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

I suspect Management had not adjusted the schedules for a team to be able to bring up the tools needed *before* everyone else begins to work on the game. That would have saved plentiful work-hours that didnt need to be spent on the project.

Instead, I suspect they started everyone working on it, so you had multiple other teams cobbling together less than desirable work artifacts using what was available. Then when the tools and solutions they needed were done months later, if not a year later, they have a crisis option to make: redo the work done with the new tools and better solutions, or carry on because of asinine project milestone deadlines picked out of thin air without support from reality.

One gives a far better result likely with less total work-hours but at a later date. The other gives the suboptimal results we have seen.
 
I agree that time rules all, and I have little doubts that they are capable. But time is the largest constraint they are working against, and time is money.
The problem is that the time that should/could/would/ have to be spent on modifying the engine wasn't there because of non engine related development shenigans especially at Bioware (& also Visceral). Every game developer has to modify the engine to suite its needs (Frostbite literally powers sports games, FPS, & open word racing games with no issues at all). Army of Two: The Devil's Cartel which is a 3rd person action cover-shooter was developed by Visceral using Frostbite more than 6 years ago. For example, the use of Frostbite in Andromeda isn't the reason why the game shipped with shitty dead eyes on the model (no AO contribution from the eye lids,quickly fixed in the first patch post release) or shitty animation...the game simply wasn't finished after being rebooted again a little more than a year before release... Bioware started using Frostbite internally more than 8 years ago...it's enough time to ship a perfectly working AAA game on any engine.
 
A 3rd person adventure is a completely different beast from a football or racing game. There's hardly any interaction with the environment in those (open doors, pick up objects, place traps, etc.).
Dragon Age Inquisition has very little of those and still Bioware complained a lot.

Ragtag IIRC was going to be very dependent on gadgets for its gameplay, so Visceral's job might have been particularly hard on that front.
Hennig claims the game was pretty advanced in its development. Makes us wonder if EA couldn't have just picked her up again and finished what was started.
 
Bioware started using Frostbite internally more than 8 years ago...it's enough time to ship a perfectly working AAA game on any engine.
I agree with that actually. Which is why I asked whether some studios were better suited for tools and engine modification better than others.

I’m not exactly understanding the issue unless of course it just comes down to leadership and management. Which is perfectly fine as an answer for me.
 
Dragon Age Inquisition has very little of those and still Bioware complained a lot.
All the BioWare titles with frostbyte has had a few glaring short cuts.
I don’t know if you recall how they did horses. Like as soon as you mounted everyone merged with you. And as soon as you dismounted everyone exploded out.

Animations were largely limited. Modifying the environment also limited compared to what we can see being done on battlefield respectively.

And load screens. Everywhere.
 
In that case I think Andromeda is. Bigger maps, denser details, higher polygon count, more AI on screen .. etc.

I don't agree. In theory yes it should be obvious, given priorities... but in reality IMO, I don't see how andromeda assests are better than battlefield/front ( well maybe here and there) or have really bigger maps, more/faster action with destruction and less opportunities for frustum culling.etc( sorry for simplifications). All things equaled, I see nothing to justify this huge gulf in performance, barrely 30 fps vs holding 60 in single and 24 maps it can be close to 3 times of performance. My point, original creators still know better how to extract renderer/engine.
 
The problem with these statements is that no one that is still working there is usually allowed to confirm or deny them and companies prefer to let these statetements be forgotten.

Has anybody checked any confirmation on Glassdoor reviews?
 
BioWare brought an outside help to aid with the facial animation, they brought in a company with a solid background in facial animations called Snappers (they worked on Injustice 2, Infinite Warefare, Mafia 3, all games with excellent facial features), and they couldn't help with Andromeda still,

but there were lingering questions about how to implement those animations into the engine and how to scale across the whole game
https://www.kotaku.com.au/2017/12/t...ct-andromedas-troubled-five-year-development/
 
Dragon Age
Andromeda
StarWars
C&C Generals 2
Anthem

That's too much scapegoating going on, and from game developers no less. Even Battlefield still has mabor engine flaws that never seem to get fixed despite multiple game iterations. Not to mention the awful DX12 problems that never seem to get fixed, for comparison the much smaller Metro team made a capable DX12 path, yet DICE never managed this feet.

Something that might be a little more interesting is comparing and contrasting the various games made or attempted by other studios than DICE.

List of games for Frostbite 3.0

Bioware
  • DA: I - meh. But at least it works and was relatively trouble free.
  • ME: A - a mess, technically and from a gameplay standpoint.
  • Anthem - good graphics, but a mess WRT design and implementation.
  • Shadow Realms - Cancelled
Victory Games
  • C&C Generals 2 - Cancelled, an anonymous person who claims to be part of development claimed it was due to politics and meddling on the part of EA.
Visceral Games
  • Battlefield Hardline - basically just another Battlefield in a different environment
  • Star Wars - cancelled despite this company having extensive experience with the engine from making Battlefield Hardline.
EA Vancouver
  • FIFA 17, 18, 19 - graphically relatively simplistic games. Static arenas with limited texture variation, etc. About as easy as it gets to graphically enhance.
EA Tiburon
  • Madden NFL 18, 19 - same situation as FIFA. Relatively simplistic from a graphical perspective compared to even corridor games much less open world games.
  • Rory McIlroy PGA Tour - I didn't even know this game existed. :p
EA Spearhead, Ghost Games, EA Black Box
  • NFS games. I don't do racing much anymore so don't know much about these. I know a few of these had issues on PC. But graphically I think they were fine. Still racing games are generally as linear as you get with some exceptions. I don't know if any of these were open world or semi-open world?
It's interesting to see that the best results with Frostbite by studios other than DICE are the rather more simplistic games graphically and games that are basically only doing what Battlefield already does.

Once teams start doing more complex rendering loads (open world, for example) it appears they start to run into trouble.

This could quite likely be more about not having the tools available to do what they want. If production is rushed, it's quite likely that time wasn't allocated to develop the tools for the job prior to work beginning on a project. Trying to create the levels, implement systems, etc. and develop the tools to do those things simultaneously just seems like a recipe for failure unless your game isn't attempting to do much (NFL, Madden, NFS for example).

It isn't like Bioware doesn't have (or didn't have) people with extensive experience in the types of games they were crafting. I may have issues with the direction the company took with their games, but they had competent people there.

Likewise, Visceral games has the experience and the talent for large story based games with complex systems.

Meh, too many things in play, IMO, to blame the developers over potential issues with the Engine...or more likely the tools available for the engine and the timeframe within which EA expects the game to be developed.

While I like what DICE has been able to do with the Frostbite engine, I think it's unfortunate that the engine was forced on the developers rather than the developers being able to choose the engine best suited to their project.

Regards,
SB
 
While I like what DICE has been able to do with the Frostbite engine, I think it's unfortunate that the engine was forced on the developers rather than the developers being able to choose the engine best suited to their project.

Regards,
SB

How do you know that was not the case?
 
I think part of the problem with EA's studios use of Frostbite is that game developers do not seem to be used to collaborate on and/or use the development/HEAD version of an engine. Game developers have historically forked off an engine and then made the game on top of that. If you are going to collaborate on the actual engine with the upstream developers you really need a different mindset. Instead of forking you can maintain a patch set for example.
 
How do you know that was not the case?

I just remember a couple years back it was widely reported that EA was moving to have all their studios use the Frostbite engine. Perhaps that fell through and they backed away from it?

I believe all of the major studios that EA owns outright are currently using Frostbite and that the only major studios that are published through EA that aren't using Frostbite are ones that aren't owned by them (Respawn, for example).

Regards,
SB
 
I agree that time rules all, and I have little doubts that they are capable. But time is the largest constraint they are working against, and time is money.

I don’t think Amy is scapegoating frostbyte, at least not intentionally, but the argument she made without getting lost in the details is that to make the game they needed, they needed more time and money. If frostbyte already shipped with a lot of those items, they would have needed less
Time and thus less money.

If you want to make a game that is like the Uncharted of Star Wars, you're up against Probably 11+ years of tools that Naughty Dog created for the Uncharted series. I honestly don't know how long it would take to fill the gaps in Frostbite to make a cinematic 3rd-person adventure game, but I imagine EA just wasn't willing to wait for as long as it was taking.
 
I think part of why this discussion is going nowhere is that people have different views what an engine really is and what it should bring to the table.

When most people here talk about a game engine they think along the lines of general purpose game engines like Unreal or Unity, i.e. a framework or a container that provides a means to write a game and provides the critical infrastructure like a rendering pipeline and a main loop along with components/middleware and tools that you may or may not use (like physics and animation) and a means to extend the engine with components/middleware and tooling But the engine usually does not contain any game logic.

And I think what Amy Hennig refers to when she is using the word engine she does actually mean an engine made for a specific type of games that already contains (customizable) game logic, more a skeletal game than a general purpose engine.

But Frostbite seems to be more like Unreal or Unity. You probably can do about anything with it but you there is a lot of stuff you don't get out of the box.
 
If you want to make a game that is like the Uncharted of Star Wars, you're up against Probably 11+ years of tools that Naughty Dog created for the Uncharted series. I honestly don't know how long it would take to fill the gaps in Frostbite to make a cinematic 3rd-person adventure game, but I imagine EA just wasn't willing to wait for as long as it was taking.

Henning specifically mentioned Uncharted 1, which Naughty Dog had all new tools for (since they throw away everything LISP/GOAL-based that they had made).
 
I don't agree. In theory yes it should be obvious, given priorities... but in reality IMO,
Character polygon count alone is significantly higher in Adnoremda than Battlefield, same for the environment. Not to mention the much more huge draw distance and denser details scattered around the map. Battlefield looks barren in comparison.
 
Character polygon count alone is significantly higher in Adnoremda than Battlefield, same for the environment. Not to mention the much more huge draw distance and denser details scattered around the map. Battlefield looks barren in comparison.

In case I was imaging things, looked up videos with somehow similar environments and now i am even more convinced.


I just don't see that assets quality/density and draw distances explaining andromeda performance. Then let's imagine them in third person on bf side which would make them apear even better. Looks pretty comparable to me. Sure few key models might look good in cutscenes but bf has that too in single player. I would even count, better lighting, AO, tessalation, particles for battlefield side and then there is destruction which if used would probably trigger flat... 25 in andromeda vs quite steady 60.

Something like anthem reveal would be just result for 30fps frostbite if BF isdoing all this at 60,but again even good engine is not magical switch for everyone.
 
"
  • ME: A - a mess, technically and from a gameplay standpoint."
To be fair, I started the game after all the patch, and It was pretty polished then. In that case, I believe it was rushed, the devs just needed more time.
 
Back
Top