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

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.

I believe they had insomniac'' engine but due to license reasons had zero starting assets so even prototyping was slow.

Whilst trying to define what the game should be like a dev made a prototype in source as he was familiar with it and progress was much much faster. The team liked it when they saw it on the Monday and this made them re-evaluate their current engine choice.

I think time frame mostly led the final decision, most of the team was familiar with it and it was proven on last gen consoles to deliver 60fps with low input latency. I think insomniac's engineengin by the wayside due to the extra time it would have taken to use. Unreal was not proven on PS3 at 60 I don't think .

There is a posr mortem of the whole dev process that details engine choice and shows early prototypes. It also goes into the Xbox One and how it ended up an exclusive.

http://finalhoursoftitanfall.com

It's a paid app but it's good content from what I recall.
 
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.
X25 has nat :)

and everyone hates ipv6.

One big upside of nat is smaller ribs/fibs
 
Did I suggest that?
It's certainly implied. :-? At no point does it seem logical that J:FO could have ever used Frostbite, so how is it 'important that EA are letting an internal studio release a game not on Frostbite'?

What's the 'important' part? That EA didn't force Respawn, an external studio, to use Frostbite, the internal engine, when EA agreed to publish the game? Or that when EA acquired Respawn, they didn't force the game to swap to Frostbite? Or something else I'm completely missing?
 
What's the 'important' part? That EA didn't force Respawn, an external studio, to use Frostbite, the internal engine, when EA agreed to publish the game? Or that when EA acquired Respawn, they didn't force the game to swap to Frostbite? Or something else I'm completely missing?
The important part is that EA didn't force Respawn to change engines during the acquisition talks.
With TitanFall 2 releasing late 2016 and Apex Legends releasing early 2019, I doubt the Star Wars project had gone past pre-production in late 2017.

I didn't suggest EA should have done it, but I'd say one of the the main reasons they didn't is because back then they already had Disney breathing down their neck for all the Star Wars games they kept cancelling, so they needed to release the game ASAP.
 
I believe they had insomniac'' engine but due to license reasons had zero starting assets so even prototyping was slow.

Whilst trying to define what the game should be like a dev made a prototype in source as he was familiar with it and progress was much much faster. The team liked it when they saw it on the Monday and this made them re-evaluate their current engine choice.

I think time frame mostly led the final decision, most of the team was familiar with it and it was proven on last gen consoles to deliver 60fps with low input latency. I think insomniac's engineengin by the wayside due to the extra time it would have taken to use. Unreal was not proven on PS3 at 60 I don't think .

There is a posr mortem of the whole dev process that details engine choice and shows early prototypes. It also goes into the Xbox One and how it ended up an exclusive.

http://finalhoursoftitanfall.com

It's a paid app but it's good content from what I recall.

I forgot I had that. xD

Side-note: the Bluepoint folks also just recently talked @ GDC a bit about what they did to get the first game running on 360 too. Something something something memory. :V
 
The important part is that EA didn't force Respawn to change engines during the acquisition talks.
Why's that important? What does it show?

The discussion around FB has thus far been, "it's crap and EA shouldn't force studios to use it." J:FO could be taken as evidence of a change in EA's attitude, allowing a studio to not use FB, except that it'd be business suicide to force the swap, incurring huge additional costs without any gains. So, AFAICS, it's not indicative of anything. It'll only be interesting if Respawn are allowed to continue to use other engines for future games.
 
The discussion around FB has thus far been, "it's crap and EA shouldn't force studios to use it."
Not from me it isn't. I don't even recall using such a poor description, as I haven't seen such statements coming from developers either.
I just wrote a couple of posts ago about why so many problems with Frostbite persist across studios, and it's not because it's crap.

J:FO could be taken as evidence of a change in EA's attitude, allowing a studio to not use FB, except that it'd be business suicide to force the swap, incurring huge additional costs without any gains.
It would be suicide because EA's handling of the franchise has been critically panned and they need to release Star Wars games pronto. Had EA picked up production of 1313 and released it in ~2014, released RagTag/Yuma in 2017/18, plus a KOTOR 3 that apparently started getting several pitches and prototypes from BioWare since the exclusivity deal in 2013, and Battlefront 2 hadn't released as a pay2win casino game for kids, then perhaps they would be in a position strong enough to demand an engine change from Respawn that would delay the game a year.

There are gains in standardizing Frostbite across the games from a major publisher, so it definitely wouldn't be "without any gains".
 
Indeed, what would be the gains for EA, Respawn, and J:FO swapping from UE4 to Frostbite? As i see it, at best you end up with basically the same game and same product, but having spent months translating it.
 
I suppose the main one is not paying Epic’s licensing terms and some cut of the profits, although I’m not sure what the latest conditions are these days. Although as implied, there would be some non-trivial cost getting back up to speed on the programming side while the content pipeline may suffer additional delays.

In the very long run, it might be a forward looking amortized cost should they decide to switch anyway, but... we’d need to see the receipts. On the other hand, we are looking at a Star Wars branded product, so the bean counters could potentially make dubious cost savings decisions in the future that end up costing more elsewhere.

/CynicAl

¯\_(ツ)_/¯


--------

That said, we already have seen a couple Star Wars games on Frostbite, and then there's Mirror's Edge (perhaps relevant for Respawn's shtick with enchanted player movement). The main issue is retraining the employees for Frostbite's tool chain :?: I do wonder if it's something they're debating for an actual Titanfall sequel - not sure how big Respawn is now.


EA would not have to pay any licensing costs for UE4.
Beaten because I was Bing'ing while squatting. >_>

ahem. /shakes fist @Shifty Geezer
 
Last edited:
What are these gains you speak of?
Indeed, what would be the gains for EA, Respawn, and J:FO swapping from UE4 to Frostbite?

Is this really the question that's being made? What are the gains for EA putting all their dev teams working on Frostbite?

Andrew Wilson said:
On Frostbite, again, we have a very large team across a number of different locations who are constantly and continually investing and innovating in that engine in terms of visual fidelity, in terms of physics, in terms of animation, in terms of AI. Every aspect that a game team needs to develop the most innovative and creative experiences.

And when you look at what we’ve been able to do with Real Player technology in FIFA this year, when you look at what we’ve been able to do with Madden visual fidelity and Madden cinematics and Longshot or story mode, when you look at the expanding dynamic worlds that we’re showing in Star Wars Battlefront 2, that is really demonstrative of the level of innovation and advancement we continue to make in that engine.

And at some level, because we have a dozen teams building on top of the platform and a core team feeding all of that development back into the mainline of code, that engine is actually making leaps and bounds faster than many engines would, which is one or two games working on them.

And so we are very, very happy with our core strategy of moving to a single engine. We’re starting to see real leverage in the engine, not just in terms of our ability to build wonderfully large and creative and entertaining new games efficiently, but because we’re able to do things that are far more innovative than things we’re seeing anywhere else in the industry.
 
Did you read the statement from EA's CEO regarding the reasons they're pushing Frostbite publisher-wide? It's right there in my post. Only thing missing is of course the fact that they don't need to pay for UE4 royalties.
 
It remains to be seen if any of that is even true, given the statements against the engine. Those CEO statements are about strategy and don't provide evidence of actual benefits. In particular all the work put in by the FIFA team wasn't a benefit to the Anthem team, it was a huge detriment since it caused engineering resource starvation.
 
The gains are diminished or even nullified by the fact that EA didn't scale up the support team for the engine. I wrote exactly that too.

The idea that frostbite's development and implementation could be a product of contributions from the several dev houses that belong to EA is a good one. Especially considering the fact that frostbite is arguably the multiplatform engine that provides the best visuals on lower end hardware (most importantly consoles).

Sum that to the fact that they'd probably save a lot of money on royalties and tech support and the concept is sound.
The process was just severely mismanaged.
 
Is this really the question that's being made? What are the gains for EA putting all their dev teams working on Frostbite?
That's creating games on Frostbite, not swapping engines. If Respawn were to swap from UE4 to FB, what are the gains for the game? Or for Respawn?

Not to mention that PR piece is untrue - Unreal Engine has far more than 'one or two' games being made on it. ;) This whole thread argues some of the downsides with forcing a one-engine-fits-all solution on devs.

The idea that frostbite's development and implementation could be a product of contributions from the several dev houses that belong to EA is a good one.
Hmmm. It sounds like one of those things that's good on paper but doesn't necessarily work in the real world when different studios have different ideas on how things should operate. If you try to force everyone to do things The Proper Way, you'll generate friction, and if you try to create an engine so wide and flexible that everyone is supported in their preferred manner, you create a whole load of work for yourself.

Other engines develop and grow beyond just the people making it in the same way, as dev conferences and best practices get shared among devs and find their ways back to the engines. So it's not like Unreal Engine is the product of the 1000 minds at Epic, while Frostbite is the product of the 9000 minds at EA, but both are the products (we hope) of the tens/hundreds of thousands of devs and software engineers doing work and sharing knowledge and coming up with solutions.

The one clear advantage I see for Frostbite is less legacy to worry about. It was a clean slate design in 2008 so was founded on the ideas and future expectations from then, whereas Unreal 4 had to be someways backwards compatible so people didn't have to rewrite their entire games and/or learn a whole new engine when moving from UE3.
 
Back
Top