Discussion - Unity tech demo "Book of The Dead"

Hi everyone,
As i was rather curious to check the other side of the fence for the latest Unity tech demo. I loaded up the “Book of the Dead” scene and recorded some footage available below.

Amazing photogrammetry assets aside, it looks great - But - couple issues i have are temporalAA not working well enough, Depth of field quality rather average (especially after testing latest 4.20 UE4 Dof !).

Two things i truly appreciated are the immersive audio design and dynamic lightning evolution, performance was also better than i expected.

One noticeable user experience feedback, you still feel the engine editor is struggling a bit with this kind of ambitious advanced project. Unity became completely unresponsive a couple times while i was opening this project, had to restart the editor once.

Yeah, every once and awhile the Unity marketing team peaks my interest enough to get me looking at Unity again. And then I try out the latest hot new release and I remember why I’ll never use the engine for anything serious. Like you said the performance of the editor is an absolute joke. When using it even for this demo you can feel it chug under the stress of doing anything substantial. I have a pretty beefy system and getting this small demo scene up and running was like trying to swim backwards through frozen molasses. Even once everything was loaded just moving the camera around the scene gave me slideshow frame rates on a 1070.

Better luck next time Unity!

Yeah Quixel’s Megascans are awesome; expensive, but awesome.

Unity’s renderer is almost there, they “say” they are investing huge to improve editor performance in future and support huge level compositions.

I hope they do, I like Unity’s support and sorry their forums are far more responsive than these ones. I had Zero replies on my last four posts in the past few months and i have seen things die down signifcantl

I think the renderer as it stands now surpasses UE on many levels, it also surpasses on anything stylized non PBR and 2d, Also their MSAA is better and faster from what i could tell something I’ve been struggling with lately in Unreal.

I think they are almost there with graph shader. But things are still in preview mode. I already like how it saves and loads as well displays material much faster in the editor. i Unreal I have to wait up a good while just to apply material changes and save inside the editor. I still don’t get why its so slow, maybe some internal thing going on.

One thing they are seriously lacking is something similar to blueprints but more versatile (like a direct C# link up) instead of crossing fingers and trying to convert BP to C++ in UE. Unless they invest in a Unity BP they wont attract as many newcomers as they hope even with many high end assets they are putting out there.

I find it kind of funny when discussing Unity editor performance, but can someone please tell me why is it that in UE i can NEVER depend on the fps in the editor to a large margin and i have to build my game over and over again to actually test it out, while in Unity it almost always tells me the correct framerate inside the editor with a margin of error of around 10% with the final packaged game?
And I too have a beefy system.

Edit: Oh and UE performance hits bottom on dual monitor displays, I don’t care what anybody says about this, there is a serious problem here, we have enough machines with different cards and drivers and all of them suffer running this thing. and god forgive if you have a material editor or a particle editor open on the side, that will kill framerate in game preview inside editor.

I can confirm that, within the few things i made with Unity, C# made my life so much easier. You think about something and basically implement it at the same time, without wasting hours trying to understand UE4s odd C++ implementation or relying on limiting blueprints (which again are very odd to use for anything more than simple “classes”). If it wasnt for the shader editor, we already would have switched tbh … Performance is also by far better in unity (but again; the editor sucks).

I believe in a year or so time they will have the graph shader stabilized as well as some other improvements.

I think as it stands now Unity will only need a BP like C# link up as well as improvements in their animation system and some of their sequencer features, and things will be a direct head to head if not better in some instances than UE.

In the animation department Unity needs a solid IK system (which is in their to do list), I hope they will implement something like what epic did for paragon (and never released) animation blending . Also more flexibility with conditions in their animation graph with more variables to play with. Faster more flexible import settings for FBX.

Haven’t tried Discord didn’t know it existed.

True regarding Fornite, It’s also worth noting Unity CEO’s comments regarding Epic which is not far from the truth: “They have a different revenue stream in that they make games. I could argue they compete with their customers.” This is a marketing point to Unity’s favor.

To look back at Unity seriously i would have to become frustrated with something regarding Unreal Engine in the first place. Which is, as of now, far from being the case.
Working with Unreal Engine, there’s a lot of things we’re taking for granted.

Yeah … the grass is always greener on the other side of the fence. Until you look more closely and objectively.

No one said it is… Personally I want to see a much tighter head to head between Unity having been using both for a few years now, while UE I have my doubts about how Epic would deal on longer term projects and support, as well as their over all split dedication between what’s good for the development of their games instead of whats good for general development outside that sphere.

One small example is the amount of struggle you need in order to cheat effects from a tech art perspective just because the engine is built for specific PBR workflow while in Unity we could fake shadows decals reflections, isolate highlights etc… on unlit surfaces and apply loads of other effects as needed without a line of code or touching source.

As an artist I appreciate some of the features in UE but I’m hoping with Unity’s flexibility, I feel i may get a better bargain if they catch up to certain areas of development that tech artists would require as in UE, plus all the other flexibility that come with it that Unity can provide which currently are either not found in UE or very time consuming to achieve and most definitely not in the plans to make.

Ex. If it wasn’t for VR they wouldn’t even have touched Forward shading (half baked), so we would all be stuck with Differed and temporal. Nothing wrong with that but very very limiting for many types of projects.

Unity desperately needs a standard (opt-in) Gameplay Framework package.
I have no idea why they never do it.

Well, why really? This is something thats actually really good in Unity compared to the UE4 way. You have Objects and start to give them life, and thats it.
For beginners, getting into PlayerControllers/Pawns/etc isnt so easy when you are not familiar with the whole concept. And still i am not sure if i really like it, despite working with it since UDK.

On the other hand, i really really hate the editor in Unity on so many levels, but i guess most of it is due to being used to UDK from the beginning. Not having a material editor in unity doesnt help much (and no, i dont wont to buy additional plugins for that, that may be outdated a few releases later).

Then again, not having to build lighting for hours in unity is really nice; and when working with shaders and getting used to NOT having “compiling shaders (2419)” again and again and again…dammit, i just changed a constant in a single material from 1 to 0.7…

So both engines are great and not at the same time.

There is a node based material editor now in Unity, works similar to UE it even has Material functions. It needs a few more months i assume till its out of experimental. But I tested it and it works very fast.

For me I like the editor and the timeline because:

A - Interface wise it is simple and light and reminds me of regular 3d apps I’ve been working with for a long time.

B - Timeline doesn’t have a mind of its own that keeps expanding my hierarchies every other time.

Lightning and working with shaders are the 2 major points i see regarding Unity as favorable ground. I do like the BP/C++ workflow of Unreal Engine a lot.

From what i understood Niagara allows a lot more “hands on” work with shaders, i’m rather curious to experiment with it. There’s also been some work started regarding Lightmass which is extremely good news, looks like they brought some talented people together to figure it out.

Because every Unity-based studio out there end up having to build one… They build the same thing for the same purpose except each one do it their own way.

That lack of standard made me grow tired of relearning how to setup the exact same things for a particular project, this time someone did their own way. Not to mention productivity goes down until a framework is in place if said studio didn’t create one yet.

Whilst compared to Unreal the Framework classes are exactly the same for every game studio using it so I can be productive from day one joining them :slight_smile:

And you will find that those Unity studio’s still use pretty much the same underlying principles. Which should not be all that difficult to get into once you worked with it for a bit.
For the other side of the spectrum I often find that UE4 teams use indeed the framework classes, then build so heavily onto(or alter) it that it is not just the base framework anymore. Which pretty much leaves you with the same learning curve for both engines much of the time, depending on the team and/or project.

Seriously the “lack” of an “framework” is by all means not necessarily a negative point at all, this highly depends on the project and the needs of the team. You will always need to read in the code before you are 100% productive in whatever team you join, regardless of the engine they use. I have seen great Unity codebases that made me productive in no time, and horrible UE4 code that just did not make sense(which sometimes can be attributed due to the framework and workarounds). Obviously this works the other way around just as much.

I love unreal just as much as the next person here, but you really undersell Unity.

The fact the editor almost bursts at the seams shows it’s still not ready to make full blown projects at that level of asset fidelity without major headaches. It takes more than simply having the proper rendering and lighting features.

Unity’s advantage for smaller projects is undeniable, of course. Unreal’s intricate systems do have a learning curve while you can get results right away out of Unity.

Unity also has a quite more robust QA cycle due to both being longer on the market (the market outside stablished studios, that is) and its heritage as a closed-source “game maker”. Unreal’s heritage as an AAA engine carries with it the assumption people can and will modify its source code to fit their project’s needs, which is an alien concept for developers coming from a Unity-only background.