous developers have faced is programming absolute bare bones, basic features required for their game into the engine.

Click to shrink...

For the record, when BioWare started working on Mass Effect 1, UE3 was barebones and basically the same, built for their upcoming game, and it could not even do certain basic mathematic equations, which BioWare actually implemented into the engine and has become what Epic has used for it since then, including UE4. Contrary to working with DICE who made Frostbite, it was also unusually hard (and still can be!) to get Epic Games' reps to explain how the engine worked, especially back then.

I don't doubt many aspects of Frostbite are unintuitive and possibly worse to work with across genres than the UE3 engine used to be, but APIs such as FaceFX (third party lip-sync and audio-generated face-gesture engine) which was used in Mass Effect 1 was also implemented into Frostbite when BioWare created Dragon Age Inquisition and worked pretty marvelously compared to later games for unknown reasons (My technical eye guesses it's the face meshes and rigging that causes incompatibility across games).

Point being that while Frostbite surely requires the developer to start from scratch with some basic features in a lot of these early non-Battlefield projects, so did Unreal 3 and I often question whether the difference is really that the engine is somehow worse or just that the general developer competence inside of EA's subsidiaries are up to the task.

Bear in mind some of the animators and cinedesigners that worked at BioWare back in 2007 are people who since moved on to studios like Naughty Dog and Quantic Dream and other places that are known for exceptional polish with their tools.

 

[Gloomz]

Have any of the fucking suits ever commented on how Frostbite has fucked almost all of the non-shooter games they've forced companies to make?

I get using an in-house engine saves money, yadda yadda.

How many projects is it going to take where the devs say 'Having to use Frostbite fucked us.' until EA considers other options?

 

[UnderNewOwnership]

Have any of the fucking suits ever commented on how Frostbite has fucked almost all of the non-shooter games they've forced companies to make?

I get using an in-house engine saves money, yadda yadda.

How many projects is it going to take where the devs say 'Having to use Frostbite fucked us.' until EA considers other options?

Click to shrink...

Well, ironic though it may sound, all the mess-ups caused by early implementations of the engine, and the workarounds/toolsets created to fix them, do make the engine easier to use in the future. Issues that appear partly because, from what I understand, EA has a relatively small specialised engine design team. Though this is also relatively understandable, as engine iteration consumes vast amounts of money, time and engineer-hours. It's why Unity Technologies and Epic are entire companies basically dedicated to continuously polishing a specific engine.
Moreover, large games tend to take a long time to make. The engine you start on has to be chosen out of what is available at the time on the market, not what the engine design team might come up with in 2 years' time. Seems like people forget that Unity was relatively widely criticised for animation stutter issues around 4 years ago (I think it was fixed in the meantime). As a game dev, you can't fully rely on the engine team's assurances that issues will be fixed in the future. What if they won't be? What if they are, but the fix point is beyond your game's shipping date?
OTOH, the whole "Bioware chose the engine" seems to me somewhat incomplete an explanation. I suspect the 'choice' was inherently biased. In that, they were given a choice between using Frostbite for free (as the in-house engine), or having to spend X amount of $ out of their already alotted budget to licence a different one. And extra budgeting provision for the outsourced engine was obviously off the table. They most likely decided to go with Frostbite and spend the money elsewhere.

 

[Duderino]

I assume all the loading times and many of the technical issues are problems that wouldn't exist if they were using UE4 or Source 2.

Click to shrink...

Hard to say, there's nothing quite like Anthem built with these engines available for comparison. No telling what kind of challenges Bioware would have faced with a different engine.

 

[Miscend]

Sounds like whinging to me. What's the job of programmers, isn't it to develop new tools for the artists and to implement new features in the engine?

Titanfall 1 and 2 are built on a modified Source Engine. Respawn customized the engine and built new features in accordance with their own requirements. I believe the COD games are built on a modified Quake 3 engine. The Arkham games are built with UE3 but use an in house developed renderer. All these other teams have had to customise engines to meet the requirements of AAA game development.

It should be less effort use something like Fostbite as a starting point for development, than it is to write a whole new game engine from scratch. Apart from BioWare many of the Frostbite based games we've seen this generation run at 60fps and have good performance. We've seen loads of UE4 titles with poor performance especially on base consoles like Xbox one.

 

[RedSwirl]

Been thinking back to how other big publishers these days handle engine tech, and EA seems to be just about the only one with this issue -- where all its studios are taking one engine that was apparently originally made for one type of game and trying to apply it to different genres.

• Big Japanese publishers like Square Enix and Namco just use Unreal now.
• Bethesda has a lot of its studios use idTech or forks of it, but almost all its games are first person. It lets Bethesda Game Studios keep using its own tech (GameBryo forks right?), let RAGE 2 use Avalanche's engine, and let Arkane use CryEngine for Prey. The only real outlier there is Evil Within -- a third person game, using idTech.
• Ubisoft I think is still basically using different engines for each of its franchises: Creed and Wildlands are Anvil, Division is Snowdrop, Far Cry is Dunia (CryEngine 1 fork), Splinter Cell was on an Unreal 2.5 fork, etc. They probably all share tech where they need to though. The only outlier there is R6 Siege being Anvil, which mostly seems to be used for Ubisoft's third person open-world games.
• Activision lets each COD studio run its own fork of Quake III.
• With Sony, it looks like Guerrilla is sort of their "DICE" but to a lesser degree. Other studios have used Guerrilla's Decima engine but I think they mostly just share stuff from it with other Sony teams. They let Bend use Unreal 4 for Days Gone.

 

[Kumomeme]

I wonder if similliar situation happened to square games with crystal tool engine before..before they claim had difficulty to render town in ff13 that lead no town existed in the game..one of main problem of ff14 first launch failure also due to the engine...i curious if similliar situation happened to versus before since the game also originally developed using same engine

there quote from toriyama and ahikio maeda

"Another issue was the universal engine. Because we were so focused on creating an engine for next-gen hardware that could be utilized across all platforms, we made the mistake of trying to accommodate every single project that was in progress at the time."

Click to shrink...

and now,currently we see the studio under SE utilizing different engine..ff14 had it own engine,UE4 and Luminous studio, like how people here pointing out that sony first party studio use their own different engine to suite their need

Well,different situation happened at capcom with mt framework engine before...and now they had RE engine

 

[Deleted member 9317]

It's safe to assume that Frostbite is a huge problem for EA. I'm glad Amy spoke against it and I hope people defending the engine realize that Amy and the developers have no reason to lie about the engine vows.

Even Anthem failed heavily partly due to Frostbite engine: https://www.resetera.com/threads/ko...hem-went-wrong-jason-schreier-article.109134/

"Frostbite is full of razor blades," one former BioWare employee told me a few weeks ago, aptly summing up the feelings of perhaps hundreds of game developers who have worked at Electronic Arts over the past few years.

"Frostbite is like an in-house engine with all the problems that entails—it's poorly documented, hacked together, and so on—with all the problems of an externally sourced engine," said one former BioWare employee. "Nobody you actually work with designed it, so you don't know why this thing works the way it does, why this is named the way it is."

Is it safe to assume that Frostbite needs to evolve/improve and that resources need to be better shared? Unreal Engine 4 is masterclass, maybe it's time to look at what they're doing right and implement it.

Maybe put Frostbite in the back burner...

 

[Embedded]

So EA can't afford to have a team constantly work on their engine to keep it up to date and add more functionalities?
The shareholders must be so detached from gaming...

 

[GrrImAFridge]

There's merit to the idea of mandating the use of an in-house engine -- it's a higher initial cost that, at least in theory, saves money long-term as you're not paying licensing fees or royalties -- but when that requirement spans multiple studios across as many countries, support becomes extremely important. Ubisoft appears to have figured this out as it juggles no less than three internal engines, seemingly with few issues.