Microsoft and Game Engines *branch*

[liams]

What I mean that no matter how they overhaul their engine it always shows its age. Though the ability to pick every single thing will never get old.

I do wonder why TES is so buggy each release. It is literally the game that is designed to be finished by the community

true, you can always tell its a bethesda game from its engine

the problem is not that its base is old, its that they dont overhaul it enough. Look at unreal, version 1 came out in 1998, and you could say each subsequent release has just been an overhaul of the previous version with some added features, but no one complains about that

 

[ChuckeRearmed]

the problem is not that its base is old, its that they dont overhaul it enough. Look at unreal, version 1 came out in 1998, and you could say each subsequent release has just been an overhaul of the previous version with some added features, but no one complains about that

Does the current Unreal 5 has any connection to the first unreal? I thought most of its codebase is completely unrelated to the first one.

 

[fehu]

An engine is not just rendering but all the tools and the interoperability with 3rd party tools.
For example, I remember a dev commenting on the engine that they were forced to use, and how much slowed them down with little things like the ineffective search engine for the shaders, so instead of reutilizing one they add to write it again.

Microsoft is directed toward a future autarchic ecosystem, where they have the platforms, the infrastructure, the store, the talents, and the ips.
What they lack is just a resident engine, to acquire a little independence from Sony's participated Unreal, and leverage on Series' features and quirks.

On the pros: share code and assets between internal studios, and speed up development.
On the cons: anyone uses UE, developers, artists, toolchain... To switch they must have something competitive and flexible.
Enter ID tech, the engine known as doom engine because it's been used for.no driving games, no fighting, no lot of things.

 

[cwjs]

I don't see it happening right away. The big multi purpose engine companies (epic and unity) succeed for a lot of reasons that are hard to replicate. These companies have big staffs of technical writers, designers, engineers, technical artists, etc, building on decades of solving problems actual users have, being able to fit in to any company's workflow, investing in training/education pipelines so new hires can join companies already able to work with them, making tradeoffs that work for a wide variety of games, and so much more. Bootstrapping an internal engine up to that level is expensive and difficult -- and microsoft already owns a lot of studios with good internal tech, so there's no EA-esque reason to invest all that money just to save a few fees.

But on the other hand, microsoft is playing a super long game with all these acquisitions, so maybe in the end it would be worth the hundreds of millions it costs. Just seems unlikely to me.

 

[liams]

An engine is not just rendering but all the tools and the interoperability with 3rd party tools.
For example, I remember a dev commenting on the engine that they were forced to use, and how much slowed them down with little things like the ineffective search engine for the shaders, so instead of reutilizing one they add to write it again.

Microsoft is directed toward a future autarchic ecosystem, where they have the platforms, the infrastructure, the store, the talents, and the ips.
What they lack is just a resident engine, to acquire a little independence from Sony's participated Unreal, and leverage on Series' features and quirks.

On the pros: share code and assets between internal studios, and speed up development.
On the cons: anyone uses UE, developers, artists, toolchain... To switch they must have something competitive and flexible.
Enter ID tech, the engine known as doom engine because it's been used for.no driving games, no fighting, no lot of things.

I could see microsoft making it compatible with tools that work with unreal, they have no qualms about doing things like that in the wider org so it shouldnt be a problem here. I can imagine Tim sweeney will have some choice things to say about it though.


One concern that they should have is that epic is buying up tons of middle ware companies, just recently they invested into sideFX who make Houdini, If epic purchased them and made it so you could only use houdini with unreal it would make a serious competitive moat around unreal engine. Last year they also purchased a company that specializes in capturing facial performances, they worked closely with ninja theory on hell blade

 

[zed]

Look at unreal, version 1 came out in 1998, and you could say each subsequent release has just been an overhaul of the previous version with some added features, but no one complains about that

sorry cant let this slide, any programmer that uses unreal complains about it.
... but its tools are good

 

[BRiT]

Any developer who doesn't complain isn't doing their job and is genuinely slacking off.

 

[cwjs]

One concern that they should have is that epic is buying up tons of middle ware companies, just recently they invested into sideFX who make Houdini, If epic purchased them and made it so you could only use houdini with unreal it would make a serious competitive moat around unreal engine

While im not a business expert and ultimately no epic aquisition would totally shock me, I don't think this is a realistic concern -- sidefx is a private company and its hard to imagine them giving up all their film vfx users (still this big majority of revenue i assume) and their big PR/presumably $ relationships with ubisoft, guerilla, insomniac, naughty dog, etc... If they're shopping around to be purchased it'd be from adobe or autodesk or somebody else like that. Also you gotta think about epics core business -- automatic processing or asset libraries that they can stick into unreal make sense, sure, but they're not a DCC sofftware company, and stuff like houdini would represent a big leap outside of their comfort zone.

Smaller pipeline stuff like simplygon, megascans, etc, sure.... but microsoft bought simplygon, for exactly this kind of reason.

 

[liams]

While im not a business expert and ultimately no epic aquisition would totally shock me, I don't think this is a realistic concern -- sidefx is a private company and its hard to imagine them giving up all their film vfx users (still this big majority of revenue i assume) and their big PR/presumably $ relationships with ubisoft, guerilla, insomniac, naughty dog, etc... If they're shopping around to be purchased it'd be from adobe or autodesk or somebody else like that. Also you gotta think about epics core business -- automatic processing or asset libraries that they can stick into unreal make sense, sure, but they're not a DCC sofftware company, and stuff like houdini would represent a big leap outside of their comfort zone.

Smaller pipeline stuff like simplygon, megascans, etc, sure.... but microsoft bought simplygon, for exactly this kind of reason.

just 3 weeks ago epic acquired RAD Game tools who make a bunch of different game development tools that are used by pretty much everyone, while they have said that they will support the licensing model as it currently exists theres nothing stopping them from refusing to license out the technologies RAD has going into the future.

Im not saying any of these middleware companies being acquired by epic is the death knell of other engines or anything like that, it just solidifies the existing third party engine duopoly, sure epic will happily keep selling the acquired companies products going forward, but would likely not support a new engine that they hadn't supported when the company was acquired, or they will just raise the price of licensing the technologies to the point that it costs the same as unreal engine would anyway (5% of revenue)

 

[dobwal]

I don’t believe in mandating a standard engine.

I believe MS should initially provide additional resources to the most notable engines in its stable of devs. The additional resources would be use to make each engine as user friendly as possible to outside devs.

Existing and new devs would have the option to pick and choose what engine they would want to adopt and the engines would generate licensing fees from games developed using those engines. Licensing fees would be used to maintain the engine teams and bolster the engines’ performance and capabilities over time.

In other words let a commonly used MS engine or engines be the product of talent and a competitive market within MS stable.

 

[liams]

On Major Nelsons podcast this week there was an interview with James Gwertzman who is the head of cloud gaming at Microsoft (things like playfab, not xcloud to be clear)

There were a few interesting bits of information from the interview.

He said we will be seeing more from computers and AI/machine learning in the content creation process, and used the geometry generation from bing maps in flight simulator as an example
He said that they are starting to invest in AI/machine learning content creation


also at gamestack live in April they will talk about some of the audio tech they are working on, this is a guess on my part but I would say they are likely commercializing 'Project Acoustics' that can be seen in the video linked below.




Project Acoustics: Making Waves with Triton - YouTube

for a more detailed look at project acoustics

 

[cwjs]

just 3 weeks ago epic acquired RAD Game tools who make a bunch of different game development tools that are used by pretty much everyone, while they have said that they will support the licensing model as it currently exists theres nothing stopping them from refusing to license out the technologies RAD has going into the future.

Im not saying any of these middleware companies being acquired by epic is the death knell of other engines or anything like that, it just solidifies the existing third party engine duopoly, sure epic will happily keep selling the acquired companies products going forward, but would likely not support a new engine that they hadn't supported when the company was acquired, or they will just raise the price of licensing the technologies to the point that it costs the same as unreal engine would anyway (5% of revenue)

I forgot about rad game tools -- it definitely fits into the fuzzy "tech middleware" category i had in mind when i mentioned Simplygon though. (just way bigger impact)

 

[liams]

someone on xboxera found a very interesting job posting on microsofts job site

its for a 'Senior Director, Cloud Gaming' position

Some extracts:

"We are seeking a Senior Director, Cloud Gaming to shape and drive a new initiative focused on making great games designed for the cloud. .... bringing new cloud-based technologies into our games."

"Build an internal multidisciplinary Xbox team from the ground up to advance the world of cloud gaming through partnerships with independent developers."

"Bridge the gap between our platform technology teams and our partner developers, while creating service-based business plans for the games developed by your team."


So it seems like they are expanding their business to business relationships around cloud gaming/xcloud, but are also investigating ways that they could leverage cloud stuff (Ala crackdown 3 as pitched maybe?) in their games.

I guess things like high level AI actions could work well being done in the cloud, you could have some AI system in the cloud that sends a high level schedule for all the npcs that would then execute locally?



I wonder what other ways they could leverage cloud-based technologies in games?



link to job posting:
Senior Director, Cloud Gaming - Xbox Game Studios in Redmond, Washington, United States | Engineering at Microsoft

 

[BRiT]

I wonder what other ways they could leverage cloud-based technologies in games?

Isn't that what Microsoft Flight Simulator uses for matching the weather and flights of other planes?

 

[liams]

this is more (potentially) related to playfab, but someone on resetera found something interesting on LinkedIn
Johan became the principal machine learning engineer at the Coalition this month (feb 2021), and previously, since 2013, worked as a Principal research software engineer at Microsoft, working on machine learning research.

During the interviews that the former CEO of playfab, now iirc head of cloud gaming at Microsoft did recently he said that they are starting to invest in machine learning based content generation for game production. The coalition seems to be the studio Microsoft uses to implement novel technologies, for instance they did the first implementation of project acoustics in one of the gears games. So maybe its possible they are trialing some content generation tech that they have been working on.



link to the LinkedIn
Johan Verwey | LinkedIn

EDIT:

Forgot to add how he describes his role at the Coalition, he describes it as:
"Empowering game developers to achieve more by leveraging machine learning and AI."

 

[scently]

this is more (potentially) related to playfab, but someone on resetera found something interesting on LinkedIn
Johan became the principal machine learning engineer at the Coalition this month (feb 2021), and previously, since 2013, worked as a Principal research software engineer at Microsoft, working on machine learning research.

During the interviews that the former CEO of playfab, now iirc head of cloud gaming at Microsoft did recently he said that they are starting to invest in machine learning based content generation for game production. The coalition seems to be the studio Microsoft uses to implement novel technologies, for instance they did the first implementation of project acoustics in one of the gears games. So maybe its possible they are trialing some content generation tech that they have been working on.



link to the LinkedIn
Johan Verwey | LinkedIn

EDIT:

Forgot to add how he describes his role at the Coalition, he describes it as:
"Empowering game developers to achieve more by leveraging machine learning and AI."

The stuff in the playfab interview is most likely from Ninja Theory. They have been about ML and AI for a long while and what was described in the interview matches some detail they noted on one of their job postings. The job posting by The Coalition is relatively new.

 

[liams]

The stuff in the playfab interview is most likely from Ninja Theory. They have been about ML and AI for a long while and what was described in the interview matches some detail they noted on one of their job postings. The job posting by The Coalition is relatively new.

oh interesting I didn't know ninja theory has been using ML, by a long while do you mean pre-acquisition? If so that probably explains how they made hellblade with such a small group of developers.

there is a lot of promise with ML in game development, it will be fascinating to see how it develops into the future

 

[scently]

oh interesting I didn't know ninja theory has been using ML, by a long while do you mean pre-acquisition? If so that probably explains how they made hellblade with such a small group of developers.

there is a lot of promise with ML in game development, it will be fascinating to see how it develops into the future

Either 2019 or early 2020. They have been on it for a while.

 

[cheapchips]

Probably worth reposting this here. Shows some of what NT of playing around with level of detail and production approach

Wonder how much of this ends up shared over those MS Studios that are doing realistic environments in UE. Same with realistic character work, which NT do an impressive job with.

 

[SmooTh]

oh interesting I didn't know ninja theory has been using ML, by a long while do you mean pre-acquisition? If so that probably explains how they made hellblade with such a small group of developers.

there is a lot of promise with ML in game development, it will be fascinating to see how it develops into the future

They started with Bleeding Edge and swapped then atleast to Project Mara (and probably to Hellblade 2)