Physics Test...
Community Forums/Showcase/Physics Test...
| ||
| Im currently re-working the vehicle physics for my game Island Storm, the scenery (crates, barrels, etc...) uses Tokamak for the physics, but I have found it hard to make a decent vehicle with Tokamak, so im now trying to mix ODE and Tokamak together. Here is my first attempt, its a little quad bike game thing, the bikes are ODE and the scenery is Tokamak, at the moment the collisions between to two are not great and I can't seem to get the ODE objects to transfere enough force into the Tokamak ones, but im getting there. Here is a test, let me know if it works: http://www.apaj15.dsl.pipex.com/KickstartQUAD%20Demo.zip  |
| ||
| Using two physics dlls, plus Blitz collision, sounds like hell. |
| ||
| Yep... I don't think it's a good idea... |
| ||
| Why not if it once works. I think it sounds smart. |
| ||
| Hm... It feels a bit heavy... Perhaps because there's not enough ground friction. Also a few glitches involving objects going through the dummy guy. Looks nice though (and I love how everything has turned out so far in IS), but I think you need to do some revision on this section. |
| ||
| Works fine here! Like it... |
| ||
| Pretty cool - nide to see ODE still be used for something. You are still getting that problem of rearwheel steer - all my reading on the ODE forums suggests that no-one has yet found a way around that while still using the carwheel joint type for the rear wheels - certainly I could do no better than a sort of bodge which worked most of the time, but as the speeds increased the problem came back, ususally flipping the car over. Have you tried the Stunt Car Racer demo in my sig? It uses just Tokamak for car physics, and while not perfect, it does allow for high speeds, powerslides and vast improbable jumps (through a cunning tweak!). The lack of suspension springs is the only area where Tokamak really falls over - you could probably do it with recently added sensor commands, as shown by the official Tokamak car demo, but I've not tried it yet. I will be quite happy to discuss the Tokamak code behind my demo with you if it would help. |
| ||
| Actually I've just checked my demo and it's a very old version - the current one is much better, especially with a steering wheel, but I don't have it here at work. |
| ||
| Works fine for me too. Only comment would be as Mr Picklesworth - seems a bit sluggish, almost bullet-time like, perhaps you could speed up the physics timing a little ? Also, do try Vordermans SCR demo as its good. I've also had success using the same library to simulate a wheeled vehicle. Would eliminate the difficulties in co-operating with ODE. @Vorderman - any chance of firing up a newer demo when you get home ? been following progress with interest. Funky Frank |
| ||
| > Funky Frank Let's jive Funky Frank! |
| ||
| Thanks all, the feel is slugish as it's locked at 30fps so I can see whats going on during impacts. Vorderman, I would love to discuss your SCR code, if I could get everything working through Tokamak that would be great, but im not great at physics code and it's the springs in ODE that make the vehicles look fun and bouncy. |
| ||
| lol @ ENAY I'd love to - but have no sense of rhythm (so i've been told). Funky Frank |
| ||
| If it works, who should care? |
| ||
| Yeah it's the springs that are a problem in Tokamak - ODE's carjoint is much better than anything Tokamak has at the moment. I think you would be better staying with just ODE for anything that requires a car's suspension - mixing the two systems is making it way too complex. The 'suspension' in my SCR demo relies upon the joints being forced apart by impacts rather than any real spring system, so in effect it is using the inaccuracies in Tokamak, which is not ideal. It seems to work, but not too well all in all. |
| ||
| If I understood the ODE-manual correctly the springs in ODE is an effect that is caused by the ERP and CFM. By fiddling with these values you can make the joint behave like they were connected to a spring. |
| ||
| Thats correct. One of them is the error correction parameter that brings the rogue joints together again, and the other is how far they get apart when affected. |
   |