Jump to content

Epic upping the ante again...


Recommended Posts

While it looks impressive, it is not exactly the answer how to make things more realistic, you cannot just up the polygon count and think things become more realistic. In fact that demo looks in no way realistic, it is overall too fancy, shiny etc and the game mechanics are all over the top.


Reminds me of those "unlimited detail engine" promo videos from Euclideon, which after all those years still has not produced any games. All those demos always look so amazing, yet in reality we still don't get games that look like that.

Link to post
Share on other sites

Yeah while this is impressive its really only useful for certain set pieces such as whats shown in the video. I don't think this would work too well for something more fast paced. Also if you look closely you can see a lot of pop in on the statues when they are in shadow. If this is basing polygon counts on ray traced lighting which it very well may be, anything in shadow is going to look pretty bad if you don't use conventional methods to model and texture the geometry

Link to post
Share on other sites

Yeah while this is impressive its really only useful for certain set pieces such as whats shown in the video. I don't think this would work too well for something more fast paced. Also if you look closely you can see a lot of pop in on the statues when they are in shadow. If this is basing polygon counts on ray traced lighting which it very well may be, anything in shadow is going to look pretty bad if you don't use conventional methods to model and texture the geometry

 

Focusing on the numbers, while it's what they push as "what it can do", is missing the bigger point made, I think. It's not "what the tech can do in a controlled situation". It's more "what the tech is doing, period" and the implications that has for designers and developers working with it.


Of course it's unlikely that any game is going to have scenes with over a billion triangles in a single frame. Except maybe in scripted cutscenes/cinematics. Even then, it just doesn't sound practical to throw more at the engine/hardware than is necessary. I doubut artists are going to start modeling stuff with 1 million polys, when a fraction of that will do, "because they can".


As an analogy... My video card (2070 Super) hits 200+ FPS on a number of games without vsync. However, my monitor only handles 60 fps sync'd. So, since I always play with vsync on (because screen tear haunts my dreams), anything over 60 FPS is wasted. However, I know I still have that unused horsepower available to hit 60fps in more demanding games where it's needed. Otherwise, while 200+ FPS sounds really impressive, it's not very practical in a typical gaming scenario. At least not for me.


That's how I look at this tech demo, and tech demos in general. Sure you're not going to start seeing games with billions of tris in a single scene, but the technology is there to make it possible, in the right circumstances. The net result is that games are still going to have the potential to look a lot better than the previoust tech, and at a comparable performance.


I mean, look at the

when that was released. No UE4 game, that I'm aware of, ever pushed its graphics as far as what's in that demo, either. Same goes for
. But the technology demonstrated in each was/is used in games made with its respective engine. And besides, all else being equal, would you argue that games made with each generation of the tech weren't a notable improvement over previous versions (visually, etc)? I wouldn't, personally.


The real take-away of this video, at least for me, is how it removes several steps from the asset pipeline for artists. Being able to put full detailed models into a scene without having to worry about manually creating LODs, or having to generate normal maps or displacement maps shouldn't be understated. If you've ever gone through the process of baking normal maps, and setting up LODs on even less detailed objects, the process can be pretty tedious and time consuming. So, UE4 providing all that detail and LOD in real time is a big deal.


The radiosity lighting that updates dynamically as the environment changes, and does so realistically, based on surface type, etc, is also a huge improvement that eliminates the need for a lot of the manual lighting placement that was previously required to simulate similar effects.


The flocking system they describe for particles is also really cool. That a bunch of bats flew off out of that cave and realistically avoided the environment and each other, without an animator having to manually set those interactions up, is also a huge deal.


And so on.


In other words, it's not just how cool the tech looks. It's also the benefits it provides to those using it.


Sure, they're pushing these huge numbers in the demo (because people love big numbers). But to me, the real exciting stuff was the technology behind the scenes, and how much it frees up designers/artists/modelers, while improving the quality of the end result.


These are all things that add up to an engine that is going to power some pretty awesome gaming experiences in the years to come.


Personally, I'm really impressed and excited for this.


Of course, mileage may vary :p.


As a quick aside. Back when I first saw 'Advent Children', I wondered if or when we'd ever see real-time graphics tech capable of that level of fidelity. I'd say the answer is very much yes at this point.

Link to post
Share on other sites

It all sounds good in theory, but in reality it is always a little different. For example the supposed benefit for artists in making asset creation easier which is not really the case, yes in theory it saves you work, but in reality it first increases your work. Those big engines and tools come with big libraries of professionally created assets, that the "artist" just places into the scene. Have you ever tried to create a photoscanned 3D art asset? It takes hundreds of photos and lots of work of exporting and importing, rendering etc for such an asset to be finished, it is far easier to just take a photo and slap it as a texture on a flat face of a low poly model, looks amazingly realistic unless you get close or it gets dynamic.


Yes in theory it looks and sounds nice, but the question is how much of that do we get in reality?


Do you remember the new Unreal Tournament? It was supposed the new big deal and they had amazing looking demos, it was supposed to be open source and build by the community, but recently it was cancelled, because it failed so hard. Why did it fail? Because the users could not reproduce the good quality, you had like a handful of maps that looked as good as the demo created by professional artists from Epic and then you had everything else, that looked so bad, some levels even looked worse than Unreal Tournament from the year 1999.


Even our Torque3D demos from 2009 or whatever still look better than almost any game that was produced with the engine until now, that is the difference between theory and reality.


If you up the tech it also becomes harder for artists to compete, because you raise the entry barrier.

Link to post
Share on other sites

Do you remember the new Unreal Tournament? It was supposed the new big deal and they had amazing looking demos, it was supposed to be open source and build by the community, but recently it was cancelled, because it failed so hard. Why did it fail? Because the users could not reproduce the good quality, you had like a handful of maps that looked as good as the demo created by professional artists from Epic and then you had everything else, that looked so bad, some levels even looked worse than Unreal Tournament from the year 1999.

 

Incorrect.

UT4 didn't "fail so hard". It was put on hold indefinitely when team members working on it were shifted over to Fortnite after that game blew up. 'cause, you know.. Money. This did not happen recently. It happened in 2018.


As for your "version".. Citation Needed. Unless you can link to an official statement citing "poor results from community contributors" as their reason for halting it, this is just you "inventing facts" to fit your narrative. It's fake news.


You know, this thing called "search engines" exist. They make fact checking very quick and easy. So, you should probably stop making stuff up.


Regarding the rest of your hyperbole, okay sure.. whatever you say, Din.

 

Even our Torque3D demos from 2009 or whatever still look better than almost any game that was produced with the engine until now, that is the difference between theory and reality.

 

Highly subjective and not at all surprising, considering the source. I half expected you to say T3D looks better than the UE5 demo.

From a few of the maps I played on, personally, I would strongly disagree with you. But, you know.. opinions.


I just hope you're not referring to the still early in-dev maps for your comparison. It's not exactly a compliment to T3D if you are.

 

If you up the tech it also becomes harder for artists to compete, because you raise the entry barrier.

 

Another silly and poorly thought out argument.


You mean like what's happened with literally every new version of the engine since Unreal Engine 1? Or IdTech? Or any other engine with major numbered releases for that matter?

Yeah, of course that happens. Every new engine introduces new or otherwise increases existing requirements for asset production. You know what else happens? Artists adapt, learn the new skills and rise to the new requirements. New tools are created to assist in the creation process. People make it happen.


Also, because UE5 can handle so much more detail in real time, does not mean suddenly all assets are going to become that much more detailed. If an asset designed for UE4 is adequately detailed for something in UE5, there's no reason for them to make it more complex. Difference is, unlike UE4, they won't have to worry about LODs, Normal Maps, etc which would effectively reduce the time to get those assets done and into the engine, not increase it.

Edited by Mitovo
Link to post
Share on other sites

I do not need an "official statement" from an "official source" to tell me what is real, I can see reality with my own eyes. Of course UT4 failed, because I saw it with my own eyes, I played it a few times myself and the player count was always close to zero, maybe around 5-20 at halfway good times and regarding the community contributions, they almost all sucked, I tested almost all of the community made levels and with a few exceptions none of them even came close to the demo material put out by Epic.


You know, I can do and see things for real, I do not need an authority to tell me what reality is because I cannot do and see things for real like other people.


Regarding things supposedly getting easier for artists, I am an artist, I do art for real and I can tell you that things have not changed for a long time before all this fancy new inventions recently. Did you use Unreal Engine? A hell of a beast to learn and use. I used Cry Engine before and it was really easy to use and understand, at least the art part, you could get something done within a few days with no prior experience and since Torque is modeled after Cry Engine it shares a large part of the ease of use. And now you come as a probably non creator and try to tell me how everything is getting better and easier, while in reality everything is getting harder and more complicated. Things only get easier for people using proprietary software and use pre-made content from their fancy libraries of fancy stuff, but for people who do things for real from the start up, things get harder.

Link to post
Share on other sites

I do not need an "official statement" from an "official source" to tell me what is real, I can see reality with my own eyes. Of course UT4 failed, because I saw it with my own eyes, I played it a few times myself and the player count was always close to zero, maybe around 5-20 at halfway good times and regarding the community contributions, they almost all sucked, I tested almost all of the community made levels and with a few exceptions none of them even came close to the demo material put out by Epic.

 

As I figured. You can't provide a source, because none exists. You're making up your own facts, as usual.

Color me shocked.


 

You know, I can do and see things for real, I do not need an authority to tell me what reality is because I cannot do and see things for real like other people.

 

The degree of shameless, arrogant self-delusion in this statement is amazing.

Sums you up perfectly, though.


Translated: "I see things, and then invent my own reality of why they are to fit my personal narrative, which I then assert as fact... because I say so".

 

Regarding things supposedly getting easier for artists, I am an artist, I do art for real and I can tell you that things have not changed for a long time before all this fancy new inventions recently. Did you use Unreal Engine? A hell of a beast to learn and use. I used Cry Engine before and it was really easy to use and understand, at least the art part, you could get something done within a few days with no prior experience and since Torque is modeled after Cry Engine it shares a large part of the ease of use. And now you come as a probably non creator and try to tell me how everything is getting better and easier, while in reality everything is getting harder and more complicated. Things only get easier for people using proprietary software and use pre-made content from their fancy libraries of fancy stuff, but for people who do things for real from the start up, things get harder.

 

Oh wow, really? You've used Unreal Engine? No way! I bet you're the only person outside Epic Games who's done that! I can totally see now why you're qualified to overrule Epic's official reason for halting UT4. Obviously you're right, and they're clueless. Clearly.


I will never doubt you again, Din. In my book, your word is as God's from now on.


You used Unreal Engine. Wow. Incredible.

Link to post
Share on other sites

I'm the source and I exist where is the problem? You could also others who have played UT4, but it is hard to find them, if I was not cancelled and I had not deleted it, I could log in right now and show you, but you could have be your own source before, since to see something with your own eyes for real is a better proof than any source. Apart from all that, the fact that it was cancelled should be proof enough that it has failed. On top of that it was only officially recently cancelled, but in reality the development was pretty much halted right when it was released, since it was never developed much further than the demo you saw initially, so it was pretty much stuck in alpha forever.


But whatever you are obviously not able to understand the concept of reality and it is probably futile to debate with you.

Link to post
Share on other sites

Certainly impressive! I haven't watched the full presentation yet but did they offer any details on how they intend to keep file sizes in check? I'm hearing about thousands of models with millions of triangles and 8k+ textures everywhere and it seems to me that a full AAA game could take up hundreds of gigs worth of HD space. Downloading games that size, even at fiber speeds, would be annoying and 1TB HDs seem small by comparison.

Link to post
Share on other sites

I think some of it is boasting, the whole billion polys being crunched to 20 million is really just a decimation algorithm at this point probably being run on some custom and special processing units, in the long run, it will be interesting to see how this translates to the PC sphere and whether or not we are starting to move into the world of even more GPU threads or processor units for high-specification games. It's certainly interesting to watch, especially given the uptake of photogrammetry within gaming.


8K textures are pure boasting, apart from rare and specialist areas, utterly useless at the moment, water physics on a small sample, and tbh it didn't look awesome, so more word-salad and buzzword hyping here.


Rumour has it that the PS5 specific SSD/Memory interoperability stuff is what makes loading etc less of a problem for this stuff, again, will be interesting to see how this starts to translate to the PC sphere.


It's nice to see the software world driving the hardware world phase again, I think we've been away from this for too long in some ways.


It's easy to miss some of the points of discussion when everyone is engaging in 'game engine jingoism' Duion is correct in the assertion that developer usage of new technology does lag somewhat, we do have several examples of this over the years. He is also correct that in some ways this does raise the bar for artists, sadly at the same time, it lowers the bar knowing that the algorithms can clean up the 'mess' to a degree.


It would be much more interesting to see how this stuff actually works in practice. It's easy to get a good tech demo when you have the game developers the engine developers and the hardware developers all working together on a single project, but this is a million miles away from 99% and probably a few decimals thrown in for good measure of games being developed right now, I mean that specific demon has probably more man-hours in that 3-minute section than most indie games ever get in total.

Link to post
Share on other sites

The irony is with polycount and texture resolution, there is a maximum and pushing it beyond that does not have any effect except using more resources, they even say some triangles are just a pixel big, then why have a triangle to begin with? And an 8k texture fills a regular full HD monitor with 1920x1080 32 times with pixels, so you have to stack 32 monitors to display that texture.

Link to post
Share on other sites

Focusing on the numbers, while it's what they push as "what it can do", is missing the bigger point made, I think. It's not "what the tech can do in a controlled situation". It's more "what the tech is doing, period" and the implications that has for designers and developers working with it.

 

But the demo is a controlled situation... The thing is that whole scene screams to me that as what Bloodknight had said its just flexing and boasting. And as Duion said textures at 8k textures and certain polygons that are only a pixel size, really isnt useful to an artist or a developer, this tech demo i think was for end users people who are going to be playing games. I have no doubt this is probably running on a PS5 and in real time, but it is a completely controlled scene, it would be interesting to know the techniques that were used to make the rest of the scene, they only talk about the opening scene as being dynamic, but the rest of it im not too sure about.

 

That's how I look at this tech demo, and tech demos in general. Sure you're not going to start seeing games with billions of tris in a single scene, but the technology is there to make it possible, in the right circumstances. The net result is that games are still going to have the potential to look a lot better than the previoust tech, and at a comparable performance.

 

this is what is going to really bake your noodle, the ability to run a billion tris has always been there. You ever tried to push torque to its limit? It can render scenes like that and it wont bust the engine it just doesnt have the dynamic model and textures in it yet as far as i know? But it can do it! push a 20 million polygon model into torque and you will see what i mean it will run just as smooth as that does but again it just is not practical for anyone let alone end users. Torque can also take 8k textures, its nothing new.

 

The real take-away of this video, at least for me, is how it removes several steps from the asset pipeline for artists. Being able to put full detailed models into a scene without having to worry about manually creating LODs, or having to generate normal maps or displacement maps shouldn't be understated. If you've ever gone through the process of baking normal maps, and setting up LODs on even less detailed objects, the process can be pretty tedious and time consuming. So, UE4 providing all that detail and LOD in real time is a big deal.

 

Again sure if you want to develop games for people with the latest hardware top of the end stuff then sure, but as we all know you will only be getting into what 10% of the market? Maybe even less. Of course you could develop solely for the consoles too but then again they have to have the latest playstation or latest xbox to run this smoothly without LOD's and other methods. Also dynamic LOD's and dynamic tessellation is also old tech, but there is a reason why artists choose to do it manually. More control over the final outcome. I remember a directx 12 demo about dynamic tessellation but artists quickly realized that alot of the polygons being added while useful for displacement maps they weren't really adding any extra detail that couldn't be done with LODs that are controllable.

Link to post
Share on other sites

Does someone have access to good photoscanned 3D models, so he can place them into our engine? Jeff posted a shot in his work blog and it looks quite nice already: https://forums.torque3d.org/viewtopic.php?f=8&t=159&start=450#p12831

Only a direct comparison of the same scene can really show a game engines capabilities. Such direct comparisons already exist for the major engines and Unreal always seemed to me like one of the less realistic looking engines, I wonder how Torque would compare to that.

Link to post
Share on other sites

engine demos are sometimes freely available to test and play around with, the assets might be able to be converted over to torque, but if its all dynamically created i don't know even where to begin to take it into torque. And from their website unreal 5 wont be out until 2021, waiting that long sounds boring lol I wonder if we could work together to design a similar tech demo. It will take time but imagine an open source engine being able to re-create the unreal demo before it is released to the public. Could get a lot of attention but again it would have to look the part. With the work JeffR and the steering committee is doing it is very possible to make it look as realistic if not more realistic but it all falls on the artwork created for it. Does anyone here have Zbrush? Torque4.0 pre release has directx12 if im not mistaken? Because torque is open source we could create a demo similar for non profit educational means without causing any problems. And release it for people to play around with.

Link to post
Share on other sites

@marauder2k9

You are dreaming, we did not manage to put out any demo material as a community for promotion so far, how do you think we should be able to do it now? We did not even manage to create a steering committee, so we not only have nobody doing anything, we do not even have people who could coordinate doing something to begin with.

Link to post
Share on other sites

@marauder2k9

You are dreaming, we did not manage to put out any demo material as a community for promotion so far, how do you think we should be able to do it now? We did not even manage to create a steering committee, so we not only have nobody doing anything, we do not even have people who could coordinate doing something to begin with.

 

nobody has made a demo because there is nothing to make a demo of. there hasnt been any major updates to the engine since 3.1 with 4.0 being the next major update there would be demos made for that to be sure. But 4.0 isnt released yet and is a bit unstable at times so no one wants to make a demo for it yet apart from maybe the steering committee, but they are making simple things to show off certain features no fully fledged demo yet but there will be. I'm merely suggesting ideas that i wouldn't mind helping on but i wont be able to do it alone. A demo similar to that wouldnt be too hard or take too long to make when you think about it. Ive often been thinking a viking style demo would be amazing for torque white cliffs grassy fields a temple in the mountain etc

Link to post
Share on other sites

Did you not read? I said we do not have a steering committee doing things, which is also the reason why we did not have any releases for 2 years, while originally we had 4 releases per year. And of course there is things to show off, there are constantly new features being implemented, why not show them off?

Link to post
Share on other sites

Let's steer back on topic shall we. . .


Tech demos such as these are in essence a flex and often are never intended to suggest how things should be done. developing a framework that can handle millions of polygons and massive textures with dynamic scaling without a noticeable performance hit is impressive. Full stop. The best way to showcase it is to push it millions of polygons and massive textures. To me, the point of them demo seemed to be letting the engine do more and more of the heavy lifting. Dynamic global illumination, cave echos, flocking "boids," IK animations; it's largely taken off the plate of the artists and level designers and put onto the shoulders of the engine. Of course it won't be perfect and changing the industry standards will take time (if it ever happens at all).


@Duion is probably right about a Torque showcase. There are things that Torque can do well and should show those things off. Being "small" or unique is immaterial. Having a good pool of art that is at least visually comparable to other engines is important, but I don't know if trying to prove to be Unreal's contemporary is a good idea.

Link to post
Share on other sites

Let's steer back on topic shall we. . .


Tech demos such as these are in essence a flex and often are never intended to suggest how things should be done. developing a framework that can handle millions of polygons and massive textures with dynamic scaling without a noticeable performance hit is impressive. Full stop. The best way to showcase it is to push it millions of polygons and massive textures. To me, the point of them demo seemed to be letting the engine do more and more of the heavy lifting. Dynamic global illumination, cave echos, flocking "boids," IK animations; it's largely taken off the plate of the artists and level designers and put onto the shoulders of the engine. Of course it won't be perfect and changing the industry standards will take time (if it ever happens at all).


@Duion is probably right about a Torque showcase. There are things that Torque can do well and should show those things off. Being "small" or unique is immaterial. Having a good pool of art that is at least visually comparable to other engines is important, but I don't know if trying to prove to be Unreal's contemporary is a good idea.

 

You are right, but i will say my idea was not to prove to be unreals contemporary more along the lines of just showing that it can be done in torque

Link to post
Share on other sites
  • 2 weeks later...

I've got my own impressions of how they're doing the tech(some of which more nerdly technical analysis people online look to generally align with, such as Digital Foundry). It's undeniably cool tech, but people tend to forget in the moment of a glitzy demonstration that literally everything has compromises and downsides. From the looks, the 'nanite' tech can push an incredible amount of tris to the GPU efficiently, which is awesome. But it looks to only work on un-animated objects(as in no bone animations, i'm sure you could move around the objects like any instanced piece), and unless the artist is slapping procedural generation noise on everything, or you're doing photogrammetry, the likelyhood of anyone but the AAA of the AAA is going to dedicate thousands in man-hours to make art THAT detailed everywhere(presuming the game disk sizes don't, predictably, balloon yet again as well) is blindingly low.

It's primary advantage is ultimately dropping photogrammetry or special set-piece art in and just having the art pipeline figure it out without extra steps like normal map baking.

For the huge majority of art assets and projects, it's not going to offer anything regular content pipelines do, with perhaps a few more steps in the creation tools.

All you need to do is look at other games that already used photogrammetry for their environment geometry like Battlefront 2, or compare AAA character models to their demo character to see that it really isn't some space magic 'this revolutionizes game art as we know it' so much as a very, very slick art pipeline for special-case art that most development teams won't have the budget to really utilize effectively.


Still, it'll be interesting to see the tech trickle into the knowledge-sphere of gamedev as a whole for integration down the line, and see if art tools do anything to make that level of art any easier.

Link to post
Share on other sites

They say somewhere that they bring the billions and trillions of polys down to 20 million that are actually rendered in the scene and 20 million is very possible to render, I already did a scene with 20 million polys in Torque and there are some games that already push close to that number.


Regarding if creating assets that detailed is harder I would say yes and no, no because a good artist usually starts with high poly and then reduces it for the final game and yes because to create photoscanned objects you need new hardware and software and put more work in there than before. It can take hundreds of photos to create a photoscanned art asset and later it has to be optimized by hand again, well in case of Epics super tech maybe that step can be skipped. The problem is still that creating photoscanned assets you are limited to things that exist in the real world for real and if you want to scan larger objects you need aircrafts to scan them.


The game engine market is a bit of a scam anyway, since indies will not be able to create such high quality, only if they use the big proprietary engines, with other proprietary software that already has pre-build assets that cost extra money in most cases and even then, big studios can afford much more of those.

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...