Digital Foundry Article Technical Discussion Archive [2011]

Status
Not open for further replies.
Do you have any video of that? I'd like to see how dynamic their GI is.
This is all I could find for you.
http://www.youtube.com/watch?v=iYyKDUKsR8Y&playnext=1&list=PL146E23837907AA56

Renting it maybe your best way of analyzing the lighting.

You keep saying this but, how many sacrifices did Crytek actually made? Slight subHD, temporal AA and draw distance for some objects/effects. Not only that, but how do you assume it was the realtime GI that caused those trade-offs?
Yes. Sub-HD, temporal AA, draw distance, texture and geometry pop-in, animation, audio, effects, etc. I guess that's not a lot to you. To me, that's way too much to sacrifice or to be poor in to have great lighting. Those things directly effect gameplay and the IQ.

Because it doesn't need to run in realtime LOL It's like asking why Pixar doesn't use MLAA :LOL:
Hmmm, there are gaming doing it in realtime as I type this. It's not at all like asking why Pixar doesn't use MLAA. They could use it, but it's inferior. Therefore, they pass on it. The same applies to Avatar. That's why James Cameron passed on what Crytech is using. ;)
 
Yap...1ms for real time GI is hardly trade of.Same goes for their SSAO,also 1 ms.
Then, why are all these other techniques mentioned are average or below average? If those techniques require so little processing power, Crytech is leaving a lot of power on the table. It has to be one way or the other. Either sacrifices were made or processing power was left on the table. It can't be both, can it?
 
Yes. Sub-HD, temporal AA, draw distance, texture and geometry pop-in, animation, audio, effects, etc. I guess that's not a lot to you. To me, that's way too much to sacrifice or to be poor in to have great lighting. Those things directly effect gameplay and the IQ.

I'm going to assume you didn't play the demo because none of what you describe effected my game playing.

I also think you greatly exaggerate what is poor.

Hmmm, there are gaming doing it in realtime as I type this. It's not at all like asking why Pixar doesn't use MLAA. They could use it, but it's inferior. Therefore, they pass on it. The same applies to Avatar. That's why James Cameron passed on what Crytech is using. ;)

I believe his point was there isn't any point in using performance saving techniques when you're not rendering in real time. Offline rendering can afford to go balls to the walls.

What you're saying here makes no sense.

Then, why are all these other techniques mentioned are average or below average? If those techniques require so little processing power, Crytech is leaving a lot of power on the table. It has to be one way or the other. Either sacrifices were made or processing power was left on the table. It can't be both, can it?

Below average compared to what? If they can iron out some small issues, the game could be in the top tier of graphics for a console shooter. I thought the game was gorgeous even in the demo.

Also, things can be scaled back for reasons other than processing power, memory is likely as big (if not bigger) of a factor for them.

Why not wait until the final game is released before assuming so much about it? If the past Crysis games are anything to go by, the campaign should look a whole lot better than the MP.
 
This is all I could find for you.
http://www.youtube.com/watch?v=iYyKDUKsR8Y&playnext=1&list=PL146E23837907AA56

Renting it maybe your best way of analyzing the lighting.
There are no changes in lighting in the video. Static GI is nothing new.

animation, audio, effects, etc.
Animation is pretty decent, specially for a game with such freedom of movement. Audio sounds pretty good too. What's wrong with the effects?

Hmmm, there are gaming doing it in realtime as I type this. It's not at all like asking why Pixar doesn't use MLAA. They could use it, but it's inferior. Therefore, they pass on it. The same applies to Avatar. That's why James Cameron passed on what Crytech is using. ;)
See, you answered you own question.

+1 to kagemaru's post.
 
Sub-HD, temporal AA, draw distance, texture and geometry pop-in, animation, audio, effects, etc

what? you act as if the demo was affected by this issue. they did not affect the experience in any way as it looked great and played excellent if you like that kind of game. AND this is not the final product nor is it Single player. Even the other map looks much better.

You appear to be pointing out things that you read or heard because I do not know how you can list those as game breaking mis-prioritizing of resources. especially when it's been pointed out that the lighting solution you are complaining about takes relative little resources.
 
There are no changes in lighting in the video. Static GI is nothing new.
There are no lighting changes in the Crysis 2 demo either, so let's not jump to conclusions here. Also it's a baseball game with realtime global illumination, which probably means it's dynamic, otherwise why even bother?
 
Yes. Sub-HD, temporal AA, draw distance, texture and geometry pop-in, animation, audio, effects, etc. I guess that's not a lot to you. To me, that's way too much to sacrifice or to be poor in to have great lighting. Those things directly effect gameplay and the IQ.

Draw distance aint that low to begin with and singleplayer footage has shown far larger draw distances. Audio effects are great, DSP effects used are quite fleshed out and well done so I guess your sound setup must be faulty or low quality. Also Pier 17 is the better sound showcase experience with more ambient sounds.

Pop-in though is quite brutal but I wonder how much game being on HDD affects this as HDD is very unpredictable vs fixed data layout on disc. This might affect timings and thus create delayed streaming.

Hmmm, there are gaming doing it in realtime as I type this. It's not at all like asking why Pixar doesn't use MLAA. They could use it, but it's inferior. Therefore, they pass on it. The same applies to Avatar. That's why James Cameron passed on what Crytech is using. ;)

You mean like how other devs sacrifice perfomance to have HDR vs games like KZ2/KZ3 that has LDR?
 
Then, why are all these other techniques mentioned are average or below average? If those techniques require so little processing power, Crytech is leaving a lot of power on the table. It has to be one way or the other. Either sacrifices were made or processing power was left on the table. It can't be both, can it?

Are you seriously saying that except a few effects the rest is mediocre or below standards in Crysis 2 and jsut saying lighting is the only thing that sticks out?

Things like shadow filtering and quality, material shading, mapping, complexity, soft material rendering, SSS, glare effects, lensflare effects, specular quality, godray rendering, water etc etc?

To me it seems you are using PC standards to compare vs Crysis 2, the outcome is brutal and given. But news is PS3 isn't PC standards, not even close and PC screenshot thread will lecture you in that. While you put the effects I mentioned above as mediocre/low standard they are for console very high standards. Else if we going to have such high standards we could pick out lots of things in games like KZ2/3, U2 etc... lots.
 
Then, why are all these other techniques mentioned are average or below average? If those techniques require so little processing power, Crytech is leaving a lot of power on the table. It has to be one way or the other. Either sacrifices were made or processing power was left on the table. It can't be both, can it?
I thought that demo had fantastic motion blur implantation for example,possibly the best i have seen on consoles(along with GOW III).Textures also seemed very good,albeit filtrating could be notch higher.

All Im saying is that we are looking at one map,from not yet finished game and it happens to be the map that is by far the worst representative of the games visuals and we have yet to see SP which was reported a couple months ago to look considerably better than MP version of the game.Particles also ain't bad,they are no Reach level but neither are KZ3s.
 
There are no lighting changes in the Crysis 2 demo either, so let's not jump to conclusions here. Also it's a baseball game with realtime global illumination, which probably means it's dynamic, otherwise why even bother?
It's shading the characters in realtime, that's not contested. Just how does the system aquire the lighting information, precalculations or at runtime.
 
I'm going to assume you didn't play the demo because none of what you describe effected my game playing.

I also think you greatly exaggerate what is poor.
Assuming I didn't play the demo or or didn't see the demo being played means you think they were made up. If these issues do exist, how does physically having the controller in my hand change what I see? Another way of saying this would be as follows. If I can see issues happening, holding a controller changes nothing. This is why videos are such strong evidence.

I believe his point was there isn't any point in using performance saving techniques when you're not rendering in real time. Offline rendering can afford to go balls to the walls.

What you're saying here makes no sense.
I know exactly what his point was. It seems you just couldn't absorb my point. I will break it down for you, so it makes more sense to you. Other games are using the very same technique he says is too expensive for console games. That means his explanation does not absolve him from my original statement/point.

Below average compared to what? If they can iron out some small issues, the game could be in the top tier of graphics for a console shooter. I thought the game was gorgeous even in the demo.
It's below average compared to 1st party exclusive offerings. Do you think you could logically argue how Crysis 2's animations, texture filtering, physics, streaming, dynamic lights, AA, geometry, characters on screen, draw distance, gun models, etc are near Killzone 3's? If so, I would love to hear it. Of course, Killzone 2 or 3's lighting isn't near Crysis 2, but that's mainly it. I'm only talking about the MP of each game. Since you wanted to know "compared to what", you have it.

If you just compared each system/feature separately, do you think Crysis 2's animation is different from the first round of launch games on consoles? If so, tell me about it. The same goes for the 6 vs. 6 MP. Please come up with something better than the "6 vs. 6 is the perfect MP gaming size" line from years ago. The same goes for the AA in Crysis 2.

Also, things can be scaled back for reasons other than processing power, memory is likely as big (if not bigger) of a factor for them.
Whatever the technical issue, memory or processing power, all these other systems/features are scaled back to such a degree in a lot of games. That's the point. If drastic scaling back of these systems/features wasn't for the lighting, what was it for?

Why not wait until the final game is released before assuming so much about it? If the past Crysis games are anything to go by, the campaign should look a whole lot better than the MP.
Hey, I'm just looking at the MP and comparing it to other MP experiences. That should be completely fine. I'm not comparing MP to other game's SP experience or anything. I tend to be fair about these things.

what? you act as if the demo was affected by this issue. they did not affect the experience in any way as it looked great and played excellent if you like that kind of game. AND this is not the final product nor is it Single player. Even the other map looks much better.

You appear to be pointing out things that you read or heard because I do not know how you can list those as game breaking mis-prioritizing of resources. especially when it's been pointed out that the lighting solution you are complaining about takes relative little resources.
When you have an enemy of in the distance behind geometry that hasn't popped-in yet, how do you see him/her? That's a graphical and gameplay issue, is not not? Unless the defintion of a gameplay issue has changed in the past couple decades, that would be a gameplay issue.

Why would I point out something I heard or read about, when there are plenty of X360s and videos around? How does that make sense? Again, if the lighting system takes so little resources, it strengthens my argument. It doesn't weaken it in the slightest.

There are no lighting changes in the Crysis 2 demo either, so let's not jump to conclusions here. Also it's a baseball game with realtime global illumination, which probably means it's dynamic, otherwise why even bother?
+1
 
I thought that demo had fantastic motion blur implantation for example,possibly the best i have seen on consoles(along with GOW III).Textures also seemed very good,albeit filtrating could be notch higher.
Fine. I say we should explore the motion blur implentation between Crysis 2 MP demo, Killzone 3 MP demo, and GoW3. I think the results of that comparison would be interesting!

All Im saying is that we are looking at one map,from not yet finished game and it happens to be the map that is by far the worst representative of the games visuals and we have yet to see SP which was reported a couple months ago to look considerably better than MP version of the game.Particles also ain't bad,they are no Reach level but neither are KZ3s.
Who's comparing Crysis 2's MP is anything other than other MP game experiences? I believe Reach particles were higher-rez than KZ3's, but are they anywhere near as plentiful? That could be considered a dead heat, or it could be broken down into 2 categories (higher rez and most particles). In that case, Reach would win one, and KZ3 would win the other. Another thing is that the physics on the colliding particles is not faked in KZ2 like it is in Reach (IDK about KZ3 on that one).
 
I know exactly what his point was. It seems you just couldn't absorb my point. I will break it down for you, so it makes more sense to you. Other games are using the very same technique he says is too expensive for console games. That means his explanation does not absolve him from my original statement/point.
Actually, you didn't know what my point was at all. My point is: it's all about trade-offs and they're dependent on the objectives and limitations developers have/encounter. Asking why avatar doesn't use Crytek's system was really silly.

It's below average compared to 1st party exclusive offerings.
And, why are you comparing them to 1st party exclusives? Obviously focusing in one system only has it's advantages, but then we could say that KZ3 is not impressive because it only runs in one platform whereas C2 runs on three.

It's below average compared to 1st party exclusive offerings. Do you think you could logically argue how Crysis 2's animations, texture filtering, physics, streaming, dynamic lights, AA, geometry, characters on screen, draw distance, gun models, etc are near Killzone 3's? If so, I would love to hear it. Of course, Killzone 2 or 3's lighting isn't near Crysis 2, but that's mainly it. I'm only talking about the MP of each game. Since you wanted to know "compared to what", you have it.
So, how about you describe objectively what's below average on those systems, specifically:

-Geometry
-Gun models
-Physics
-Dynamic Lights
-Animation

Numbers would be appreciated.

When you have an enemy of in the distance behind geometry that hasn't popped-in yet, how do you see him/her? That's a graphical and gameplay issue, is not not? Unless the defintion of a gameplay issue has changed in the past couple decades, that would be a gameplay issue.
Any videos of this actually happening? Crysis 2 popup problems are related to textures, shadows and small geometry, nothing as dramatic as you mention.

http://forum.beyond3d.com/showpost.php?p=1521041&postcount=1292
 
There are no lighting changes in the Crysis 2 demo either, so let's not jump to conclusions here. Also it's a baseball game with realtime global illumination, which probably means it's dynamic, otherwise why even bother?

The primary difference is that MLB's solution could be a mix of baked and realtime because they customized it to work on their game and their game only. Crytek on the other hand is going for a realtime approach because they want to dethrone Unreal as the next default middleware to be used. This means that their approach must work in a very general manner across all types of games to be of use as said middleware, including supporting time of day, realtime level changes for content creators, destructibility, 360/PS3/PC support, etc. That's what makes their approach so interesting because it has implications on many games beyond just their own. That's also why Crysis 2 on console is being so closely scrutinized not just by us on forums, but very likely by video game developers around the world.
 
Yeah, it's a very interesting question - can Cryengine replace Unreal on the current generation of consoles? Or are they going to have to duke it out on the next gen systems, where everything might change once again, knowing that Epic's been at work on their new iteration for at least 1-2 years by now?

And, yeah, my opinion on the first issue is that it's too late to challenge UE3 on the current consoles at this time. There's literally thousands of developers familiar with the engine, its capabilities and its toolset, as well as its weaknesses. Studios that have already licensed it have a large library of various assets from models and textures through shaders and animations to scripts and who knows what else, they're invested beyond the point of no return.
Whatever advantaged CE might offer, the costs are too high IMHO - the time to ramp up production would move most titles out of the launch window anyway, 2-3 years from now is far too far.

Now for the X3/PS4, a lot of the current code can be re-used while scaling up everything from image quality to GI fidelity and so on, so they have a better chance. But we don't know anything about how far Epic is with their stuff, and they already have the client relations and other bits in place, so it's still going to be quite interesting... ;)
 
Who's comparing Crysis 2's MP is anything other than other MP game experiences? I believe Reach particles were higher-rez than KZ3's, but are they anywhere near as plentiful? That could be considered a dead heat, or it could be broken down into 2 categories (higher rez and most particles). In that case, Reach would win one, and KZ3 would win the other. Another thing is that the physics on the colliding particles is not faked in KZ2 like it is in Reach (IDK about KZ3 on that one).

http://www.youtube.com/watch?v=LgjylRdDrIA

They look pretty plentiful in these first 10 seconds, although they dissipate quicker than they do in KZ2. How does Halo: Reach fake its particle collisions? The video offers a quick explanation, but I was hoping you knew a bit more.
 
Why would MLB09 go with a realtime solution if none of the lighting is dynamic? Might as well go for a higher quality offline solution if it's completely static. Plus, baseball games are definitely ideal candidates for time of day lighting changes.

It's shading the characters in realtime, that's not contested. Just how does the system aquire the lighting information, precalculations or at runtime.
The link Nebula posted suggests lighting calculations for the environment as well as the characters are performed at runtime. Specifically this part:
We then trace the sunlight’s path as it bounces off the ground and stadium. When sunlight hits a surface, it takes on that surface color, bounces off, and lights objects around it. We add that indirect light into the scene.
 
Do you think you could logically argue how Crysis 2's animations, texture filtering, physics, streaming, dynamic lights, AA, geometry, characters on screen, draw distance, gun models, etc are near Killzone 3's?

So it all makes sense now, I knew you've been trolling Crysis 2 for a reason, and now I'm done attempting to have a conversation with you.

BTW it's amusing how many of your points contradict each other in your post SMH

Yeah, it's a very interesting question - can Cryengine replace Unreal on the current generation of consoles? Or are they going to have to duke it out on the next gen systems, where everything might change once again, knowing that Epic's been at work on their new iteration for at least 1-2 years by now?

And, yeah, my opinion on the first issue is that it's too late to challenge UE3 on the current consoles at this time. There's literally thousands of developers familiar with the engine, its capabilities and its toolset, as well as its weaknesses. Studios that have already licensed it have a large library of various assets from models and textures through shaders and animations to scripts and who knows what else, they're invested beyond the point of no return.
Whatever advantaged CE might offer, the costs are too high IMHO - the time to ramp up production would move most titles out of the launch window anyway, 2-3 years from now is far too far.

Now for the X3/PS4, a lot of the current code can be re-used while scaling up everything from image quality to GI fidelity and so on, so they have a better chance. But we don't know anything about how far Epic is with their stuff, and they already have the client relations and other bits in place, so it's still going to be quite interesting... ;)

My thoughts are exactly the same
 
Last edited by a moderator:
http://www.youtube.com/watch?v=LgjylRdDrIA

They look pretty plentiful in these first 10 seconds, although they dissipate quicker than they do in KZ2. How does Halo: Reach fake its particle collisions? The video offers a quick explanation, but I was hoping you knew a bit more.

They both seems to use terrain + static collision detection with simple "physics". Might even just be simple collision based "physics". Just saying you use physics means little when complexity can be like 10+ year old games or expensive like PhysX GPU driven particle physics.

Somebody tell me what type of physics this is.
http://www.youtube.com/watch?v=naZWQ2S0Zzg

And what type of physics is used in this one.
http://www.youtube.com/watch?v=4eSG04ThpSQ

Winner gets a screenshot.
 
Last edited by a moderator:
That's also why Crysis 2 on console is being so closely scrutinized not just by us on forums, but very likely by video game developers around the world.

Also by research institutes and military. I think the market most adopting CE2/CE3 was not game developers. I am sure list has been posted of how many non game related businesses have adopted CE2/CE3, many for serious simulation "games" targeted for military and other types of use.
 
Status
Not open for further replies.
Back
Top