CliffyB: Epic Must "Drag" the Industry Into the Next Generation

[Dirty Hipsters]

So...is this iteration of the engine going to do away with the signature texture pop-ins that the Unreal Engine is known for, or are those a "feature?"

 

[fulano]

Every time someone says something stupid like "do we need a new generation SO SOON" in threads like these AFTER 8-9 YEARS I want to faceplant my head against a wall..

If you think 7 years is actually 8 or 9, I suppose you might be banging your head against a wall. You might want to stop, I don't think that's good for you.

The Xbox 360 launched in Fall 2005. That's six and a half years ago, 30% less than you claim. It was also only four years after the launch of the original Xbox, which came late that generation.

The PS2 came out in 2000. The PS3 came out in 2006, although the PS2 continued to sell strongly for quite awhile after that (unlike the original Xbox).

The last cycle was shorter than usual for Microsoft because they came last. They launched earlier than Sony this generation, and are more inclined to milk the generation as long as they can, as profits come in the end, not in the beginning.

We're barely overdue, if we are. And if we are "overdue" it is more likely an indication that at the console price point, we're approaching the limit of diminishing returns in terms of what can usefully be produced and sold for a $60 price point. Go ahead and push the next generation to start now and we'll likely see a move to a standard price point of $70 or higher.

Pretty much. You nailed it on the head.

The one good thing to look forward to is that, at leas on PC, Mac, or Linux, indie gaming is indeed taking off and thus we, the consumers, can focus on actually looking for decent games instead of over priced, overproduced messes that sell over sixty U.S Dollars. And since their graphic capabilities are rather low-end in comparison we can get a product whose focus is on gameplay b/c it is actually cheaper to do it that way.

I'm sure Halo 6 is gonna look very pretty and all, but if it is expensive as fuck I'm not buying it. The same goes for anything Sony publishes like The Last of Us 2X Reloaded Sigma, etc.

 

[Akisa]

Go ahead and push the next generation to start now and we'll likely see a move to a standard price point of $70 or higher.

Hopefully PC games will still 50 dollars and people will migrate elsewhere ^^.

 

[Sylocat]

I saw the title and thought, "Probably an out-of-context quote in the title, blowing things out of proportion."

Then I clicked and read, and I facepalmed.

Every time I think this guy can't get more cartoonishly egomaniacal, he surprises me.

And... how do they plan on keeping game budgets from skyrocketing out of control?

I think we should solve that problem before we tackle making games look even more graphically advanced.

This is how. The vast majority of a time a high poly, cinematic quality model is made first (some vehicles and buildings usually the exceptions) which is then baked onto a low poly mesh (which takes time to build in its own right) to create the normal map. With more advanced game engines and hardware you completely cut out the low poly mesh and constraints and in turn save time there. Additionally higher resolution textures are much more easier to make due to a lack of any constraints (speaking from experience here)

Um, no. You're not going to get rid of poly-count in real-time rendering (at least not within the next 10 years), so you'll still need a low-poly mesh, regardless of how many actual polys you're able to put in it.

You may have a point about the textures, but with each generation dev costs have increased exponentially.

Every time someone says something stupid like "do we need a new generation SO SOON" in threads like these AFTER 8-9 YEARS I want to faceplant my head against a wall.

As opposed to "Do we need a new generation in the middle of an economic crisis when the spending power of the average consumer has been shot to hell?"

 

[Baresark]

I agree, it's way past time for consoles to stop holding back PC's. It's a pain in the ass watching games fail on multiple levels because the PS3 or 360 can't handle all of it.

All that said, Cliffy B should be banned from speaking in public, and also from making games since he doesn't really know how to make more than Gears of War. When he opens his mouth, the human race looks much more stupid. As evidence, in this one article he said, "...our studio to drag this industry into the next generation", and "because I want it and gamers want it-even if they don't know they want it." He needs to get his head of his ass because the last time I checked, while awesome graphics are always a plus, they don't make games. And check the ego on this guy, Cliffy B, Sweeney and Epic are the most important thing in gaming. Only, they aren't. They built and awesome engine, no one can doubt that. But gaming can certainly do well without folks like him telling gamers and developers what they want.

 

[Saulkar]

I saw the title and thought, "Probably an out-of-context quote in the title, blowing things out of proportion."

Then I clicked and read, and I facepalmed.

Every time I think this guy can't get more cartoonishly egomaniacal, he surprises me.

And... how do they plan on keeping game budgets from skyrocketing out of control?

I think we should solve that problem before we tackle making games look even more graphically advanced.

This is how. The vast majority of a time a high poly, cinematic quality model is made first (some vehicles and buildings usually the exceptions) which is then baked onto a low poly mesh (which takes time to build in its own right) to create the normal map. With more advanced game engines and hardware you completely cut out the low poly mesh and constraints and in turn save time there. Additionally higher resolution textures are much more easier to make due to a lack of any constraints (speaking from experience here)

Um, no. You're not going to get rid of poly-count in real-time rendering (at least not within the next 10 years), so you'll still need a low-poly mesh, regardless of how many actual polys you're able to put in it.

You may have a point about the textures, but with each generation dev costs have increased exponentially.

Every time someone says something stupid like "do we need a new generation SO SOON" in threads like these AFTER 8-9 YEARS I want to faceplant my head against a wall.

As opposed to "Do we need a new generation in the middle of an economic crisis when the spending power of the average consumer has been shot to hell?"

Ok, I fucked the wording on that... a lot. What I meant to say is that you bypass... you know what I will just compare work pipelines. (Many artists/companies have their own unique workflow but this is still a competently accurate one)

Right Now.
1.Low Poly Model
2.High Poly Model
3.Retopologising High Poly into Low Poly
4.Baking High Poly onto Low Poly
5.Texturing Normal Mapped Low Poly Model

What can speed things up.
1.Low Poly Model
2.High Poly Model
3.Apply Tessellation Map from High Poly onto Low Poly
4.Texture High Poly Model
5.?
6.Profit!

You only cut out one step but the pipeline is far more streamlined and efficient and you change up a couple and in turn speed things up. You use the same low poly and high poly models without the need to retopologise a new low poly model (which is what I meant) and applying a tessellation displacement map is not without its challenges but is sometimes faster than baking a quality normal map.(depending on the tools and skill of the individual) I do this myself on the amateur level.

This in ten minutes (not including modeling or sculpting time)

Spoiler

 

[Reincarnatedwolfgod]

well he talks a big game about unreal engine 4 but i suspect he is full of shit
i kinda have a feeling he is just stroking his massive ego with claims like this but
then again i have that feeling every time he opens his mouth

 

[VoidWanderer]

I wonder if Cliffy B would be interested in paying rent for his massive ego, of would at least see a psychiatrist for his delusions of greandeur, that should solve the American Financial Crisis quite nicely.

 

[userwhoquitthesite]

CliffyB's lips are in need of a few staples.
Forgetting the massive ego in the statement for a moment, let's consider that Gears of War, Epic's biggest moneymaker, LOOKS LIKE ASS. And if he simply means pushing the limits of game mechanics, well, that's back to the arrogance. Particularly because Gears of War also is written, sounds, and plays like ass.

Didn't this guy used to be a gaming hero?

 

[ThriKreen]

Sigh... you'd think people here would also pick up that Epic would also respond on feedback from their various UE3 licenses and improve the editor side of things too.

There's more to UE4 than graphics, and people in this thread are missing a lot of what he's saying.

For instance, texture popping: UE3 supports texture streaming to lessen how long the game is displaying a loading screen, so it can load the textures on the fly as you move throughout the level. If it does happen, it is a fault of the DEVELOPER for overloading the level with a lot of high resolution textures and models, not giving it enough time to load what it needs (especially off the slower optical disc instead of HDD). It is especially prevalent if you immediately start the player in a highly detailed starting area, instead of a more bland one and take advantage of the streaming system to load more in the background.

What Cliff is implying about dragging the industry into the next generation, is improving the hardware with things like mandatory HDD for game installs, that way it speeds up load times vs. loading off the much, much slower optical disc. And larger RAM capacity in both system and video RAM, to reduce to need to have a lot of level streaming volumes so the engine can load and swap textures and portions of the level as you play.

Another thing about improved graphics and speeding up development time is that with a much more powerful engine and hardware backing it up, you can spend less time having to adjust lighting on a level. You could switch to dynamic lighting completely, doing it on the fly and not have to make multiple passes of the level in whatever scenarios are required, baking different lighting and shadow maps, and wasting storage space for said shadow maps. Heck, not having shadow maps can speed up aforementioned texture popping too, as it is one less thing to load.

If you've ever mucked with the UDK, one of the most boring, lengthy things one can do is to hit the "Build Lighting" option then walk away for a drink, food, play another game, raise a kid from birth to graduation, then it's done.

The level designer or artist can make the level, place a handful of lights, instead of dozens of lights to handle all the main, reflected, and back lights, and hit play right away. Speed up iteration process, less time and money wasted on development time.

You'd still need render meshes over skipping them and just using the detail mesh though, a good artist can fake a lot using decent textures, but textures are much larger than mesh information.

And they were just talking graphics! I can't wait til more information on the editor shows up, and if there is any news on a UDK4 release!

 

[Grey Day for Elcia]

I saw Cliffy's name and lost all interest. The guy remade Unreal over and over, then went on to make the same chest high wall simulator three times. His opinion has about as much importance to me as that of the brown and grey walls he loves.

 

[Deadyawn]

I'd just like to throw in my contempt for cliffy here. I don't really have anything new to say as everyone seems to have beaten me to the punch.
I don't want better graphics and I don't care what you say. I'd rather we concentrate on improving the weak areas of games rather than just putting more effort into redundant prettyness.
I simply don't see how graphical fidelity is important. I mean, you look at deus ex. It's often cited as one the greatest games of all time and its also really goddamn ugly. And not just graphically ugly but aesthtically too. Compare that to, say, Crysis. It was extremely tech heavy, really pushing graphical boundaries but it wasn't all that fun. Not bad, just mediocre. How does this happen? Shouldn't games be progressively growing in quality? No they aren't because developers are focusing on graphics instead of substance and that sucks.
It also sucks because developing games for all this new hardware is even more expensive exarcerbating the problem (and also driving up prices).
So how about you make a good game before you decide to push graphics forward again?

 

[Chairman Miaow]

"It needs to be a quantum leap."

People keep using that phrase... I do not think it means what they think it means...

Epic will drag the industry into the next generation, with gamers kicking and screaming, but like Gears of War 4, it's going to happen...

It's become so commonplace for people to say that, I had never considered the actual definition, it's rather hilarious when you think about it.

 

[weirdee]

I'd just like to remind people that this is the same person who thought that Fatal Frame - a horror game - needed a multiplayer aspect.

We can handle the next generation just fine without him.

What, like co-op? Wasn't there twin sisters or something like that? It's not like he said it needed team deathmatch or something.

You do understand that merely adding a second person would kill the horror? A major aspect is the feeling of being all alone. It's why teaming you up with an AI partner in horror games is also almost universally stupid.

On a side note, I believe two player horror games ARE possible, if you use those two players properly, instead of treating it like a one player game with an extra person riding in a sidecar. Let's remember that two player games are of their very nature a delicate relationship to begin with because it's easy to turn it into something competitive or downright enraging when you have a teammate who doesn't play nice and you are given the capacity to mess with each other like in say, Super Mario Bros Wii (google some vids if you need context). Playing those people off of each other to amp the terror is a tricky, yet very tempting prospect due to those interactions. With the right types of cues and misdirection you could send those two players down a path of insanity, paranoia, dementia, and all sorts of hilarious chaos (hilarious to the outside viewer) that you wouldn't be able to do with one person. Just imagine twice the amount of girly screams you see in Amnesia play vids, but caused by each other's mortal faults twisting each other into their own nightmares.

Okay, maybe that's not a good idea, I don't want to end up accidentally killing anybody or scarring them for life. Just keep in mind that you, and perhaps a lot of developers and film makers who play at creating horror but don't grasp the concepts, are underestimating the human psyche's capacity for torment.

 

[weirdee]

Back on topic, considering that real life has hundreds upon thousands of independent/codependent variables upon which we only grasp the concept of perhaps a few hundred or such at all, much less more than a handful at any given time, the idea of trying to understand how to manipulate or design something like that that isn't already just real life (we already have trouble dealing with three dimensional space with tools that only operate in two dimensions at any given time, and it's still only barely tolerable with all of the extra software that has been created relatively recently), much less simulate it on that level IN FUCKING REAL TIME, is pretty much a huge waste of time when the average person will find it impossible to appreciate that level of detail (can you even tell that your computer runs on individual frames rather than a dynamic image without having been told first?), when you could instead just work on a few points of easily grasped perception to the point where they become precision tools that your audience will focus on, rather than try to do everything and either burn them out on the matter, or end up not making a strong impression on anything.

Wow, my run-on sentences are getting way out of control.

 

[Strazdas]

It's not MUST it's want to. And why do we need another generation so soon anyway?

The next generation of consoles have been due for the last.... 5 years? its not soon, its late, way too late already, and its amazing how so many people stil clinge on the current consoles. the limitation in them is HUGE, a good exmaple is DUST514, where creators had to cut amount fo eniromets 8 times because "PS3 didnt ahve enough memory to render it", and therefore we are left with bland "normal planet" look. instead we could b walking on gas giants, flaming lava planets, ect, but consoles cant handle that.

Things like SpeedTree make plant development easier,

dont know whether it is my problem, but whenever the plants in game is developed by speedtree, my FPS drops drastically around trees and bushes, so i dont endorse this method at all.

 

[weirdee]

Hey, if you want to be doing those kinds of things like exploring the universe, start developing for the PC customers who actually have bought the cards capable of doing that next level crap, not something whose strangely specific limitations actually made the "high definition" remake of silent hill 2 to look several times shittier because the ps3 couldn't work with the same types of visual effects in the original.

i mean it's just fog for pete's sake

 

[Dryk]

People who work at a company that makes all its money licensing a series of engines that push tech want people to give them more powerful tech to work with, not seeming to care that pushing graphics is part of what's inflating budgets and ruining AAA gaming. What a scoop.

Every time someone talks about rendering Avatar in real time I just have to point and laugh. Yes, you're really going to be able to render Avatar, a film that took 40 000 processing cores, and each frame took about 30 hours to render.

I'm sorry, but I have to call bullshit. You will never have realtime film level CG in a video game.

Then again, I really shouldn't be surprised when CliffyB spouts off bullshit anymore.

That was my first thought too, if Cliff wants the next generation of consoles to be orders of magnitude more powerful than current supercomputers he needs to take a step back.

"It needs to be a quantum leap."

People keep using that phrase... I do not think it means what they think it means...

Epic will drag the industry into the next generation, with gamers kicking and screaming, but like Gears of War 4, it's going to happen...

Lets be honest, home consoles going from an Xbox 360 to the most powerful supercomputers ever would be a quantum leap