Poly Count, RagD Phys etc...

[nikhil]

hi guyz..

i'm into a game projects course.. and we guyz are using Ogre for building our game.. alongside Newton etc.

and as its a class assignment IT HAS TO SEE THE DAY LIGHT.. otherwise my grades are screwed

i'm on the programming side and my modellers have asked me this:-

1)If you have a preference for polygon counts, texture size and bone count for character, backgrounds, or models -- please let us know early on.

and

Will you be using ragdoll physics for deaths, explosions, etc?

Can the experienced guyz tell me how do i determine these things..? What polygon count I want.. I've no idea.. I'm happy thinkin that i know what polygon count actually means, or i think i know, but what count i want:?:..

The game is about racing on roller skates with people running after a motorball.. And its important to hold the ball to win.. People can also engage in melee fights with other racers..

Does this game require ragdoll Physics or is it just a question of how much are we looking to put in our game..

i'm sorry if these questions seem like passing off my work.. bt i'm nothing more than a programmer learning OGRE and these questions've completely stumped me ... Any guidance as to how to answer these questions will be appreciated..

Thanks
Nikhil

[xavier]

The better question is "does ragdoll physics have any benefit for gameplay, or is it there for eye-candy only?" Your designer(s) and/or project managers/producers (or whoever is fulfilling that role in your course) have to answer that question; it's not something for the engineers or artists to decide.

The other question can be answered more than adequately with a search (Google box above) on "poly count" and/or "polycount". The short answer: polycount is a meaningless metric -- worry more about batch count. The hits on your search will more than explain why that is.

[nikhil]

Hey thnks Xavier..

OH.. BTW.. i told my professors to get ur book for the class.. i'll be gettin it in few days time .. and being a complete noob to ogre.. u can be assured of a fair response from ur target audience (ME) ...

Can you plz briefly tell how does gameplay benefit frm ragD phys.. just a short example will do plz...

reat i'll google for poly count as u suggested..

regards

nikhil

[Alexander]

Unless you intend for player-controlled characters to recover from a ragdoll state to a playable state (very tricky to do from an animation as well as networking standpoint) or perhaps become zombie food or weapons after death (where their ragdoll-physics-determined position matters for gameplay reasons) then it is pretty much certainly going to be eye candy. This is especially true if you game is multiplayer where games will simply not send the expensive ragdoll calculations over the network.

If you are using Newton however, you might want to check to see if that feature is supported out of the box. It might make the decision easier for you and I seem to remember seeing something on Walaber's site perhaps were just that was done.

[nikhil]

Theres this ragdoll demo in OgreNewt.. but i cann't get it to compile. it needs TinyXML and the thing gives linking error tryin to complie that demo.. gives me gazillions of error saying:-

Code: Select all

XXXX stuff already defined in [b]MSVCRTD.lib(MSVCR80D.dll)[/b]

thats another reason i couldn't actually see for myself what ragD actually meant..

and as for out of box support.. i dont think theres any in OgreNewt, though i could use the demo and use the code in building one for my game.. but as far as i can see.. its not easy to accomplish, so i'll put the RagD thing as a NO NO..

Though I'm still curious, knowing of using RagD at the start really reduces the amount of work or will it require the same effort if we later on decide to go with it ????


thnks guyz
nikhil

[Alexander]

Ragdolling is something you could choose to implement at any reasonable point, I should say, but you should take a closer look at your libraries and tools to determine how much time and development effort should be spent on the endeavor.

As for the artists' part there is little they have to do up until they rig the models to the ragdoll primitives and joints of which you will need an application to handle. Some physics libraries come with tools for this and I know there is a free one called Scythe out there which may be able to do the job. You could also put together a solution using Maya or Max along with the model format COLLADA to hold the physics information.

The alternative would be that they would have to animate the death animations, perhaps with IK on the limbs to help with the illusion. Keep in mind that even some newer and realistic games don't use the RD technique. I remember seeing the death animations in Ghost Recon for the 360 recently and noticed that they were doing it the "old fashioned way."

[jacmoe]

Theres this ragdoll demo in OgreNewt.. but i cann't get it to compile. it needs TinyXML and the thing gives linking error tryin to complie that demo.. gives me gazillions of error saying:-

Code: Select all

XXXX stuff already defined in [b]MSVCRTD.lib(MSVCR80D.dll)[/b]

thats another reason i couldn't actually see for myself what ragD actually meant..

If TinyXML is used as a library (ie linked in), you need to recompile TinyXML after setting it to use multithreaded runtime libraries, just like Ogre.
And - to avoid memory manager problems in debug mode - modify the TinyXML main header to include ogre.h, preferably as the first include there.
Recompile it, and rest assured that you've made your own Ogre-friendly TinyXML library.

[Vectrex]

ragdoll in ogreNewt is really easy and really good (it's NOT using the newton ragdoll stuff which is too basic)
Go over to the ogreNewt forum (ogreAddons) and search for your problem, it's been addressed before.

[nikhil]

Theres this ragdoll demo in OgreNewt.. but i cann't get it to compile. it needs TinyXML and the thing gives linking error tryin to complie that demo.. gives me gazillions of error saying:-

Code: Select all

XXXX stuff already defined in [b]MSVCRTD.lib(MSVCR80D.dll)[/b]

thats another reason i couldn't actually see for myself what ragD actually meant..

If TinyXML is used as a library (ie linked in), you need to recompile TinyXML after setting it to use multithreaded runtime libraries, just like Ogre.
And - to avoid memory manager problems in debug mode - modify the TinyXML main header to include ogre.h, preferably as the first include there.
Recompile it, and rest assured that you've made your own Ogre-friendly TinyXML library.

Hi

"after setting it to use multithreaded runtime libraries, just like Ogre." If you meant setting tinyxmlSTL_Project_Property->Configuration_Properties->C/C++->Code_Generation->Runtime_Library to "Multi-threaded Debug (/MTd)" , it was already setup like that.

Also I included "ogre.h" in the "tinyxml.h" .. and recompiled..

and, you must have figured it out by now, i still get the error like:-

Code: Select all

1>LIBCMTD.lib(_ctype.obj) : error LNK2005: _isspace already defined in MSVCRTD.lib(MSVCR80D.dll)
1>LIBCMTD.lib(hooks.obj) : error LNK2005: "void __cdecl terminate(void)" (?terminate@@YAXXZ) already defined in MSVCRTD.lib(MSVCR80D.dll)
1>LIBCMTD.lib(stdexcpt.obj) : error LNK2005: "public: virtual __thiscall std::exception::~exception(void)" (??1exception@std@@UAE@XZ) already defined in MSVCRTD.lib(MSVCR80D.dll)

Wht am i doing wrong now...??? Y did tinyxml had to define these things again anyway.. Its Microsoft's job to redefine things as their own, and i think its offended by the fact that somebody (i.e tinyxml) redifined their already redefined code ..

regards
nikhil

[jacmoe]

It should be DLL: Multi-threaded Debug DLL, etc.