Theoretical question: I am just wondering if a high budget company were to attempt to start from scratch and develop a competitor to BeamNG, are they best to start with a Torque or CryEngine3, or is there other options out there to build a proper real time soft body physics car crashing game?
Can't answer the question directly but recent devs/support comments indicate a jbeam physics type scheme for any engine other than T3D would be almost impossible. It appears to be the biggest reason why they are tearing the base T3D code apart and rebuilding it. I will leave the tech stuff to others but the details of why this is happening with T3D are very specific and very intentional. Conversely there are other attempts at this sort of thing and I personally believe a good representation (visually at least), of jbeam type of deformation is close at hand by someone, somewhere. It seems to me if a jbeam 'copycat' was to be easily made in another engine we would have seen it by now? I also don't think there is another game developer on earth that has the interest or energy to try to replicate what BeamNG is trying to do with physics on the scale they are doing.
The Cryengine was probably dropped because its a pretty shitty engine to develop for and on top expensive. I've read it not only once over the years that the cryengine is just a clusterfuck of spaghetti code and mostly only experienced crytek devs can work with it "efficiently". With that out of the way lets not forget that the switch to Torque3d happened somewhere 2012 or 2013. At that time todays big engines werent released or were still super expensive. So i dont think T3D was chosen because its the best thing since sliced bread. It was the best choice for a small startup of 4 guys in that time period. But if somebody wants to develop a competitor to Beamng.Drive now they have a lot of other choices compared to 5 years ago. Also it kind of depends on the budget. Eg a big publisher wouldnt give a flying ... about T3D being open source. They have enough money to buy a closed source engines or develop their own inhouse engine. A small indie dev probably would look into UE4. Maybe some maniac would choose Lumberyard. IDK. But T3D surely isnt simply the best engine for softbody physics car games.
Any engine would work. Unity (probably w/ a native physics plugin), Unreal Engine, etc. If you were to develop a softbody physics engine, it'd hopefully be like other physics engines (PhysX, Bullet), which are decoupled from the engine itself, and could be moved from engine to engine freely.
Sure other games have pretty graphics, as they should in 2018. However, if you have to rip so much apart in the game engine, as with BeamNG, to get it to actually do what you want it to do, you might as well just improve 'what you have' that works with everything you ALREADY have, and improve upon it. Bonus points awarded for already being familiar with it's pros and cons and how it works. These graphics in Torque engine look 5~10 years old in some places, but not so much 'old' on newer maps... it's the lighting that's the shortcoming. Surely the developers have acknowledged and are working on improving the function and form of the engine - but function comes first. Then we can dress it up. For now the pig with lipstick will have to do as it surely has great physics and the mappers / vehicle artists have done a wonderful job so as it is.
If a small team had a blank check to find a proper developer for this who could meet all of their expectations in regards to building a proper soft body physics car crash game, maybe evcen using Unrela Engine 4, who would this team want to contact first to get the ball rolling? one thing I have learned is throwing a ton of money at a project isnt the end all, be all - you still need to have the right team of guys who know what they are doing
Thats a oddly specific question i cant answer. But i agree that simply throwing money at a project wont make it magically a success. Nonetheless i'm pretty sure i wouldnt choose t3d as the graphics/game engine. Especially since a now established company has already a headstart of >5 years.
The physics engine we have is a proprietary in-house solution. It could technically run on any engine, if you have the time and will to do that. Mind that behind our physics engine there's an immense amount of work. (How it works, how fast it can work, what is being simulated, and a ton more technical details that @estama could end up making the longest post here yet ) So not exactly a quick task to take this approach.
Well, with unlimited budget, everything would be done from scratch of course, but no one has unlimited budget, most big ones might have money and resources, but they are after big profits too which means deadlines that would drop most of the realism.
Our game engine is not really Torque3D anymore, we forked 5 years ago by now and nearly everything is changed, so it's essentially a new engine. In the code its called "GE" GameEngine nowadays as well
This is game-engine related content here, I'll post a safe-for-work link. Maybe the developers might be able to get pointers or reference some things. Link - Nvidia PhysX engine is now open-sourced. https://www.guru3d.com/news-story/nvidia-physx-engine-now-is-open-source.html Again, this link is safe for work and links to the Computer Hardware review site Guru 3D.
Posted this in another thread but i think it suits here too. If someone thinks the T3D engine lacks on graphics compared to UE4 or CryEngine then have a look at this recent blogpost form JeffR, PBR (Physics Based Rendering) is almost here (for T3D that is). https://forums.torque3d.org/viewtop...5d20a7ec46b381800e0a926be9cc&start=300#p10741
We have our own approach for new graphics. Not everything that T3D 'shows off' is production quality at all...