Unreal Engine 5 = Game Changer

Honestly?

Nah.

I use Blender day-in and day-out and if there's one thing that anyone who has used Blender in this capacity can tell you, it's that as soon as you fire up the compositor, Blender falls to absolute pieces. There are only like three compositor nodes that are even GPU accelerated. When you venture outside of those, you're starting to see several second per frame render times, minimum. I ran headlong into this issue when I tried to recreate the image effects I use for my game (a game that is targeting integrated graphics cards at 30fps) in the compositor and ended up seeing the render times jump to 2.1 seconds per 960x540 frame.

This is a problem when the entire reason I implemented it in the first place was because I wanted to be able to test my image effects before export so I'd know immediately how they'd look in game?

 

Loads of people say this and I have yet to hear a compelling response to my primary question:

Why would this make a difference when it hasn't for any of the dozen demos they've created where they end up making custom workflow and production tools that never end up in the engine.

 

Because a demo (demo cinematic, not even demo games) is not a fully fledged game? And like I said, making workarounds for themselves, for their specific cases (and keeping it for themselves or simply not going into the engine) won't fix the problems with the engine/editor... But, it's just my 2 cents. It's just that... from time to time while working within the editor or coding we encounter stuff that we think: If they used the engine and needed to deal with this they may already have fixed this. But again, just a bet.
Also, open sourcing would be cool, found it very useful that they open sourced parts of the engine for reference, it help us a lot too.

 

Good point!

 

If we ignore the whole problem they have where everything ends up with a custom workflow and production tools there are some aspects of the engine that can't be properly tested with just demos that have no user control or interaction. An easy example is the new input system but you could make a case for the runtime component of UI Toolkit.

 

I haven't even thought about open source lol

BUT yeah, they need some amount of dog food, while it might not end up in a real workflow 10 years from the release of the game, they at least would have an insight in solving the right problem, instead of proposing silver bullet solution that only goes half way there. It would also help to integrate all problematiques into a single vision (make a game).

The great thing about nanite/lumen isn't that it's a great technical tour de force, is that it show many problem solved at once (technical meet workflow) presented in a way that actually make sense (both game and filmic) in a way that convince people of many walk (game, technical, artist, producer). The clarity breadth of the presentation is something unity don't have.

In general I only tangentially pay attention to unity demo, I really really get into features when you have unite (or anywhere else) talk from actual dev, and generally they bypass unity's feature entirely, but tell how you can replicate it. They really are the highlight to me, the @jbooth talk were spectacular too. Scriptable architecture was more hype than anything unity proposed, even with the inherent flaws.

IMHO the ideal game unity should do is basically fortnite but on a planet instead of island. Ie:
- a game as services, so they continue update across section (and dog food their services like ads and all),
-with dynamic environment
-time of day and weather,
-with ai scattered on the map,
-multiplayer, with dynamic events,
-non flat terrain to LOD, stream, populate and navmesh,
-that scale from mobile to high end.

 

Probably because it's not actually open source.

 

I still don't get this idea why Unity should make full games. They always showcased what a small team of professionals can achieve with a Unity version. If they made a game, they would still showcasing the same: what a small team of professionals can achieve in Unity.
This does not make any easier for the lone non-professionals anything.

And obviously they always have internal and external dogfood teams, but apparently no one cares or no one believes them?

 

IT's not about showcase, it's about proof testing and getting insight in roadblocks ahead rather than later. The only moment they decided to revise the mecanim workflow after years of complaining was after that mess about that xbox game, which is the first and last time AAA use the engine. It also cascaded into performance by default being a priority, and probably started the whole hdrp mess too. Because that was the closest they were to making a game. It really felt after they were in saving face "uh oh" moment. THere was also instance like the bomberman switch who made it very clear they were not up to the game, even after the legendary hexadrive dev get to optimize the game, the only thing they could do was put the game in a lower resolution. Meanwhile fortnite is 30fps locked on a dynamic open world on switch.

In fact playing fortnite on switch was basically seeing the team solving problem day after day, it really made a case for unreal team.

 

Open source is term I was looking for? I always got those wrong, anyway, you know, do it like UE does.

 

One benefit to consider is that while making the game they will work their way through "Polish" phase.

Currently Unity has a well-known problem where you project can be finished by a bug you'll encounter late into development, or even during polishing phase, and you won't be able to do anything about it (no source code access, after all).

Lifecycle of a demo is shorter than that of a game, has lower requirement, and there's no "polishing" part. For a cinematic part you're basically doing cinematographic VFX work, and underlying mechanics could be a large scaffold of hacks, and that wouldn't matter one bit.

 

OpenSource is a specific term associated with Free Software Foudnation and Opensource movement.

It typically refers to permissive licenses, which may or may not be viral, and you can reuse code under such licenses in your projects.

Unreal engine is proprietary. Code reuse is not allowed, even though you can see the source code.

 

One precision, when we say unity don't make game, that's not LITERALLY true, they do make games, those inconsequential demo like the boat race, the mobile runner, and various small tutorial. It's just they don't have any sorts of stakes that push for any sorts of quality, except dazzling visuals.

ninja'd by @neginfinity

 

Sure, that always an option. However, the entire point of the UE5 video was that they were able to throw a bunch of movie quality photorealistic 3D assets into a scene and it all worked without the usual fuss of building the so called game ready assets.

 

Thanks!

 

Again, this isn't compelling because NONE of their custom stuff gets backported to the engine. Ever. Unity doesn't eat their own dogfood for any of their projects. They just look at the bowl and then go to a restaurant to get something there instead.

 

I still don't get it, one thing to acquire the information and another to pull away a bunch of talented people from the engine development to a game development, which is not the profile of the company and would not benefit anyone.

I understand what you're referring to, but I think you fall for the same thing like when the gamers try to tell the developer what's wrong.
Your intention is valid, your proposed solution is probably not.
Unity is not a game developer company, practically never was.
They chose to fill this gap by multiple ways:
- they have small teams on the inside dogfooding each other's work
- they have "friendly" game studios who work closely with them on things (do you still remember Nordeus, they worked together on the first iteration of DOTS at the first place)
- they have a lot of surveys (on and off of forums)
- they have a lot of other devs whom visited by Unity people from time to time (evangelists, experts, etc... Ian Dundore anyone?)

The problem is not that Unity doesn't get the info, the problem is in the execution or (as I suspect) tremendous baggage from the "old days" and the backwards compatibility requirement.

Considering all this, I'm glad that they don't waste tens of thousands of (wo)man-hours on something they don't do seriously and focus their effort on the engine and services.

 

If you want to be gracious, Madfinger is their unofficial game team, I'm still referencing shadow gun's blog, which is like unity at its best, and they were the first to test stuff like webgl and online play.

 

I'm not understanding something. A user has to be able to choose graphical settings to push a game to different devices. That's not some uncommon part of the workflow, and I can't imagine this new setup making that difficult.

The only situation I can think of where you're not building "game ready assets" is one where your Switch port has to store the same high res textures you're using on the Xbox Series X, which sounds...dumb.

 

They did release a game in 2005, which was a commercial failure. That's before they renamed the company.

I'm not sure to which degree the "baggage" part is true. I vividly recall amazing bug aroudn 5.0.x release which caused transition events to fire twice, which produced double fireballs. Definitely wasn't the case before that point and definitely wasn't result of a baggage.

Also.. I'm not sure if current way is working well. As you may be aware, unity has a bit of a bad rep in gamer community. Typically it is perceived as an engine with poor performance. While there are nutcases which vow to never buy or play a unity game, there were multiple releases where performance weren't that great at release time. I recall Wasteland 2 being sluggish, and more recently sabnautica and its below zero iterations had issues with lit scenes being seen through windows.

While unity does attempt to maintain backward compatibility and generally does decent job, there are still circumstances where this fails. For example, since some of the more recent upgrades, it is possible that the engine won't be able to open older versions of terrain assets.

So maybe putting some hours into actual game project wouldn't be a bad idea.

 

Don't forget the mess with managing big project, like file size limitation, streaming hitch, etc ... sometimes you need scale to find problem that happen in real project. And unity itself is too fragmented, a game would unify their focus on the various moving part.

But I'm just participating in this thread because i'm impatiently waiting for the nanite lumen technical debriefing ...

 

Hence "practically".

Yeah, we all well know the problems, although I'm not sure that all of these individual cases are the failure of the engine. It may or it may not.

Obviously I'm not sure about the baggage thing either, that's why I added the "I suspect" part, which apparently fell off the radar somehow.

It still is. We're humans, we are capable of empathy, that makes us capable of do things for each other even if we have never experienced the circumstances other people were in.
Like if a social worker who works with homeless people bad at their job, the solution isn't that you throw out the social worker to the street to experience homelessness.

They have a lot to work on, I hope they will concentrate on the thing which matters, making the engine and the editor better for all of us. And not on another commercially failed game which has no meaning to anyone.

And to be honest, I don't see this all that dark as one may take away from this thread, because although there are numerous problems on many levels, they are doing great things as well, and they actually are showing improvement. Both on communication and on actual development level.

Obviously whatever I write here is only my opinion.

 

I disagree. Making an engine and not having people at the company that use it all the way through is the same as making a game the developers only play during scheduled play test sessions. Or designing mountain bikes which you only ever ride on a mock obstacle course in the company's yard.

Unity had (and has) ridiculous grievances for years that wouldn't have lasted anywhere as long if someone they couldn't ignore as easily was pestering the dev teams to fix. One example is the fast platform switching: it's good we finally have it, but damn, did it really need to take that long to do when the primary Unity market is mobile, where you need to constantly switch between platforms?

The small teams can only dogfood small projects. Scaling up is exactly Unity's biggest problem: everything grinds to a halt when your project becomes too large and people need to resort to ridiculous workarounds like breaking things up in multiple sub-projects, essentially fighting against the engine. And their dogfooding is incomplete, since these small projects never go through the most grueling phase of any game: actually shipping the darn thing to consumers. Want an example? We had to hack our way through C# decompilers to get Unity to produce valid PS4 patch builds because their standard packaging procedure does not fill in a required parameter correctly, writing scripts to fix the issue and repackage the game after Unity does it's own packaging. We had to do something similar with Xbox One because Unity wouldn't copy a DLL we needed into the correct place. They never go through this stuff themselves. It's not under their behinds the fires lit up under when the crash rate of your Android game goes up by 300% after upgrading to a newer Unity version. If Epic screws up something big in their code, the first victim will be Fortnite.

 

You're implying that asking unity to make a game is comparable to throwing people out one the street to live as homeless.

This is kinda ridiculous.

• It is also worth keeping in mind, that assuming the numbers mentioned previously are correct, unity has more staff than Epic, makes less money.
• The demo videos released previously to advertise the features likely took as much time a small game to produce. Adam demo took 5 months to produce, apparently. It could've been a game produced in that timeframe instead, and it would've been more relevant to engine's target audience.
• It is worth keeping in mind that while trying to repeat success of fortnite is likely fool's errand, game production is additional revenue stream in the end. Fortnite wasn't even supposed to be a battle royale, I think it was supposed to be a zombie game.
• Unity also already attempted to make their own linkedln/facebook, or whatever it is supposed to be, in form of connect. Maybe making a game could've been a better idea.

Given that information, I'm not seeing why making a game would be such horrible thiong to do that you'd need to pull empathy card and bring up homelessness.

For example, those demos could be games people could play at their home computers and see the power of engine in person.

-------

Also, you would expect a car company to have someone who performs a test drive, right? Why shouldn't the same thing apply to game engines?

 

This is a really bad analogy. It's more akin to, "This is a company who makes tools and guidelines for social workers, maybe it's a good idea that they also consist of practitioners from the field of social work, who actually understand the field and use the tools and guidelines they provide. Because otherwise, they're glorified consultants who have no practical experience."

 

This is factually not true. I wasn't implying anything like that. I merely demonstrated that in the real world, humans are capable doing great job for one another without experiencing the same thing. Not more, not less.

It is, so why did you try to frame like this?

Yes, which conveniently left out the sea of contractors at EPIC (yeah, those, whom being abused).

Adam demo initially wasn't in-house demo. It was a "friendly" developer (or rather film-enthusiast). And why a director/film-maker would develop a game?
And it is a really bad example, because a great tool came out of it. Cinemachine anyone?

You are implying that those web developers, who made the Connect should have made a game instead? Are you sure about this or just throwing random thought on the screen?

What? Please, read my post again, try to see what I actually wrote, not what you wanted me to write.

Maybe it wasn't the best example, but the idea still stands. You don't have to practice the things you try to help. That's why we even capable of art and help and many other things.

 

I'd honestly argue that it is the opposite. There's a reason authors are told to "write what you know." Perspective is pretty important to art. And if you're making tools? Even moreso.

 

No idea,
We are re-creating all our current cinematics in Blender for our consoles and PC port and they are running flawlessly fine (and looking awesome).
[EDIT] Btw, our cinematic workflow is not real-time, they are baked videos because importing and setting up all that in Unity is a problematic problem for us, plus our game expect videos to play, not real-time data, hence my comment about it.

 

Yet we have great feudalistic fantasies and great horrors and all that. Which obviously no one (almost) currently living human experienced. Oh, and magic!

But this is a side-quest.

Honestly, if my boss were approaching me with the idea that we need to deploy a new Netflix, I would be baffled. We are working on streaming video things (NDA don't allow me to share more).

 

Cinemachine comes from a game dev who developed it for the new homeworld games, they made an asset unity gobble up. The Adam association is opportunistic, but the adam's demo was created through the connection of one unity employee to the team doing it, since they made district 9, it was a great marketing opportunity to support them and start a small team to AAA hollywood machinima talking point.

 

On homelessness, it's actually a perfect example, sadly. As someone who tether that line constantly, the clueless of the yet empathic people who measure success on the absolutely wrong metric is something I have to endure too many times.

 

And are those cinematics using the compositor at all? Because the compositor is ultimately where the major failure point comes in for me. My game relies near entirely on post processing effects and implementing those same effects in Blender spikes render times into the seconds, making it about as efficient as a production tool for me in that context as just exporting the models directly into Unity.

 

While web and native apps are two different beast this made me laugh my *** quite a lot:

 

Web devs have had it coming for a decade plus now

 

The less said about the sorry state of our lord and savior webgpu (one reason to hate apple apparently).

 

Unreal Engine 5 Demo (Made in Unity)

not convinced, some are kinda trying

 

You could make a similar claim for UI Toolkit (previously known as UI Elements). It has a document descriptive language like HTML, it has a styling language like CSS, and we can script it with C# like you'd script a website with JavaScript.

 

To the people who trash blueprints and generally UE eco system. Let me tell you a personal story.
I started with Unity before I jumped to unreal engine last year. I am mostly an artist but before that I knew basics of programming. Anyway, I started using Unity and I couldnt help but notice how much experimental and cumbersome ways it had of doing so. Alot of the tutorials I was watching had different functionality/UI while I realisef my unity version had already changed them. I got frustrated with the engine not because I wasn’t good at it because simply I couldn’t find stuff. There are so many buttons, rollouts, ticks and all that really badly labeled and put together while in UE4 you have a very nice and clean structure for anything. I honestly didn’t need to look up anything again after going over the basic UI ONCE. Lightmapping in Unity, please spare me that memory. I had tried to get decent lighting in unity for a while but every damn attempt it would do something unexpected while when I tried the exact same thing in unreal it would work flawlessly in a much better way. But the most irritating was the lack of features the engine had out of the box. Shader graph wasnt a thing back then. If i wanted something like tesselation or other features I could easily graph up in unreal I had to learn the shader language and not to mention how alot of thing unreal had out of the box were market place assets in unity. That were some really big red flags to me. Im new to the engine, why doesnt the game engine provide me the tools to get great visuals out of the box? Im trying to make a game not coding my own volumetric lighting solution. Anyway, the summary was that I concluded that unity was just alot of mess. I can understand why its more appealing to programmers (C#) but as an overall engine i found it a mess. I felt like every update something was breaking.
I downloaded unreal and I actually made my first start to finish playable game within 2 month. And its not a jump and run but actually has decent substance. And all in blueprints. People really underestimate the power of blueprints, there have and are great looking titles being done by solo developers entirely in blueprints! When someone says blueprints are limiting then its very apparent that he hasnt given them a chance and has blindly disregarded them for whatever reason. To name a few games that were entirely done in blueprints by One developer is (Omno, Lost soul aside, bright memory, recently Lucen and many more). Those game are far from simple games but to repeat, entire complex games are being developed solely with blueprints. Artists like me are able to whip up our own games without being dependent on a programmer. With unreal our, or at least my creativity really skyrocketed and I cant think of any game I could NOT create in blueprints

 

Wrong, and ignorant comment. Plenty of complex impressive games made entirely on blueprints

 

You wouldn't know what's in the heart of a programmer or certain dev ... and the level of control and complexity they want to achieve ... you have no idea, but it's fine because you do what YOU want If you are new, there is much more than you ignore before processing those criticism.

 

I want to make games, thank you and blueprint just do that . Anything else, a normal c# dev probably would have a hard time doing as well. All im saying is stop disregarding blueprints. There are plenty of incredible games made with it. Seeing comments like above are just irritating because they seem so ignorant. I know people who are solo developers doing c# and then have artists doing their thing with bps. Guess which one delivered beautiful games. At the end of the day we re all here to make games. Im not in it to make high level engine modifications or make my own complex shading model.

 

u should avoid creating all game in blueprints sir. it is not advisable other than for prototype.

 

Nativize blueprints to c++ (yes it an in built option in unreal which improved performance ALOT)

check the performance when bps are nativized. Also the example is extreme on its own. And i like to quote „blueprints are slower than c++ but they re NOT slow“. It was one of epic games slide presentations, can’t remember which. But saying blueprints aren’t battle ready for making entire games is false.

edit: of course if you have some experienced coder, he surely can do complex stuff in c++ to get an extra edge but as a solo developer or even as a not so experienced team, BPs are capeable

 

no sir, not talking about speed optimization but clarity of code maintenance and level design.

 

also just to add most advice here is now obsoleted with new version of unity (see signature for details)

 

That can be an issue, yes. Though one could also say the same about badly written code. At the end of the day it all depends about how clean someone works. Badly written text code can also be confusing. Bp can get messier faster if you dont watch out but for solo developers who want to utilize the power of unreal like me, bps are just fine. And I also understand the desire of people to write code but please don’t trash systems you dont have real experience with

 

mam u need to enable ray tracing but perf will take a hit

 

Just to ride the point home, some games using entirely blueprints (confirmed by the dev himself):

lost soul aside (solo dev)

Omno (solo dev)

Bright memory (solo dev)

Lucen (solo dev)

https://youtu.be/h5CCDNmm7zM

And many more. Do these look like simple games to you?

 

unreal users showing up to unity forum touting blueprints as best thing since sliced bread = time to lock thread

 

Nobody is doing that, what they're doing is saying that "nobody uses them" and "visual scripting is only for simple projects" are completely off-base statements.

Which they are.