Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Fuzzy Avatars Solved! Please re-upload your avatar if it was fuzzy!

Massively & CU's rendering engine + Andrew Meggs chimes in...

NegativeXNegativeX Denver, COPosts: 100Member

http://massively.joystiq.com/2013/04/03/camelot-unchained-discusses-the-trouble-with-rendering-engines/

Andrew Meggs adds in a couple more detailed pieces of information and rebuttals in the comment section.

Camelot Unchained Fanpage

https://simply-gaming.com/camelot/

Comments

  • gylnnegylnne South Hutchinson, KSPosts: 320Member
    Originally posted by NegativeX

    http://massively.joystiq.com/2013/04/03/camelot-unchained-discusses-the-trouble-with-rendering-engines/

    Andrew Meggs adds in a couple more detailed pieces of information and rebuttals in the comment section.

    Appreciate the link Negative, thank you.:)

  • DMKanoDMKano Gamercentral, AKPosts: 8,518Member Uncommon

    The secret sauce is revealed:

    "Phase two, which I'm working on RIGHT NOW, is to reduce that number of draw calls. The plan is to stream multiple meshes into a single vertex buffer, and then fire off one draw that covers 10-20 characters at a go. It's a generalization of the WAR character art system (which I wrote), in which all the models attached to a character (gloves, boots, helm, etc.) were merged into a single unique per-character mesh so it could be done in one draw. By extending that to work across characters rather than just within a single character, we'll go even further. Of course, the downside of that approach is that in order to get high performance I'll have to treat multiple copies of the same mesh as if they're unique meshes....you're saying that's the problem, but I'm saying it's the solution. :-)"

    So the engine will duplicate player models in RvR instead of drawing individual models. 

    I am not a fan of this approach, as you won't be able distinguish individual player by their unique look. Seeing copies of 5 base models sounds sort of a letdown, I hope I am wrong.

  • belatucadrosbelatucadros Toronto, ONPosts: 263Member
    Originally posted by DMKano

    The secret sauce is revealed:

    "Phase two, which I'm working on RIGHT NOW, is to reduce that number of draw calls. The plan is to stream multiple meshes into a single vertex buffer, and then fire off one draw that covers 10-20 characters at a go. It's a generalization of the WAR character art system (which I wrote), in which all the models attached to a character (gloves, boots, helm, etc.) were merged into a single unique per-character mesh so it could be done in one draw. By extending that to work across characters rather than just within a single character, we'll go even further. Of course, the downside of that approach is that in order to get high performance I'll have to treat multiple copies of the same mesh as if they're unique meshes....you're saying that's the problem, but I'm saying it's the solution. :-)"

    So the engine will duplicate player models in RvR instead of drawing individual models. 

    I am not a fan of this approach, as you won't be able distinguish individual player by their unique look. Seeing copies of 5 base models sounds sort of a letdown, I hope I am wrong.

    I don't think you are understanding correctly. This shouldn't be a problem

    Eldritch, Nightshade, Warden, Druid, Mauler, Mentalist
    Sorcerer, Necromancer, Theurgist, Armsman, Cleric
    Healer, Warrior, Skald

  • SBE1SBE1 New York, NYPosts: 335Member

    I think what the developer is saying is something akin to this example:

    Suppose you have a green helm, red torso, blue pants and yellow boots.   That's 4 different "mesh" patterns that would ask the  computer to draw 4 different things when it moves.  Instead, think of it as outfit #2310, which is an outfit predefined as a green helm, red torso, blue pants and yellow boots.  Thus, the computer is asked to draw outfit #2310, which is simply 1 drawing.  As a result, you've decreased the workload to 1 draw versus 4 draws for the same character.

    Now, that's like a super duper simplified example, but that's what he means by having the whole thing as a unique mesh.  Hence, people will still look different, but from a computer point of view they are just changing outfits.

     

    Also, you have to remember that in an early update about a month or so ago they said that they going with performance first, so initial testing is being done with very basic landscape, textures, etc.   They will then add as much of that stuff as possible but performance (FPS) has to remain above a certain threshold once you have around 150-200 players in a select region.  So, you might have detailed characters and less detailed landscapes or vice versa. Every level of detail reduces FPS, so it's a tradeoff.

    The other big tradeoff they have to deal with is how much to put on the client and how much to put on the server.  People can't be trusted, so most of this has to go to the server. So, even if they have a good engine, crappy server code will cause ability delay lag, which is just as bad as low FPS.  That's why he also addresses the way the server has to change how it calculates things like AoE effects and such. 

  • Plastic-MetalPlastic-Metal Highland Heights, KYPosts: 405Member

    ^ SBE1 is close to the target.

    Mesh instancing has been within the industry for years and every game uses it.  The most common use is through environmental objects, like trees, fences, brush, crates, etc.

    No reason for alarm. :)  I am, however, glad some people are asking questions rather than spouting off misinformation and pretending they have experience and knowledge of the technology behind the system(s).

    My name is Plastic-Metal and my name is an oxymoron.

    image

  • meddyckmeddyck Athens, GAPosts: 1,140Member Uncommon
    I don't care about the details of how the engine works. I care about what kind of gameplay the engine will be rendering. Once you get over a few dozen players from each realm fighting, the impact any individual player has on the battle drops dramatically. It quickly becomes an AE spamfest. Anybody who is focus fired dies so fast no healer has a chance of keeping him alive. Those kinds of huge fights are not something that I find fun on a regular basis. Sure you deal with them for the odd relic raid. But I don't want to be engaged in fights featuring 1000s let alone 10s of 1000s of players on a daily basis. CSE needs to clarify if that is the type of game they are trying to create because if so, it's not for me and I'll need to adjust my donation accordingly.

    Camelot Unchained Backer
    DAOC [retired]: R11 Cleric R11 Druid R11 Minstrel R9 Eldritch R6 Sorc R6 Scout R5 Healer

  • belatucadrosbelatucadros Toronto, ONPosts: 263Member
    Originally posted by meddyck
    I don't care about the details of how the engine works. I care about what kind of gameplay the engine will be rendering. Once you get over a few dozen players from each realm fighting, the impact any individual player has on the battle drops dramatically. It quickly becomes an AE spamfest. Anybody who is focus fired dies so fast no healer has a chance of keeping him alive. Those kinds of huge fights are not something that I find fun on a regular basis. Sure you deal with them for the odd relic raid. But I don't want to be engaged in fights featuring 1000s let alone 10s of 1000s of players on a daily basis. CSE needs to clarify if that is the type of game they are trying to create because if so, it's not for me and I'll need to adjust my donation accordingly.

    Technically DAOC was that type of game too.

    8vs8 worked out just fine

    Eldritch, Nightshade, Warden, Druid, Mauler, Mentalist
    Sorcerer, Necromancer, Theurgist, Armsman, Cleric
    Healer, Warrior, Skald

  • gylnnegylnne South Hutchinson, KSPosts: 320Member
    Originally posted by meddyck
    I don't care about the details of how the engine works. I care about what kind of gameplay the engine will be rendering. Once you get over a few dozen players from each realm fighting, the impact any individual player has on the battle drops dramatically. It quickly becomes an AE spamfest. Anybody who is focus fired dies so fast no healer has a chance of keeping him alive. Those kinds of huge fights are not something that I find fun on a regular basis. Sure you deal with them for the odd relic raid. But I don't want to be engaged in fights featuring 1000s let alone 10s of 1000s of players on a daily basis. CSE needs to clarify if that is the type of game they are trying to create because if so, it's not for me and I'll need to adjust my donation accordingly.

    I don't think Andrew was trying to clarify anything at this point, just to show you how he plans to build a game engine that works within an rvr situation.

  • meddyckmeddyck Athens, GAPosts: 1,140Member Uncommon
    Originally posted by belatucadros
    Originally posted by meddyck
    I don't care about the details of how the engine works. I care about what kind of gameplay the engine will be rendering. Once you get over a few dozen players from each realm fighting, the impact any individual player has on the battle drops dramatically. It quickly becomes an AE spamfest. Anybody who is focus fired dies so fast no healer has a chance of keeping him alive. Those kinds of huge fights are not something that I find fun on a regular basis. Sure you deal with them for the odd relic raid. But I don't want to be engaged in fights featuring 1000s let alone 10s of 1000s of players on a daily basis. CSE needs to clarify if that is the type of game they are trying to create because if so, it's not for me and I'll need to adjust my donation accordingly.

    Technically DAOC was that type of game too.

    8vs8 worked out just fine

    The server size (max of 3000 to 4000) and fact that ~80% of the population was doing PvE at any given time made it much easier for DAOC to support soloing and group vs group fighting. If CU will have 10,000 player servers where you get 3,333 vs 3,333, vs 3,334 fights and there's no PvE, that's an entirely different situation.

    Camelot Unchained Backer
    DAOC [retired]: R11 Cleric R11 Druid R11 Minstrel R9 Eldritch R6 Sorc R6 Scout R5 Healer

  • belatucadrosbelatucadros Toronto, ONPosts: 263Member
    Originally posted by meddyck
    Originally posted by belatucadros
    Originally posted by meddyck
    I don't care about the details of how the engine works. I care about what kind of gameplay the engine will be rendering. Once you get over a few dozen players from each realm fighting, the impact any individual player has on the battle drops dramatically. It quickly becomes an AE spamfest. Anybody who is focus fired dies so fast no healer has a chance of keeping him alive. Those kinds of huge fights are not something that I find fun on a regular basis. Sure you deal with them for the odd relic raid. But I don't want to be engaged in fights featuring 1000s let alone 10s of 1000s of players on a daily basis. CSE needs to clarify if that is the type of game they are trying to create because if so, it's not for me and I'll need to adjust my donation accordingly.

    Technically DAOC was that type of game too.

    8vs8 worked out just fine

    The server size (max of 3000 to 4000) and fact that ~80% of the population was doing PvE at any given time made it much easier for DAOC to support soloing and group vs group fighting. If CU will have 10,000 player servers where you get 3,333 vs 3,333, vs 3,334 fights and there's no PvE, that's an entirely different situation.

    For sure. You should hop on IRC if you're capable, but basically, I think everything will be OK

    Eldritch, Nightshade, Warden, Druid, Mauler, Mentalist
    Sorcerer, Necromancer, Theurgist, Armsman, Cleric
    Healer, Warrior, Skald

  • meddyckmeddyck Athens, GAPosts: 1,140Member Uncommon
    Originally posted by gylnne

    I don't think Andrew was trying to clarify anything at this point, just to show you how he plans to build a game engine that works within an rvr situation.

    The same point could have been made by showing 1,000 characters running around instead of 10,000 which is way more than other RvR games handle well. That they decided to show 10,000 suggests but doesn't prove that they want fights of that scale to happen in CU. Hence why I would like them to clarify what the scale of combat they are shooting for is.

    Camelot Unchained Backer
    DAOC [retired]: R11 Cleric R11 Druid R11 Minstrel R9 Eldritch R6 Sorc R6 Scout R5 Healer

  • belatucadrosbelatucadros Toronto, ONPosts: 263Member
    Originally posted by meddyck
    Originally posted by gylnne

    I don't think Andrew was trying to clarify anything at this point, just to show you how he plans to build a game engine that works within an rvr situation.

    The same point could have been made by showing 1,000 characters running around instead of 10,000 which is way more than other RvR games handle well. That they decided to show 10,000 suggests but doesn't prove that they want fights of that scale to happen in CU. Hence why I would like them to clarify what the scale of combat they are shooting for is.

    I'm not sure it does. Consider that is just pure rendering, not of the effects/data processing that needs to go on.

    You want to rendere to be as high performance as possible, regardless

    Eldritch, Nightshade, Warden, Druid, Mauler, Mentalist
    Sorcerer, Necromancer, Theurgist, Armsman, Cleric
    Healer, Warrior, Skald

  • mklinicmklinic Pottstown, PAPosts: 1,435Member Uncommon
    Originally posted by meddyck
    Originally posted by gylnne

    I don't think Andrew was trying to clarify anything at this point, just to show you how he plans to build a game engine that works within an rvr situation.

    The same point could have been made by showing 1,000 characters running around instead of 10,000 which is way more than other RvR games handle well. That they decided to show 10,000 suggests but doesn't prove that they want fights of that scale to happen in CU. Hence why I would like them to clarify what the scale of combat they are shooting for is.

    I didn't get that from the video (intent to have 10k person battles). I got that they were using a pretty simple model and, knowing the actual game models might have a bit more, just loaded a bunch of them as a means to show "we can do this simple scene to this scale so think a more complex scene at a smaller scale should be fine". Just my take-away though based on an extremely early presentation.

    I do agree that'd be nice to know what their "sweet spot" is intended to be for large battles. Given all the thought they've put in to the design thus far, I imagine they would have defined this though maybe they were waiting to gauge the potential population (using KS as a factor). I guess it would also be a "damned if you do, damned if yo don't" scenario though. If they stated "we're shooting to handle 1000 v 1000 battles smoothly", and then revise that down at all based on technical or design considerations, I could see the criticism pouring in. :P

    -mklinic

    "There's a point I think we're missing.
    It's in the air we raise our fists in."
    -from Behind Closed Doors by Rise Against

  • meddyckmeddyck Athens, GAPosts: 1,140Member Uncommon
    Originally posted by belatucadros
    Originally posted by meddyck
    Originally posted by gylnne

    I don't think Andrew was trying to clarify anything at this point, just to show you how he plans to build a game engine that works within an rvr situation.

    The same point could have been made by showing 1,000 characters running around instead of 10,000 which is way more than other RvR games handle well. That they decided to show 10,000 suggests but doesn't prove that they want fights of that scale to happen in CU. Hence why I would like them to clarify what the scale of combat they are shooting for is.

    I'm not sure it does. Consider that is just pure rendering, not of the effects/data processing that needs to go on.

    You want to rendere to be as high performance as possible, regardless

    It's a balancing act. There's the maximum possible characters to render like in this 10,000 model demo. Then there's the target number of what they realistically would like the numbers in big fights to be. And there's what the numbers could be if they made the graphics state of the art; this would be a lot lower than the others but still might be high enough to be perfectly fine for me. Then they tweak the engine to find the right mix of performance and beauty. I just want to know roughly what that second number is that they are targetting.

    Camelot Unchained Backer
    DAOC [retired]: R11 Cleric R11 Druid R11 Minstrel R9 Eldritch R6 Sorc R6 Scout R5 Healer

  • SeariasSearias Edmonton, ABPosts: 712Member Uncommon
    Originally posted by meddyck
    Originally posted by gylnne

    I don't think Andrew was trying to clarify anything at this point, just to show you how he plans to build a game engine that works within an rvr situation.

    The same point could have been made by showing 1,000 characters running around instead of 10,000 which is way more than other RvR games handle well. That they decided to show 10,000 suggests but doesn't prove that they want fights of that scale to happen in CU. Hence why I would like them to clarify what the scale of combat they are shooting for is.

    I don't think that is what they are trying to imply. Think of it this way, showing 10,000 units on screen is more impressive than showing 1000 units on screen. Since the models used were pretty low quality and showing 1000 units would not really impress many people. I think this is why they went with 10,000.

    <InvalidTag type="text/javascript" src="http://www.gamebreaker.tv/cce/e.js"></script><div class="cce_pane" content-slug="which-world-of-warcraft-villain-are-you" ctype="quiz" d="http://www.gamebreaker.tv"></div>;

Sign In or Register to comment.