Rage (id Software)

We'll have to wait for Rage to come out and see if megatexture is indeed a better solution compared to other streaming systems out there. U2 has repetition but the texture artwork is so good that you hardly notice while playing.

I'm sure that if id Tech 5 is the technically better solution other studios will license it. Personally I'd love to see a Fallout game using this technique.

I'd also like to know if megatexture causes extra wear on the DVD/Blu-Ray drive compared to other engines.
 
As a real technical noob, sometimes I don't quite understand everything some of the senior members are saying.

However, I don't quite understand why people are already bashing this tech, saying it is not really anything new. Well, from what I have seen, it looks better than anything out there. On the PC, which I imagine is running the code on a lot of the demo's shown(http://goo.gl/1pRg), it looks great up close. But really how good do you want it to look?

Anyway, the steps they are taking in tech is necessary for the industry to keep moving forward in tech. They could just as easily not do it, but where would we end up?
I hope some other developers use the tech and open up avenues with it even Carmack didn't think of.

By the way, 60 FPS.
 
As a real technical noob, sometimes I don't quite understand everything some of the senior members are saying.

However, I don't quite understand why people are already bashing this tech, saying it is not really anything new. Well, from what I have seen, it looks better than anything out there. On the PC, which I imagine is running the code on a lot of the demo's shown(http://goo.gl/1pRg), it looks great up close. But really how good do you want it to look?

Anyway, the steps they are taking in tech is necessary for the industry to keep moving forward in tech. They could just as easily not do it, but where would we end up?
I hope some other developers use the tech and open up avenues with it even Carmack didn't think of.

By the way, 60 FPS.

Normal people and tech junkies have a different view on what "looks" good. :) Normal people just look at something and think it looks good or it doesn't.

Tech junkies look at something and if it doesn't have the latest and greatest tech, it doesn't "look" good.

Just different ways of looking at things. It's like music appreciation. Most people like music that sounds good. Music afficianado's don't think music sounds good unless it's also "technically" good or something that would be difficult to play.

Which is why you have arguments that Baked shadows make a game look horrible even though depending on the game, baked shadows look just as good if not better than dynamic shadows, especially if the lighting doesn't change. But then you get into arguments about static lighting versus dynamic lighting. Blah blah blah. At the end of the day most of us don't care if it's dynamic shadows or baked shadows, as long as the game looks good.

But here's the kicker. You're on a mostly technical site now, so there's going to be a lot more discussion about the tech involved than just whether a game looks good or not. :)

Regards,
SB
 
You can do that to all textures - the only advantage is that you have 1 "texture" that the game automatically composes of tons of tiny ones. The only thing you save is artists having to break down the world(-geometry) in small enough tiles to fit texture boundaries - but thats something the toolset could do automatically.
Streaming in unique textures is nothing new.

Alucard and Laa-Yosh already mentioned some of the advantages but here's two others: that saving of artists having to break down the world geometry isn't just a benefit to artists: that means there's less "invisible" polygons to handle at run-time. Invisible polygons because they do not add extra geometry and are only there to establish the UV coordinate boundaries.

Then because everything is using a single texture, you could in theory load everything in a single batch and lower the CPU overhead dramatically. I say in theory because you'll still want to check for occlusion of detailed geometry pieces and dynamic lights, etc. but otherwise, there's no extra draw calls due to "oops, have to load a new texture for this surface".

Then there's the other part of the graphics equation people often forget. Let's say for argument's sake that Uncharted 2, RDD, etc. do look as good or even slightly better than Rage: do they run as fast? Maybe Rage's graphics ought to be compared to other games' that run at 60fps locked.
 
Normal people and tech junkies have a different view on what "looks" good. :) Normal people just look at something and think it looks good or it doesn't.

Tech junkies look at something and if it doesn't have the latest and greatest tech, it doesn't "look" good.

Just different ways of looking at things. It's like music appreciation. Most people like music that sounds good. Music afficianado's don't think music sounds good unless it's also "technically" good or something that would be difficult to play.

Which is why you have arguments that Baked shadows make a game look horrible even though depending on the game, baked shadows look just as good if not better than dynamic shadows, especially if the lighting doesn't change. But then you get into arguments about static lighting versus dynamic lighting. Blah blah blah. At the end of the day most of us don't care if it's dynamic shadows or baked shadows, as long as the game looks good.

But here's the kicker. You're on a mostly technical site now, so there's going to be a lot more discussion about the tech involved than just whether a game looks good or not. :)

Regards,
SB

Well some of the more knowledgeable people are just jiving on the game because it doesn't look so good up close (to them). But they are not offering any arguments on how this could be fixed, if possible. So they are in effect doing the same thing the noobs are, except they know what a texel is.
 
IMO, those who are inclined to rag on Megatexture do so because some say Rage is the "best looking game on consoles", which puts certain people in defensive mode. They are just "defending the faith", so to speak.

Richard's point is well worth considering. Rage is a 60 Hz game, and it looks sexy.
 
U2 has repetition but the texture artwork is so good that you hardly notice while playing.

U2 is a linear, relatively confined game; and there's significant texture repetition (e.g. I remember a scene where you're stalking/stealthing some guards in a large hall with stairways and columns, near the beginning of the game - was that a museum? - which was essentially the same handful of textures everywhere).

I'm sure that if id Tech 5 is the technically better solution other studios will license it.

Yes, that would be the proof of the pudding, to some extent. However, here we're getting out of technical and artistic into business territory. If idTech5's message is "we let you spend 50 mln on texturing alone", who will bite?

I'd also like to know if megatexture causes extra wear on the DVD/Blu-Ray drive compared to other engines.

Probably lower than equivalent non-megatexture tech streaming the same world; after all, significant work has been spent baking all the artists' contribution to a single texture.

Then there's the other part of the graphics equation people often forget. Let's say for argument's sake that Uncharted 2, RDD, etc. do look as good or even slightly better than Rage: do they run as fast? Maybe Rage's graphics ought to be compared to other games' that run at 60fps locked.

I think this is very hard, almost impossible to explain to non-technical people, or even to technical people who haven't stared at milliseconds in the profiler with despair: how much harder 60 fps is than 30 fps, and how much you have to "discount" the visual quality because of it. So hurry up and bring the 30 fps Doom to the genitalia-measuring table!
 
Supposedly, Id promised some Doom for Quakecon...

As for Rage, it keeps winning various "best game" awards from E3; Gametrailers has just given it the PC, X360, PS3, and overall categories. IGN has already decided as far as I know, and there are some more still pending. And these journalists have all seen the game in person, not just on a few screenshots or compressed videos, so it seems to be quite convincing.
Looks like Id has a good chance to surpass Doom3's success.
 
Well some of the more knowledgeable people are just jiving on the game because it doesn't look so good up close (to them). But they are not offering any arguments on how this could be fixed, if possible. So they are in effect doing the same thing the noobs are, except they know what a texel is.

I don't think anything stops a megatexture game from having sharp textures up close as far as the engine goes, but I think the storage requirements would increase so dramatically, perhaps 4x+ as much, that it's just not practical right now. I have ragged on megatexture a bit myself but only because of two reasons. Time of day support is critical to me so that alone would rule out use of Rage at this time. More importantly though is that you do not want to make 25gb+ games, anyone that does will cry once digital downloads become the norm on console. Fortunately few games need that kind of storage, but any game based on Rage does which also makes it really hard to consider using it.

Aside from that it looks pretty cool, and doing that at 60fps is killer. I'm always a fan of what Carmack does so i'll likely get the game one way or the other.



U2 is a linear, relatively confined game; and there's significant texture repetition (e.g. I remember a scene where you're stalking/stealthing some guards in a large hall with stairways and columns, near the beginning of the game - was that a museum? - which was essentially the same handful of textures everywhere).

Wow, you are a brave man posting that on a public forum :)
 
Er, I don't see anything stopping people from using the tech5 engine to do dynamic lighting and time of day. It's just probably not likely to work at 60fps. Especially because virtual texturing doesn't help at all with such a high refresh rate, in fact it's the other way: Carmack has to spend a lot of extra shader instructions compared to other engines because of the new tech.
Oh, and having a 1:1 resolution lightmap for "free" is just very tempting, compared to spending all the effort on dynamic lighting if the gameplay itself wouldn't really benefit from it.

Regarding the large downloads... Rage has a (supposedly) big open environment - but most corridor shooters would probably be able to work with far, far smaller game worlds and thus use a single 128k texture or even less. You can also break up your game into separate levels and each could use 64k or 32k texture atlases. Or, they can increase the texel density in the relatively smaller environments.

What I'm saying is that the megatexture tech doesn't determine that much about the type of games that could be implemented with it. Id probably chose to build a huge open world environment so that they can demonstrate the advantages of the engine to the full extent.
But I fully expect Doom 4 to be on the other end of the scale, at 30fps with dynamic lights and a much more detailed world (and enemies).
 
I have a feeling we need some direct feed videos where you can really see that 60fps smoothness that is really "game" changer and its visual fidelity because all we have is some low quality E3 footages which look like they run at ~20fps and that does not do RAGE justice.And what is all about bashing baked shadows and dismissing RAGE because of it?Assassins Creed 2 has all of them dynamic if Im not mistaking and i think they would be better of with baked shadows and without dynamic time of day,in the end very little people notice it but those are big performance hamper that would result in better looking game if it wasnt for them.KZ3 or something like Crysis 2 is another story...
 
I seem to remember Id said that dynamic lighting was a challenge that they simply would not be able to overcome for Rage, but expect to have in place for Doom?
 
I still think it's the 60fps, as it would allow for less then 50% shader complexity (to maintain a stable 60+ Hz refresh rate before vsync). With the extra code for texture lookups and filtering, they're probably unable to fit dynamic lighting in as well. Their materials are pretty simple as well, no fake translucency or fresnel stuff or parallax mapping or anything else.
 
Er, I don't see anything stopping people from using the tech5 engine to do dynamic lighting and time of day. It's just probably not likely to work at 60fps. Especially because virtual texturing doesn't help at all with such a high refresh rate, in fact it's the other way: Carmack has to spend a lot of extra shader instructions compared to other engines because of the new tech.

I'm curious to see if that will be the case, getting time of day support at 30fps. Still not totally convinced because of all the shader side work they have to do to support megatexture, but I guess we'll see. Have id ever done a game with time of day support? Maybe they don't consider it a must have feature.


Regarding the large downloads... Rage has a (supposedly) big open environment - but most corridor shooters would probably be able to work with far, far smaller game worlds and thus use a single 128k texture or even less. You can also break up your game into separate levels and each could use 64k or 32k texture atlases. Or, they can increase the texel density in the relatively smaller environments.

What I'm saying is that the megatexture tech doesn't determine that much about the type of games that could be implemented with it. Id probably chose to build a huge open world environment so that they can demonstrate the advantages of the engine to the full extent.
But I fully expect Doom 4 to be on the other end of the scale, at 30fps with dynamic lights and a much more detailed world (and enemies).

Hmm see I wonder about that. I wonder if they chose the big open world because in those types of games you can't get very close to the environment (except for the ground). So they can show how the canyon walls have tons of variety, and they don't have to worry about the player walking right up to a canyon wall since it's not possible. Compare that to an indoor scene for example where the player can go right up to just about everything, which would necessitate higher details everywhere. So I'm not totally sure that a predominantly indoor game would really need less texel space compared to a mixed indoor/outodor world.
 
It was probably both, but I remember an interview on this. They were pretty adamant to get it into the next Doom though for obvioud reasons.
 
I'm curious to see if that will be the case, getting time of day support at 30fps. Still not totally convinced because of all the shader side work they have to do to support megatexture, but I guess we'll see.

If the engine can already support a solid 60fps with all the extra shader work, then I'm sure that the added +100% performance from downgrading to 30fps will enable at least one dynamic light per surface ;)
Shadows should be completely independent of virtual texturing, too, so I don't see anything stopping them from implementing a completely different rendering architecture using megatexture.

Of course it would require a significant amount of modification to the Rage engine; but we don't know if Rage and tech5 are actually one and the same. Maybe Rage is a variation and Doom 4 is another; or maybe Rage is the basis and Doom 4 extends upon it? We'll probably learn more as soon as any license deals are announced. Although there aren't any, and id has pretty much stopped talking about the issue completely...

Have id ever done a game with time of day support? Maybe they don't consider it a must have feature.

Doom3 had fully dynamic lighting and they've probably learned that it's as much a problem as a completely static approach has been. But it can't really really be called time of day support yet.
Then again, I'm quite sure that Carmack could write any kind of lighting engine; the question really is whether he considers a particular solution to be necessary for the game they're making.

Doom 4 will probably follow Doom 3's direction: relatively smaller environments and small number of monsters. So they'll need dynamic lighting for the scare factor and they'll have the resources and performance for a higher level of detail and fidelity in everything.
As for time of day, it's probably not required for this kind of game either. As someone else has mentioned, it didn't add much to Assassin's Creed either, apart from the visuals of having nighttime for the Venice Carnival. But it was a full moon for every single night as far as I remember ;)

Hmm see I wonder about that. I wonder if they chose the big open world because in those types of games you can't get very close to the environment (except for the ground).

I dunno, there are many indoor levels as it seems (most missions), so the player can get close plenty of times. But outdoors are indeed an easy example to show the lack of repetition.

So I'm not totally sure that a predominantly indoor game would really need less texel space compared to a mixed indoor/outodor world.

Think about it as the relative sizes of surface areas. A 10Km x 10Km wasteland is more than likely about twice as big after unwrapping as a 50Km long set of corridors and rooms would be.
I'm not sure how long all the MW1/2 or Doom3 levels would be when stitched together - but I guess they're not that big. So a linear game could either ship with half or even a quarter of the data; or they could keep the large storage requirement and increase the texel density significantly.

Anyway, if they really show Doom 4 at Quakecon then we'll definitely know more about the potential of the engine. For now, I consider megatexture to be a single feature that doesn't determine anything else about the engine using it (also see Brink).
 
Back
Top