TGEA good for Physics Based game?
by Jamie Gibson · in Game Design and Creative Issues · 05/23/2008 (4:44 pm) · 9 replies
I have a lot of 3D game development experience and am looking to make a small game and am wondering if torque will be able to handle most of my ideas. I have some questions to ask that I hope somebody may help me with before I decide on a purchase.
A little info on the game:
8+ player team-based FPS
No terrain used
Physics based gameplay (key-factor)
1. Is it possible to have 200 cubes stacked up and will it react realistically when toppled? How's the physics engine?
1a. How are the physics online? Are they accurrate and in synch between players? Will we all see the same thing at the same time and is there any jitter?
2. How are parameters such as mass, elasticity and friction handled? Will a medium sized object slow me down some where a large object may slow me a lot or even stop me in my tracks?
2a. If an object is too heavy for one, would 2 players be able to push it? Do we combine our strength?
3. Does a moving object calculate damage on a player based on its mass and speed? Any splats?
4. Can different types of projectiles have varying degrees of force against a rigid body? Will a pebble rock it back and forth and a cannon ball knock it flying?
5. Can objects be picked up by the player and thrown or dropped?
5a. Can we give controls to rotate the object while it's held? Middle mouse for orb rotation?
5b. Can it be set up to affect movement speed based on the mass of the object?
5c. Can an object that is too heavy for one player be picked up if a 2nd player holds the lift action within a predefined area that is linked to the object; thus leaving control of the object to the 1st person who grabs it?
5d. Can an object be set to snap to other objects based on a radius around anchor points or intersecting dummy objects?
6. Do we have ragdoll deaths?
6a. Can players be broken apart into pieces?
6b. Can ragdoll be overridden? Can we blink out the player on a spectacular death and spawn a particle effect in its place?
6c. Can we use custom skeletons for ragdoll with weights and constraints? Like a limp mattress or a dead spider?
7. Can we emit rigid bodies?
So basically if I get like 9 no's and the physics aren't very good online, I still want to pursue the idea but as a single player. Have some ???'s for that too:
8. What's a reasonable number of rigid bodies on a high end machine running locally?
9. Can AI be scripted to react with rigid bodies?
9a. If they can't get around a rigid body, can they try to push it out of the way?
9aa. If it's too heavy or blocked by other objects on the other side, can they climb over them?
9aaa. If it's too tall to climb can they be scripted to attack the object and push it out of the way?
9b. Can AI be scripted to attack rigid bodies from a distance as soon as they're within their weapon's range?
9bb. Will they be able to recognize when a blocked path is now clear?
10. Can a very large AI controlled character be created? For instance, a 100ft gorrilla or a T-Rex?
10a. Can it exert a lot of force on rigid bodies either during attack or just by pushing them?
For the most part I'm expecting "if you can code it" and that's the kind of open-endedness I'm hoping for with this engine because I can code. But I admit that I'm desperately hoping that there are modules or code snippets for these types of elements in existence or integrated into the editor itself such as setting an object's mass. Some of the AI questions I'm hoping are natural behaviors already built into the pathing.
I'm sorry for the humongous list of 10 questions
A little info on the game:
8+ player team-based FPS
No terrain used
Physics based gameplay (key-factor)
1. Is it possible to have 200 cubes stacked up and will it react realistically when toppled? How's the physics engine?
1a. How are the physics online? Are they accurrate and in synch between players? Will we all see the same thing at the same time and is there any jitter?
2. How are parameters such as mass, elasticity and friction handled? Will a medium sized object slow me down some where a large object may slow me a lot or even stop me in my tracks?
2a. If an object is too heavy for one, would 2 players be able to push it? Do we combine our strength?
3. Does a moving object calculate damage on a player based on its mass and speed? Any splats?
4. Can different types of projectiles have varying degrees of force against a rigid body? Will a pebble rock it back and forth and a cannon ball knock it flying?
5. Can objects be picked up by the player and thrown or dropped?
5a. Can we give controls to rotate the object while it's held? Middle mouse for orb rotation?
5b. Can it be set up to affect movement speed based on the mass of the object?
5c. Can an object that is too heavy for one player be picked up if a 2nd player holds the lift action within a predefined area that is linked to the object; thus leaving control of the object to the 1st person who grabs it?
5d. Can an object be set to snap to other objects based on a radius around anchor points or intersecting dummy objects?
6. Do we have ragdoll deaths?
6a. Can players be broken apart into pieces?
6b. Can ragdoll be overridden? Can we blink out the player on a spectacular death and spawn a particle effect in its place?
6c. Can we use custom skeletons for ragdoll with weights and constraints? Like a limp mattress or a dead spider?
7. Can we emit rigid bodies?
So basically if I get like 9 no's and the physics aren't very good online, I still want to pursue the idea but as a single player. Have some ???'s for that too:
8. What's a reasonable number of rigid bodies on a high end machine running locally?
9. Can AI be scripted to react with rigid bodies?
9a. If they can't get around a rigid body, can they try to push it out of the way?
9aa. If it's too heavy or blocked by other objects on the other side, can they climb over them?
9aaa. If it's too tall to climb can they be scripted to attack the object and push it out of the way?
9b. Can AI be scripted to attack rigid bodies from a distance as soon as they're within their weapon's range?
9bb. Will they be able to recognize when a blocked path is now clear?
10. Can a very large AI controlled character be created? For instance, a 100ft gorrilla or a T-Rex?
10a. Can it exert a lot of force on rigid bodies either during attack or just by pushing them?
For the most part I'm expecting "if you can code it" and that's the kind of open-endedness I'm hoping for with this engine because I can code. But I admit that I'm desperately hoping that there are modules or code snippets for these types of elements in existence or integrated into the editor itself such as setting an object's mass. Some of the AI questions I'm hoping are natural behaviors already built into the pathing.
I'm sorry for the humongous list of 10 questions
About the author
#2
06/10/2008 (10:27 pm)
It's fairly easy for anyone not afraid of engine changes to get PhysX or any of the other decent physics libs into Torque. Far easier, IMO, than dealing with an engine with no source at all.
#3
Then you are in luck. The engine comes with full source code so if you can code it you can do it.
" But I admit that I'm desperately hoping that there are modules or code snippets for these types of elements in existence or integrated into the editor itself such as setting an object's mass. "
Again you are in luck. For example you can use PhysX in TGEA and there is already a resource for it. Note that you must already be a TGEA owner to access the resource.
There are many good resources available on the website. Maybe someday GarageGames will allow users to browse them (with code omitted obviously) so that users like you can see the wealth of additional resources you get when you buy a Torque engine.
06/10/2008 (11:00 pm)
" For the most part I'm expecting "if you can code it" and that's the kind of open-endedness I'm hoping for with this engine because I can code. "Then you are in luck. The engine comes with full source code so if you can code it you can do it.
" But I admit that I'm desperately hoping that there are modules or code snippets for these types of elements in existence or integrated into the editor itself such as setting an object's mass. "
Again you are in luck. For example you can use PhysX in TGEA and there is already a resource for it. Note that you must already be a TGEA owner to access the resource.
There are many good resources available on the website. Maybe someday GarageGames will allow users to browse them (with code omitted obviously) so that users like you can see the wealth of additional resources you get when you buy a Torque engine.
#4
06/11/2008 (6:52 am)
That's a great recommendation, Ben. Something like an authenticated download link but browsable resources. 
#5
Sounds like someone has been watching the Crysis Engine demos...
06/11/2008 (7:00 am)
1. Is it possible to have 200 cubes stacked up and will it react realistically when toppled? How's the physics engine?Sounds like someone has been watching the Crysis Engine demos...
#6
I had high doubts about the physics working out and I researched I guess maybe most of the engines. A lot of them seemed ok, but none of them are as complete as Torque. The good news is that it forced me to get my mind away from relying completely on the physics to drive the gameplay I had in mind. I've since come up with gamemodes that are fun even if everything in the game is completely static. Adding anything on top of that will now further enhance the now stronger core gameplay. I still have concerns about how any physics engine is going to perform online. I know the way the real world works and I haven't heard about any breakthroughs in that quite yet. It's always been a little off. However, I had always known that, so the gameplay doesn't require accuracy. Impeding characters in general is my only goal, but I do want to do some of the things I listed before with it, at least picking things up and throw or drop them like I've seen in a billion games already.
I have to wonder, though. How hard is it for the developers to implement it into the engine and give us an upgrade with a nice GUI or something? Also, it's a minefield of licensing issues with external engines, something that's almost scaring me enough to steer clear because I'm afraid I might miss something in the fine print. I would pay more for the license personally. I would also pay for an upgrade separately from the main license for it. I'd pay as much as the regular license. I'm just a regular dude, not a company. I hope to be one, one day. But not yet so I'm poor in terms of business. I can pay like a grand without thinking that I've gone and done something really risky. If it was like $250, I wouldn't even bat an eyelid. Something to think about. Till then, can the devs also consider opening up PayPal as a purchasing option?
I will be back on the boards soon to bug about the nitty gritty with hopefully some folks who are a lot better than I am.
Thanks again.
06/13/2008 (2:57 pm)
Thank you guys for the responses. As soon as I sort out some affairs, I'm going to be purchasing the license and working full-time on this project. I've done a considerable amount more research into the physics implementations and I can already see that it's quite doable to implement a very nice physics engine into this engine. The kind that can topple a lot of different objects nicely with all the good behaviors that we love. However difficult it may be in the end remains to be seen as I still await access to the real guts of the engine which will open up soon with the license. Most comments have been 50/50 on whether the person was successfull or not, so it clearly shows it can work. I must say that so far the performance of the engine is quite superb at even handling completely unoptimized and large geometries scattered quite liberally. I've done some stress tests of my own design on the demos to see how far it can go and it's well beyond the needs I have for my idea. In essence, I'm really stoked about it because I think it's going to work and even give me some room to play outside of the core gameplay with things like cool physics.I had high doubts about the physics working out and I researched I guess maybe most of the engines. A lot of them seemed ok, but none of them are as complete as Torque. The good news is that it forced me to get my mind away from relying completely on the physics to drive the gameplay I had in mind. I've since come up with gamemodes that are fun even if everything in the game is completely static. Adding anything on top of that will now further enhance the now stronger core gameplay. I still have concerns about how any physics engine is going to perform online. I know the way the real world works and I haven't heard about any breakthroughs in that quite yet. It's always been a little off. However, I had always known that, so the gameplay doesn't require accuracy. Impeding characters in general is my only goal, but I do want to do some of the things I listed before with it, at least picking things up and throw or drop them like I've seen in a billion games already.
I have to wonder, though. How hard is it for the developers to implement it into the engine and give us an upgrade with a nice GUI or something? Also, it's a minefield of licensing issues with external engines, something that's almost scaring me enough to steer clear because I'm afraid I might miss something in the fine print. I would pay more for the license personally. I would also pay for an upgrade separately from the main license for it. I'd pay as much as the regular license. I'm just a regular dude, not a company. I hope to be one, one day. But not yet so I'm poor in terms of business. I can pay like a grand without thinking that I've gone and done something really risky. If it was like $250, I wouldn't even bat an eyelid. Something to think about. Till then, can the devs also consider opening up PayPal as a purchasing option?
I will be back on the boards soon to bug about the nitty gritty with hopefully some folks who are a lot better than I am.
Thanks again.
#7
06/13/2008 (3:37 pm)
@Jamie, if you take a look at the PhysX license, I think you can rest assured there's nothing really scary there. It's free for the binary version as long as you're not targeting Xbox 360 or PS3. Getting the source license is $50k (for any of those cases, Xbox 360 and PS3 included).
#8
06/13/2008 (7:39 pm)
Yeah, you're right. Read it again and it appears clear enough. I assume the only issues will get run into if the project is highly successful, at which point I'd be happy to deal with issues.
Torque Owner Maddermadcat
So that kind of rules out most of your questions, because TGE(A) has essentially no physics, haha. You've got RigidShape which is decent for crates and the like, and then some basic gravity stuff. No ragdolls, joints, anything. You can implement it yourself, of course, but it won't be easy.
9b. Yeah, of course, a radius search, set look target, and make it shoot. Quite straightforward.
9bb. If you can code it. :)
10. Mhm, I've simply scaled my player and it continued to work just fine.
10a. Yeah, I think so.