Official Introducing Gigaya: Unity's upcoming sample game

Tat NavMeshQuery I believe it's still run at C++ engine side. I would like to have full pure dots navigation solution to get maximum performance.

Actually I really want official release for dots AI solution that can use at production. That article I think it's just introduction to AI concept and I believe it can't use at production. What I want to say is currently I dun have resource to really implement in house AI solution from scratch and keep maintaining it. I want working solution AI from official and will keep maintaining by official.

 

I'm not saying they will not deliver this, but not now. Now they are in a transition period where they are cutting heads in the company. While we talk here, there is HR team that looks at who is not important, how to merge teams to make them more "efficient", stuff like that. Firing thousands of people takes time.

I doubt that people are productive right now, because of the overall stress this situation implies.

So once, and if this situation calms down, they may deliver what you need.

Because of the PR disaster they have managed to enter in, they will try to put a lot of eggs in the next unity version. So tools and products that are off limits to talk about by staff will most likely get pushed towards that version. But, it will be a beta version will lots of bugs, and will take years to sort out.

To make it short, if you want something right now you are waiting for nothing

 

This is not bad, thinking on making money with your games is not a bad thing

 

what a shame

 

By the way, I've also started programming my own C# based engine, but in my case just for fun.

 

It boggles my mind that something as basic as navigation that almost all games need is only under consideration and not an absolute requirement. If it doesn't get considered are we the game devs seriously going to have to re-invent the wheel making pathfinders and navmesh systems rather than actually make games? It is frustrating when we have to make engine related systems rather than spend the time making games from those systems which is ultimately the whole goal of using Unity rather than making our own engine.

 

The question is why isn't Unity grabbing AAA level games like Unreal does? Thats something that Unity is really missing. When it comes to smaller indie games and mobile games, Unity dominates. But it misses the boat big time when you want to scale up to a big project - people/studios seem to just opt for Unreal in that situation...so what is Unity getting wrong here that fails to entice some of those studios?

 

I predict Epic is just going to wait for smartphones to get strong enough to handle AAA graphics and content. And its certainly possible within the next few years; especially when factoring in streaming and the leaps of smartphones every year. Would be cheaper and less work than trying to get Lumen/Nanite running on current phones; just wait for the tide to rise.

 

All the new Unity tech that can provide AAA experience is still very green and unproven in AAA. Unreal has a long history of shipping AAA titles, so there's trust that it can be done. And Unity's ability to deliver the tech they promise themselves has been spotty at best.

And for all the S*** visual scripting gets around these forums, AAA likes having artists, designers, level designers and other creative roles be productive without taxing the programming resource unnecessarily. Sure, a lot of it is rewritten to code in the end for optimization, but designers and artist types are still free to rapidly iterate.

Currently, Unity does not have a single Visual Scripting tool that can scale up to AAA. So the studio can get source access and implement their own VS or skip visual scripting which bottlenecks the programming resource, or they can choose Unreal which is designed for AAA workflows.

EDIT: To be fair, Unity are working on unifying all graph based tools under the same UI/UX with Graph Tools Foundation. And they are also working on rewriting Unity Visual Scripting - new performant runtime (per node C# gen for performance in builds, current runtime is reflection based), new API, new UI, deeper engine integration, etc. Hopefully this finally materializes in Unity 2023. It's been more than two years since they started on this quest with nothing to show for it so far.

 

Maybe because unity's DOTS was taking too much time so they just started making their own

 

They made the right decision, but I don't think it's necessarily one that makes Unity a negative. It's the sort of decision any AAA game studio kind of wants if their objectives are a bit too custom. They didn't pick Unreal either.

 

My 2 cents: visual scripting done by non-programmers often does end up taxing the programmers a lot. It's often the cause of either performance issues, architecture issues, or both. And these issues lead to programmers having to redo and refactor everything late in development, which ends up taking even more time than if programmers just did everything in the first place.

The bad thing here isn't VS itself, but it's giving programming tasks to non-programmers. VS is still just programming, but with rectangles instead of words. A superficial difference

 

High-level nodes can still be custom tailored per project, which would nullify most of the problems you outline. It's a problem when a non-coder does low level scripting but a high level node is written by a programmer, so performance and architecture aligns with the rest of the project.

 

The fact is it is silly to assume programmers aren't aware of what it is like to not know how to code. Believe me, every programmer started clueless. Then they realised what they REALLY needed from visual scripting back in the day. TLDR: trust the programmer to design good visual scripting.

It is NOT the same as asking a programmer to design painting tools. That is the problem, when artists figure somehow because they're involved, it's somehow the artist knows best. That's just silly since vs is a problem space shared by a lot of disciplines - including writers and designers.

Use a proper problem solver: a programmer. They still remember when they couldn't program. Good feedback and peer review will do the rest.

 

While not exactly the same, I know that Rust (the game) spent months switching over to URP, and then realized it wasn't viable and dropped the whole switch and stuck with builtin. Lots of money wasted on that mistake.

EDIT: HDRP, not URP

 

It was HDRP wasn't it?

 

It was.

Eventually they did an "HDRP update", which made all the visual improvements they were planning with HDRP, but in built-in, with better performance.

 

Bluntly speaking, for many developers in the AAA space Unity isn't considered a "proper" game engine. The fact 99% of the game has to be written in the "scripting language" combined with access to the source code being the exception, not the rule, are big red flags. The fact that nobody out there is shipping an AAA-grade Unity game is the final nail.

Unreal was designed from the ground up around the way AAA devs like to work, and even so it still had quite the uphill battle to lure some studios away from their in-house solutions. Unity, as it is, has no chance in that space, no matter how many good looking demos they put out every GDC.

Big publishers like EA and Activision will happily use Unity for mobile projects, however.

 

Yeah but what you have to look at is a new generation of studios. These kids, they're flush with cash and bred on Unity. That's our future.

 

Are you suggesting then Unity needs to give us more access to lower levels of the engine for customising to our needs? Even the ability to use C++ for performant critical areas to ever entice bigger studios to consider Unity over Unreal or their own in house engine?

 

Then who's going to make the next Unity when there's nobody left who knows how to tinker with the innards and everyone is entirely dependent on someone else to update SDKs, expose the shiny new graphics features or support new platforms?

(This is actually a problem across pretty much the entire software development industry.)

That's a prerequisite. Without source access Unity is seen as a super-evolved Flash (technically, a super powered Adobe Air, but still).

But these big studios can afford Unity source licenses, and they do regularly license it for their mobile games (a GaaS can't afford being at the mercy of a 3rd party to get engine bug fixes). That brings us back to the next big issue that keeps large console/PC studios away from Unity: the chicken-and-egg dilemma. Since nobody makes such games using Unity, it's unproven how well it can be used to make such games (if at all).

After all, the less common the use-case is, the more likely a piece of software is to break when applied to it.

So, Unity would need to drastically change themselves to get to a point where a company like CD Project Red would even entertain the idea of moving the next entry in their franchise to it. Personally, I don't think this is viable or even a good idea.

 

I think it's a terrible idea. But many tech companies thrive by having a particular solution. It's clear that no matter what Unity does at this point, mobile is going to be number one in every single budget choice.

So if I wanted to have fun with high end features with an indie perspective (it's not illegal to do this), I can probably spend all my time making tools and optimising around HDRP or I won't.

Recently I found out there's no real good way to mix custom generated shadows with Untiy's own. You have to replace the lot, or hack it in sideways with a bit of ugly. Source would fix that.

Thing is I don't even want to move or anything, it's just almost like I'm being shoved.

 

CD Project Red will use unreal for the Witcher 4

 

About the visual improvements, not really. They simply brought all the textures/new meshes they intended to put in HDRP into built-in. That's it. Was basically just an asset refresh.

HDRP would've made the game look quite a bit different, not just from all the additional features, but lighting itself looks better in HDRP imo. Less flat.
They posted a few pictures back when it was really in HDRP, it looked quite different.

About the switch, to be fair, this was ages ago. They did this on.. 2019 HDRP, or maybe even older. HDRP is night and day when you compare 2019v to 2020/2021/2022.
The performance issues they were having was a drop of 15-20 FPS in HDRP compared to bIn

 

I expect supply and demand to take care of it. The job by itself is interesting for many and I doubt that curiosity will be killed that easily.
I hope Unity didn't fire any of those experts, that would be a bad idea

 

Let me generalize this please, because I've been on situations like this many times:

> Recently I found out there's no real good way to [...]. You have to replace the lot, or hack it in sideways with a bit of ugly. Source would fix that.
I'm strongly resisting to list my own problems that would fit there,but there are at least 6 systems.

I would have wanted to see Unity dealing with users and their gamepads. Steam may deal with it but if they had released on other platforms, supporting only a couple of gamepads like they do now wouldn't be enough, which would require rebinding. Seeing that (and other) code made by the ones that know about it or can contact those who do would be very interesting. That's if they don't cut corners, like how the grass on top of moving things stays in place as shown in the trailer.

 

Well I've released a few things in Unity (all of them mobile level stuff) and it just seems to be an uphill battle whenever I want to get ambitious with Unity. I think Gigaya was that sort of sentiment echoed among devs.

You see, Unity STILL hasn't experienced the pain of putting more money on the line. With Epic, you see that they have to create technology that works in a live game, and invest in that, or it will harm what is now their core business, fornite. That's very strong for the engine.

Unity felt a little like it might cost some money so immediately cancelled it with "we think its fine, accelerate will keep giving us money and maybe somehow something will drip down to the scrubs". Could be imagining things.

I mean they've spun it like we truly benefit from any accelerate stuff they get up to. If so, where's the tools? With Gigaya being commercial they'd have had to invest some in tools. It's just not profitable for Unity to do games. They tried and I thank them for their response - too much of a risk using their own engine. Taught me a lot.

 

There is a wild misconception about refactoring visual scripting from non programmer. Ie the idea you could do it right in the first place. Sure if you are making a very simple clone, but most of the time, you gonna iterate wild design with architecture breaking exceptions and addition, and the loop of waiting the artist and design to make its mind about where the fun while implementing that moving target was wildly inefficient and killed creativity. You really want the designer artist to test all it's ideas and home in the final one, before you can take that idea and streamline it into code. Because you are non longer solving two things at the same time, ie efficient architecture and fun game mechanics (add figuring out workflow before actual production). It's the equivalent of don't let prototype become production code.

Stuff midcore company does that typical small indie don't nor can:

- Hired expensive software engineer
- Have those software engineer rewrite c# unity stuff (being impress it's done with URP, BUT URP was rewrote 80% by a guy and ported some HDRP part into it)
- Have the software engineer rewrote bought asset as a base to adapt to the project
- Have the software engineer wrote tools inside unity
- Have the software engineer wrote external tools to create asset for unity engine
- Have the software engineer access unity tools to fix or adapt code to their use case
- have the software engineer refactor designer's and artist technical mess after validation of concepts

- Use unity as a frontend (for artist and designer) to create asset, see, validate, tools to export to the integration build
- use unity like the unity demo team, as a fancy "viewer" of external data

The issue with unity they (midcore) tend to have is mostly what I would categorize as architecture and "IO" (input output), because they have the resources to replace almost anything and will act accordingly. IO is file management issues (build time, compile time, asset export and import, streaming issues, etc ...), input (compatibility, polling, timing, etc ...), rendering and sound, cpu and memory management, core engine bottlenecks, etc ... All are low level access and high skills issues that tend to be locked under the engine.

The issue with dedicated unity tools is how they are advertised, promise a lot and under deliver. To contrast unity is a lot more humble and front load all the limitation even when the solution is very obviously close to magical and paradigm shift (like don't use nanite for vegetion, even though it has similar performance than not using it, but the message is really don't expect same gain fthan for static). With unity what happen is that you invest a lot in a tools, only to realize it one obscure limitations that seems arbitrary, then sunk cost fallacy kicked in and you are frustrated to have invested so much time, in contrast clear warning from unreal leave people satisfied when they misused nanite for vegetation because it was a conscious decision to confronts the limitation, I had similar feeling when unity 2 to 4 didn't have grabpass nor shadow on free tier, I knew what to expect when I choose to go rogue when I realized I could still have shadow (limited to static) by simply making shader with baked depth map.

 

When Unity moved to making things free and introduced PBR I genuinely felt here was a market leader to take on Epic. But instead of Unity focusing on real games as a core business, they quietly strengthened and grew the company around mobile. We all kind of thought that was OK cos you could just scale built-in. It would go from mobile to high end and back again.

But then - plot twist - Unity felt the need to announce DOTS and SRP! And then there was HDRP! Yes everyone, Unity was taking on AAA and we all believed there would be some seriously great strides in that area. But really, since all those years ago, barely no real movement has been made on the tooling. You got shader graph (a decade later than the competition) and maybe a trickle of tools over time. To be honest I should have seen clearly things were able to move way, way faster. After all, the competition managed to release an entire new engine in the time it's taken for DOTS to be announced and not finished yet. Same for SRP. It's just taken forever and it's not the staff's fault.

It's the fact there has been no real word from Unity at all. We have asked and asked, and honestly even now it's like trying to draw blood from a stone. I don't really see the scrappy startup I fell in love with at least. Anyway, cutting a long story short... Unity is extremely cool, a brilliant engine. But it's got bollocks all tooling and there's a mountain of technical debt and no guarantee that there will be a focus on fixing that. Just more Gandalf gesturing. I seriously like Unity, just nah I can't really see it changing much.

 

Actually the fact that the technical debt of Unity is much lesser than UE. I think I already see a lot of people mentions that UE is tested with Fornite production project, UE is better than Unity and etc things like that. But the reality is UE also has lots of issues that a lot of UE developers tells me that if you working on serious production. UE from surface level u see there are lots of awesome artist tooling out of the box but the performance is just not enough for production level. Not to mention that UE also can corrupt whole game project. The even painful thing is you will spend a lot of time just to get the whole thing working before u really can work on actual game. The most painful thing is when u really want to modify engine code. Then u will see the technical debt of UE.

And also u can see a lot of UE games suffer performance issue and even FF7 remake that is created by big company also has performance issue.

 

So far the big consumer (gamer) facing issue of unreal has been stutter due to shader compilation.

Also I think Unity didn't just jumped on AAA only, they really thought they could replicate the success of mobile with VR and AR, which didn't materialized, that's when the wild "diversification" of market happen ( a lot are adjacent to VR and AR professional use case).

But their big ass AAA failure was that game ReCore, it exposed unity has not ready for the big jump.

 

@The_Island I have lot of feedback that I wish Unity official can improve it but I will mention the most important core thing first. Will all the future Unity tooling will use dots tech under the hood at low level to develop to get maximum performance and then at high level provide both game object and dots workflow i.e. hybrid workflow for unity developer to choose it? Game object workflow means it's game object wrapper for unity developer that still want to stay at game object land.

The main reason I mention is I hope it's the last time that official can come out truly working tooling that is performance by default working across all the supported platforms and high quality code that able to keep continue to improve and maintain it. Then official no longer need to keep reinvent new system and then sunset old system.

 

Think it's gonna take some time. A long time, realistically.

 

i got some enlightening info on UE's issues on Reddit awhile back. basically as long as what you want to do fits with UE's existing frameworks and such, you're good to go. that poster said there was something like 20,000 lines of C++ code making up the player controller. so if you're working with Blueprints you're fine, and even a bit of customizing is reasonably easy. but if you have to customize or refactor in more depth in C++, then good luck wading through the code to find out how to do that.

another thing i read is that UE doesn't allow that much shader customizing compared to Unity, although its default look is quite AAA looking. i have no verification on whether this lack of custom shaders is true. other things i know is that the engine can break after updates (and that this can break middleware as a result), it seems to have finicky character rigging requirements, and i'm not sure how good the camera system is compared to Cinemachine, although i think since movie companies use it, it must be decent at that.

but i think the two factors that block me are not enough depth of documentation or forums, and the excessive GPU requirements to even run the engine. i installed UE5 on my MacStudio and it took an hour and a half to compile the shaders. since i teach game audio and use game engines i need to make sure what i run will run on a variety of hardware. i know Unity can fall back to very basic GPU requirements, and maybe its possible with UE as well - but there's extremely scanty information - last i read i would have recompile from source. In Unity i just have to decrease Quality.

 

The thing that really scares Unity is not Unreal, is Godot. Version 4.0 already provides great improvements, they also got tons of money and Juan hired a new graphic team. That little boy will create some serious competition very soon. It won't replace Unity right away thats for sure but it will become a serious competitor and the easiest and lighter of all current engines.
Think of the Blender of game engines. The editor is so lightweight that it runs on a frigging Android phone...
So many people speak about build performance, but did anyone speak about workflow and iteration (where most of your time will be spent)? What about Unity editor being so bloathed and slugish on every major release?

 

There is definitely a real lack of documentation and their forums are far behind these, but the engine only appears to be excessively demanding because it's defaulting to the highest graphical setting out of the box. You can run it on older hardware. You just have to turn it down once you're in the editor.

UE 5.0 runs through Rosetta. My understanding is that is going to start changing with UE 5.1.

 

Which team are you on/feature are you working on?

 

I feel like the recent events have been a turning point for you in terms of Unity. Before, I had the feeling that you were quite an advocate for the engine (especially taking into consideration your contribution as a forum mod).

Have they ever approached you to discuss the problems you encounter?

Back on topic: The ship has left the harbor for Gigaya. Even before this happened, I had to reconsider the usage of Unity in our projects. Our final verdict was that we're using it for mobile/VR and quick prototyping but not any bigger project. I've invested A LOT of money in training my employees (artists and developers) to use the engine and to optimize the assets and code to a max (in such a way that even Unity employees asked us how we managed to get those FPS for a car configurator on tablets a couple of years ago).

We've made our choice for another engine (which requires training and thus, again, quite some investments), but for bigger projects we want to tackle in the future, Unity didn't quite cut it anymore.

 

Hah! i remember that bug, good times. Had to write a script to recover my scene file. lol

 

I see. What I want to say is at low level the new Unity tooling / feature should be implemented in dots to enable code sharing for both dots and game object lands. So u don't need to do implement double work for both dots and game object lands which is too much work and not efficient.

 

it has 20k lines of code because is a very advanced character controller. It has Ai support so you can posses a character or unposses and the AI will use behaviors trees to move the character you have left alone. if you want a similar experience in unity you need to buy a character controller from the store which you guessed, comes with thousands of lines of code. What is the difference apart the money investment?

First the asset store item have big chances to have been build by someone like me and you who was a bit earlier to the game dev, and just hack it together and added pieces over the years without a clear plan ahead. Then is the backward compatibility which would had prevent him rework the entire thing even if he wanted.

Second, as a unity team member confirmed in this thread, unity doesn't play super well with third party assets. They still have problems with bolt. But this is not really the engine fault, I'm sure unreal has same issues with third party asset. is logical that some basement dev can't compete with a team that knows the engine in and out. he can do his best but at the same time ignore something in the engine that conflicts with hi asset.

I've made the mistake at the end of 2019 to not actually try both engines but just listen to people talking about which one is better. There is a popular dude on youtube that creates tutorial videos about unity and why he switched from unreal to unity and how is better and stuff like that. I was an idiot to listen to these people. should had tried myself and invest the time to figure this out. Don't do same mistake I've made that costed me two years of my life losing time with this engine. Don't listen to some random dude on reddit and don't listen to me. Try yourself.

 

It runs on browser
https://editor.godotengine.org/releases/latest/

 

Just a tiny side note in the grand scheme of things, but of course the Starter Asset maintainer is also fired.

I look forward to see how maintained this asset publisher page would be in 6 months and whether people will rate them "unmaintained asset, 1 star"

https://assetstore.unity.com/publishers/1

 

They have, yeah.

When I try to help and muck in, I give 100%. I don't ask for much back, just the opportunity to keep on giving at my own pace. I don't think that's gonna carry on.

Really did care a lot about community's health and fellow devs. Cared a lot about Unity devs as well. Those are real people.

Unity never reached out to me or helped all these years. I have been a really useful idiot.

 

I really love unity, been using it from version 3.5..and made mill sims, some boat sims, some shooter sims, some underwater exploration vehicle sims, made some jackpot machine sims, made medical sims, made drone sims that connect to arduino and flight controller, and because unity i am able enter game industry (non sims related) as unreal engine programmer...i even made demo using unity and submit it to the recruitment people at said company and accepted, made to learn unreal in a week and go full blown productions team and i succeed and yes i often made snarky remark at company discord channel i can made that feature A with unity in matter of day not week...lol..

PS: forget to mention..when you guys hire that you know who as a CEO some ages ago..i know you people will taste this bitterness..i know i felt this as well.

 

Uff... This hit hard. You may have been ignored by the company, but I can't imagine the Unity community without your contribution. The forum is as important as the engine itself.

 

Can John Riccitiello just get fired already? We're beyond sinking ships. Unitys management is complete shambles.