I’m wondering what the future of idtech engine will be under MS.
As it stands, idtech 7 uses Vulkan API, and not DX12. Id has credited Vulkan as the reason why idtech 7 is so good.
Will MS allow the engine to stay with Vulkan as the API, or will they make it use DX12?
Will MS go all in and try and create a super engine taking the best from Idtech, Forza and Slipspace for instance, and then try and go after Unreal engine as a high end third party engine?
Will MS just promote idtech and push it out as an alternative to Unreal?
I firmly believe engine tech will be as important to game graphics this gen as the hardware.
We have seen what Unreal engine 5 looks like, Cryengine has shown Ray Tracing without the need for hardware, and MS has more ability than both Epic and Crytek to push engine tech forward.
What part will idtech play?
I am curious if dx12 offers any notable benefits over vulkan. They would need to maintain a cross platform path for the third parties that are using it on non-MS platforms.
id will exist underneath Bethesda, which isn’t immediately being killed as a brand and publisher as far as we know, so this is more likely a Bethesda decision than a top level Microsoft one.
Bethesda’s studio tech is a bit messy, but I suspect it will start turning into something a bit more standardised over time:
id develops and maintains idTech, the latest version being 7
MachineGames closely collaborates with id and uses whatever the cutting edge idTech is. For Wolf 2 and Youngbloods, it was a “halfway between 6 and 7” sort of engine, although they called it 6 still.
Arkane’s Lyon’s (non collaborative) games use an increasingly heavily modified fork of idTech 5 known as the “Void Engine”.
Arkane Austin has an unknown project in development, but job postings have requested “experience with Unreal Engine 4”. They previously used CryEngine for Prey.
Tango Gameworks just transitioned from using their own custom idTech 5 fork (“STEM engine”) to Unreal Engine 4 for Ghostwire Tokyo.
Bethesda Game Studios uses their custom great great grandfork of Gamebryo. I think Creation Engine is what they most recently called it.
Zenimax Online uses some kind of mutant derivation of the Hero Engine that they forked or something. It’s really unclear. They are apparently further developing new in house stuff to use for their new game.
Dropping CryEngine in favour of UE4 was a good decision by Arkane Austin. Dropping “STEM Engine” in favour of UE4 was a good decision by Tango. For the Japanese industry that’s doubly true, because UE has a huge number of people using it there and support for it is great.
Arkane Lyon should probably pick a lane and either go UE or more closely integrate with id and MachineGames to go for the latest iteration of idTech instead of continuing to work with Void Engine. They all work on First Person games in environments of relatively limited scope, so I don’t think it’s some kind of horrible fit being forced onto them or anything.
I think they will, yes. At least for the forseeable future. All it ever takes to turn things around is a change in management, naturally, but I don’t think upper management really cares on that level. Vulkan still runs on all their platforms, the games are already built with it, it’s more work to change than to keep it for no extra benefit since these games won’t be on non-MS platforms even if they are on a technology which could run on other platforms.
They will not do these things. Too much technological risk on the former, too unlikely to succeed on the latter 2.
A move to engine licencing would require a major increase in id’s staff count to facilitate, and probably a refactoring of the engine to be more general purpose (unless the marketing pitch was “We’re great at doing FPS, please use us for your FPS”). UE and to a lesser extent Unity are already really great generalist engines with huge market share and talent pools of people you can hire that are ready and raring to go with experience in it. The right time to make a viable competitor was probably 10 years ago, before UE went free and half the world started switching their games to it.
I think the odd engine out in Microsoft’s existing tech stack is Slipspace. For ForzaTech you can see it’s historically been a well oiled machine, specializing in doing amazing graphics and high performance for racing games. For Unreal Engine, it’s extremely flexible and there’s so much talent globally that can make use of it. For idTech 7, it’s at least got 2.5 studios using it, and it’s got a proven track record of delivering. Bethesda’s use of “Creation Engine” is questionable but the games sell 50 zillion copies and they’re very dedicated to cheese-wheel-persistence-technology or whatever so I guess they’ll probably keep on keeping on.
I think there’s nothing really in old style Halo that couldn’t have been done on idTech 7 or modern iterations of Unreal Engine. Even Halo Infinite with it’s new approach as an open world game could have been built on UE4 / 5 / whatever. It’s not like UE can’t do FPS games, it’s not like UE can’t do proper vehicle handling, it’s not like UE can’t do open world games, or has a horrible workflow or anything like that. If they are insistent on developing their own tech that’s fine, but I think internally they should be asking whether it wound up being really worthwhile to spend 5-6 years building the technology that is powering Halo Infinite, and whether in the future they should maintain their own separate tech or consider sun-setting it after this game and going for something else.
There’s no real reason to change it to DX12, other than to support DX12U features inherently for the PC and Xbox platforms. If Switch and PS are no longer targets, this might make sense.
The other thing is that MS could fork it and work on it themselves. They did the same with UE after Epic (Tim Sweeney, rather) didn’t want to support the Windows store and UWP apps.
I’m not sure if there were any plans to take id Tech and turn it into a game development engine for the masses a la Unity or UE. But it seems like if there’s any opportunity to do it, doing it with MS backing makes sense. And if that happens, DX12U support along with other targets makes sense.
Microsoft will not force any change of engine tech in Bethesdas games. They will support them as much as they can so that it runs great on PC and Xbox.
Unreal and Unity serve the market very well right now. I don’t know what opening up idTech to the broader market would accomplish? This would take a lot of ressources for documentation and international support away from Bethesda. And what would it mean for idTech to be a general purpose engine and not just great for inhouse FPS projects?
I think Id Tech 7 is fine the way it is and will be left alone. MS still contributed to an extent to the Vulkan API.
Id Tech 8 I could see a shift to DX12 with Id being able to contribute and customize it as desired.
As for licensing, I see Id Tech kept in house, or maybe folded into Game Stack with older versions. It’s a fantastic FPS engine but you run Frostbite Engine like risks forcing it to be something it isn’t.
DX12 Ultimate currently offers some notable improvements over Vulkan, such as better support for HW GPU scheduling, SFS, DXR 1.1 as well as DXR 1.1 API in-silicon acceleration on Xbox. Of course this comes with a catch - you actually have to use them to see any benefits. Other than that as both Vulkan and DX12 are derived from Mantle they are more or less interchangeable and it comes down to what the programmers prefer/are better at.
Lack of support for Ray Tracing on the same level as DXR 1.1 along with other features like Velocity Architecture and GPU Scheduling are some of the major drawbacks of Vulkan compared to DirectX 12U.
Vulkan released their extension for Hardware RT already (based on work with Nvidia but they also collabed with AMD behind the scenes too so it’s not like it won’t work). March this year I think.
Wolfenstein 2 (vulkan only) was the first PC game to release with variable rate shading, and Wolfenstein Youngbloods (also vulkan only) had support for that plus hardware accelerated ray traced reflections (Mar 2020).
I doubt these kinds of decisions are weighing on anybody in a big way. Even before this official support Nvidia had their own vkl extension for stuff like Quake 2 RTX in 2019. If there’s much in DX12u that’s missing, it’ll be included in Vulkan before the next major idTech engine revision either way. Wolf 3 wouldn’t be coming out before the second half of next year.
Simple as that. Doom Eternal was amazing, my favourite triple-A game of the year, but as the DLCs end I believe they can look at something new. Go back to Quake, work on a new IP, then reboot Doom again learning from the good and bad of the classics and the modern games alike.
They did but it does not have the same feature set as DXR 1.1 which has been built with dev feedback on DXR(I have edited my post to clarify what I meant). For example giving ray tracing tasks directly to GPU without using the CPU or inline ray tracing.
This is a new MS. They have embraced opensource. They even have a dedicated Linux team now. I doubt MS will force id to any technology they are not comfortable with.
Correct, Vulkan RT officially released this March. However note that doesn’t mean Vulkan RT is equivalent to DXR 1.1. It’s more so in line with DXR 1.0 cause Khronos Group has been dragging their feet with Vulkan RT
Maybe? It’s possible id Software is porting Doom Eternal/id Tech 7 to DX12 for the PC gamepass release (would explain why it is releasing after the console version) as I am not sure if Vulkan is supported on the windows store UWP which is how gamepass PC games are delivered. Once Doom Eternal hits PC gamepass I can check what API it uses with MSI Afterburner/Rivatuner Statistics Server if you’re interested (currently the Steam release is Vulkan powered).
