As a fellow dev on Hyenas, it's really good to finally know the details about the fast path renderer everyone loved! It's also great to finally see what the ice cream symbols meant during playtests
I would love to see these guys try to tackle rendering problems for Meta quest. This has been a lot of interesting information that I might be able to use in the future.
Curious about why they didn't mention ISM/HISMs at all. Making modular models per-se is fine, with well optimized instancing framework (there are works but much less compared to this talk) maybe the problems they encountered in the first place would not be so much of a problem. Cool engineering methodology though.
I am so confused. It looks like the art team of the project were just people who were throwing assets onto the engine as if it was a kitchen sink and then developers made overengineered solutions, reinventing the wheel on many occasions, instead of the dev team just stopping for a seconds to contemplate if the approach is even valid. Bottom line is that judging from the game's trailer, there's nothing in terms of visual quality or the world scale that would justify this degree of custom tooling on such a low level as rendering. If anything, it sheds a bit of light on why Sega cancelled the project. Must have costed excessive amount of resources to achieve average results.
I think the focus of the video is not on whether they did the correct call or not creating their game. I think the purpose of this video is to show people the customization of UE, that you can actually do anything you want pretty much. Now, most of us are aware of this already but I'm sure there are people who will have a light-bulb-moment after watching this or even just reading the title. But yes, I'm 99.99% sure your focus on the game and it's necessities are not why they had this talk. The game they worked on was simply an example to demonstrate capabilities. If the game was much larger and did look a lot prettier custom-tooling is a solution that could be and often is used. I agree though, doesn't look like it would be necessary for the project they worked on.
Not at all. Some of the things mentioned here have later been implemented by Epic themselves for example. Don't forget that Unreal is an "off-the-shelf" engine that has to cater to the masses, and while it does an admirable job, that comes with a lot of overhead many games just don't need, and it's extremely common to need to modify it - even at this level.
If they were reinventing the wheel, then it would be good to have information on what parts he proposes could be replaced with off the shelf methods. (I also think this talk is decent as a way to showcase how UE could be improved, by making more of these optimizations decently automatic)
Seemingly the art team and the game play team were supposed to work in parallel, making a 'procedural' solution necessary, which was terrible for performance. Agree with your bottom line.
I think his game scene will run smoothly by UE5 lumen&nainite or even by UE4. The result looks so bad like a mobile game. They waste lots of time modifying render pipelines like making a new engine.
Wouldn’t be cool if they added mode where it teaches you step by step how to use unreal engine. First teaches you how to change the character and how to use blueprint and the Teaches u other stuff. :D help me pls. I want to get into unreal engine. I already created my character but I have no clue what’s next.
As a fellow dev on Hyenas, it's really good to finally know the details about the fast path renderer everyone loved!
It's also great to finally see what the ice cream symbols meant during playtests
Im so sorry for the guy, that the game he worked on for 5 years didn't even survive over 1 month
5 years not 8
@@biiiiioshock Thanks for the correction
he works in the company, that means that the game is another game in their list, that be a great tool for future projects
That's what you get when you're owned by Sega
Then there's games that people work for 5 years and they never see the light of day :c
I worked on this game for a while and Grant was a tremendously smart dude, glad to see his beard is as magnificent as ever. Godspeed sir
Awesome presentation!
I would love to see these guys try to tackle rendering problems for Meta quest. This has been a lot of interesting information that I might be able to use in the future.
Was this modern rendering path done before UE5 and nanite? I wonder if they would make the same choices now as they did then.
Curious about why they didn't mention ISM/HISMs at all. Making modular models per-se is fine, with well optimized instancing framework (there are works but much less compared to this talk) maybe the problems they encountered in the first place would not be so much of a problem. Cool engineering methodology though.
where do i signup?
Good God.
Cool now i learn more of unreal enginr
Engine*
I am so confused. It looks like the art team of the project were just people who were throwing assets onto the engine as if it was a kitchen sink and then developers made overengineered solutions, reinventing the wheel on many occasions, instead of the dev team just stopping for a seconds to contemplate if the approach is even valid.
Bottom line is that judging from the game's trailer, there's nothing in terms of visual quality or the world scale that would justify this degree of custom tooling on such a low level as rendering. If anything, it sheds a bit of light on why Sega cancelled the project. Must have costed excessive amount of resources to achieve average results.
I think the focus of the video is not on whether they did the correct call or not creating their game. I think the purpose of this video is to show people the customization of UE, that you can actually do anything you want pretty much.
Now, most of us are aware of this already but I'm sure there are people who will have a light-bulb-moment after watching this or even just reading the title.
But yes, I'm 99.99% sure your focus on the game and it's necessities are not why they had this talk. The game they worked on was simply an example to demonstrate capabilities.
If the game was much larger and did look a lot prettier custom-tooling is a solution that could be and often is used. I agree though, doesn't look like it would be necessary for the project they worked on.
Not at all. Some of the things mentioned here have later been implemented by Epic themselves for example.
Don't forget that Unreal is an "off-the-shelf" engine that has to cater to the masses, and while it does an admirable job, that comes with a lot of overhead many games just don't need, and it's extremely common to need to modify it - even at this level.
@@Jambaxyea I think their descriptors or chocolate instances ended up as Lightweight Instances which is really cool if true.
If they were reinventing the wheel, then it would be good to have information on what parts he proposes could be replaced with off the shelf methods.
(I also think this talk is decent as a way to showcase how UE could be improved, by making more of these optimizations decently automatic)
Seemingly the art team and the game play team were supposed to work in parallel, making a 'procedural' solution necessary, which was terrible for performance. Agree with your bottom line.
I think his game scene will run smoothly by UE5 lumen&nainite or even by UE4. The result looks so bad like a mobile game. They waste lots of time modifying render pipelines like making a new engine.
Wouldn’t be cool if they added mode where it teaches you step by step how to use unreal engine. First teaches you how to change the character and how to use blueprint and the Teaches u other stuff. :D help me pls. I want to get into unreal engine. I already created my character but I have no clue what’s next.
Go study Computer Science. The UE is easy.
There are many tutorials on RUclips to get started! :)
DEADPOOL & WOLVERINE WILL BE AVAILABLE ON JULY 26 @ 8PM ET 🔥
i'm here!!