2d - Verlet - Proof of concept demo.

[STEVIE G]

Hello,

I've been working on this engine for a few months now when I've found the time. It's at a stage where I'd like to show a demo.

http://www.steviegoodwin.plus.com/images/Screenshot014.PNG

The engine is verlet integration based and quite versatile as you can see by some of the objects. The collision engine is ok and could do with some improvement. Some objects get stuck in others but are easy enough to free ( actually I quite like it ;) ). What I'm most happy about is the ragdolls and physics based emergent behaviour. :)

The basic idea is that you control a ragdoll and throw stuff at other ragdolls using something similar to the 'force'. Hold leftmouse button over an object to grab and drag it towards the cursor. You can only grab most objects at their centers ( neck and hip for ragdolls ). The force can only be used on objects within your 'force' radius. At the moment the only objects which do damage are the mines and you only control the yellow guy who is indestructable .

I have a pretty good idea where I'm going with this one but feel free to suggest anything. All feedback welcome. Please also report FPS, which is limited to 50fps btw.

OBJECTS
=====================
Ragdoll
Spider
Bird
Worm
Hamster Ball
Crate
Sack
Wrecking ball
Rope Bridge
Cart
Mine
Vine
Palm Tree
Grass

CONTROLS
=====================
LM = Grab/Pull
RM = Explosion test
Mouse = Move cursor
SPACE = Reset Game
=====================

http://www.steviegoodwin.plus.com/TD.zip

Cheers
Stevie

[Farbror Fazer]

Hey man, pretty good game, just got a few hints...

* Make the simulation faster so everything doesn't fly in slo-mo

* Fix so the Ragdolls can spann around instead of up-righting themselves but so they still stand by themselves
* Implement (rightspelled?) weapons in the game, preferrably used with right mouse button so youd have to use the force and take cover behind stuff

otherwise Good Job! keep at it!

[STEVIE G]

* Make the simulation faster so everything doesn't fly in slo-mo
* Fix so the Ragdolls can spann around instead of up-righting themselves but so they still stand by themselves
* Implement (rightspelled?) weapons in the game, preferrably used with right mouse button so youd have to use the force and take cover behind stuff

I like the speed as is and I'm very happy with the way the ragdolls stand upright so they're be no change to the later. I might increase the timestep slightly though. I've got plans for quite a few weapons traps and other stuff which'll kill you.

Cheers
Stevie

[Matthew]

Great tech demo! Are you planning on implementing a game on top of it?

I don't have much to add for suggestions, although I would echo that it feels very floaty. Try doubling the speed and playing with it for a day or two before you change it back. It'll feel really fast at first, but give faster simulation a shot before you decide which is better.

[juju02]

great! just make it faster, and maybe some weapons:)

[STEVIE G]

I'm curious to know what FPS everyone is getting, especially the ones who think it's slow? It only runs as fast as your CPU will allow it ( GPU will not be the bottleneck ) up to 50fps, so it may be the case that you're not playing it at the indended speed.

Let me know.
Cheers
Stevie

[fusion1224]

Good work. I like the buoyancy effect with the water. You've got everything running at a good speed, as in nothing is lagging even with all those ragdolls. On my machine things seem to be running above 50FPS even though you said that's what the FPS is limited to :).

Just a question... what Verlet algorithm are you using - normal, velocity or leapfrog? When I was making physics for my game I was having problems with length constraints with the Velocity Verlet algo. It swung as normal but slowed down to a stop very quickly, almost as if it had a very high amount of damping applied to it. Normal Verlet worked fine but I needed velocity to be calculated for collision detection. The way I did collision detection is if the simulated body penetrated the terrain - before it had been rendered penetrating the terrain - I set it's position to out of the terrain and reflected it's velocity so it appeared to bounce off. Did you do collision detection a similar way? Your collision detection appears very stable.

Keep up the good work!

[STEVIE G]

Just a question... what Verlet algorithm are you using - normal, velocity or leapfrog? When I was making physics for my game I was having problems with length constraints with the Velocity Verlet algo. It swung as normal but slowed down to a stop very quickly, almost as if it had a very high amount of damping applied to it. Normal Verlet worked fine but I needed velocity to be calculated for collision detection. The way I did collision detection is if the simulated body penetrated the terrain - before it had been rendered penetrating the terrain - I set it's position to out of the terrain and reflected it's velocity so it appeared to bounce off. Did you do collision detection a similar way? Your collision detection appears very stable.


I use the standard position based algorithm. For better air resistance don't use a viscous system described in Jacobsens paper - you need to model it correctly ... using DragCoef * ( Speed * Velocity ). Where Velocity is implied by new - old positions. I don't use any reflection in my collision with the landscape - this is all handled automatically by Blitz3d. I'm using 2d in 3d and the landscape is actually a mesh. Blitz collisions can be set to automatically prevent penetration so I don't even need to project out of collisions. I also have some code to prevent objects breaking by getting inverted due to a large impact. Only 3 iterations of the relaxation loop are required as a result.

I know that this collision method does not work with single particles but the idea of projecting both old and new positions onto the collision plane is how I'd do it. This is the same as what you describe.

Stevie

[Justin1221]

great job! but the camera control is a little tricky.

[Beakless]

Yeah i like it, I like the random little creatures ;) Umm, yeah, speed the whole thing up though, and make the poeple heavier, and you have a great engine !

[STEVIE G]

Cheers

Is everyone who says this is running slow too frightened to post their FPS cos they've got a naff Duron 1.1? I really need this number to gauge how other systems are handling it. Pretty please.

Cheers
Stevie

[Xerxes]

It runs really smoothly for me, but all the objects act like they're floating in mollasses...

[STEVIE G]

Yes but .....


Is everyone who says this is running slow too frightened to post their FPS cos they've got a naff Duron 1.1? I really need this number to gauge how other systems are handling it. Pretty please.

[Matthew]

I don't think it's a performance issue--just the way the numbers are tuned. It feels very slow to me. It's difficult to be objective when you're tuning your own mechanic (you quickly adapt to it, and not it to you). Honestly, just try doubling the speed of everything and keeping it that way for a week. If you don't like it after a week you can always change it back...

More details:

Processor: AMD Athlon(tm) 64 X2 Dual Core Processor 4400+, MMX, 3DNow (2 CPUs), ~2.2GHz
Memory: 2048MB RAM

Here's how fast it runs on my box (video capture): http://www.fun-motion.com/misc/td-test1.html

Probably easiest if you verify that's the intended simulation speed (or if others verify if that's speed on their boxes, too)...

[STEVIE G]

I don't think it's a performance issue--just the way the numbers are tuned. It feels very slow to me. It's difficult to be objective when you're tuning your own mechanic (you quickly adapt to it, and not it to you). Honestly, just try doubling the speed of everything and keeping it that way for a week. If you don't like it after a week you can always change it back...

More details:

Processor: AMD Athlon(tm) 64 X2 Dual Core Processor 4400+, MMX, 3DNow (2 CPUs), ~2.2GHz
Memory: 2048MB RAM

Here's how fast it runs on my box (video capture): http://www.fun-motion.com/misc/td-test1.html

Probably easiest if you verify that's the intended simulation speed (or if others verify if that's speed on their boxes, too)...

It couldn't be easier than posting the FPS which is sitting nicely at the top left of the screen ;)

Anywho, just doubling the speed isn't straightforward unless you're using euler. You need to bear in mind that twice the speed requires adjustments to all objects drag, friction and will probably require a rewrite of the collision routines and the control mechanism - so may end up slower. I'll maybe mess about with it at some stage but the game I'm making will be better suited to a slower pace with more strategy and skill involved. There'll be no right mouse button auto explosion in the final version ;)

[Juggz]

I got a steady 40 - 50

Do speed up the simulation though, it feels like working in honey instead of air...

[Jokker]

45-46 constantly.

Also, it is indeed slow. Speeding it up would be a good idea unless you can think of another way. :/

[fusion1224]

Maybe I'm the only one here that doesn't think you should speed it up, but even if you do wouldn't it be a simple increase in gravity?
Also, are your ragdoll physics just several length constraints joined together as described in the Jacobsen paper? I'm currently programming a physics engine that uses RK4 integration and have wondered if ragdoll physics are done that way.

[STEVIE G]

Maybe I'm the only one here that doesn't think you should speed it up, but even if you do wouldn't it be a simple increase in gravity?

At last someone who prefers a more leisurely paced game :) I've posted the demo on several different forums and this is the only one that complained about speed.

I can increase the gravity easily by increasing the timestep but this means that you can't spin the ragdolls so well. I've mentioned before that everything would have to be tweaked ... drag , mass, friction, dynamic friction , static friction etc.. to compensate for the increase in speed. It's alot of trial and error let me tell you. I'll play with it at some stage and MAYBE make it an option for this particular game.


Also, are your ragdoll physics just several length constraints joined together as described in the Jacobsen paper? I'm currently programming a physics engine that uses RK4 integration and have wondered if ragdoll physics are done that way.

Yes, but the constraints structure is different. Each ragdoll has 15, pointmasses, 15 stiff constraints, 5 minimum distance constraints and 10 rotational constraints. Like so .... the white are the stiff , green minimum and red rotational.

http://www.steviegoodwin.plus.com/images/Screenshot020.PNG

Minimum Distance or Dot Product constraints have their place for some things ( they pretty much do the same thing as each other ) but they introduce singularities when trying to prevent illegal joint rotations. e.g. Imagine you have a stick figure like this where 4 is the head, 1 the chest, 0 the hip and 2 and 3 the elbows ...

........4
........!
2----1----3
.......!
.......!
.......0

You apply a minimum distance or dotproduct constraint between 0 and 3 to prevent the elbow moving through the torsoe. If you were to pull the elbow hard enough it could end up on the left hand side of point 0 so will stay there until pulled hard back to the right.

I came up with a solution which takes 3 pointmasses and constrains the last to be within a low and high angle range, relative to the direction between the first and second pointmasses. If the pointmass has breached the low or high angle I determine which angle is closest to the current and correct in that direction. So, basically if you pull 3 hard enough to end up on the left it will spring quickly back to where it should be. As you can see in the demo it works pretty well.

Stevie

[tommyverceti]

ive been waiting for this for ages, great demo, i like the little wormy thing in the water

[X-iDuex]

This is a great demo looks like a very promising game, and of course it would be nice if things were speed up but if it takes to much work than it's all good.

My FPS is usually between 60-70.

The only suggestion I have is that it'd be cool to have dynamic break physics applied to objects like the crates, Also dynamic water physics, it my be a little pressing on the CPU but you could make it so only a certain amount of the water reacts fluidically.

Can't wait to see what else you do with this, it really looks amazing.

[Wrestler#1]

my fps: 50 (83)

PS: Are the 83 the "real" fps?
Sorry, I don't really know, how to explain the "real" :P

[STEVIE G]

Thanks. The '83' is the number of frames a second that it could render given the chance. It just helps me to see how many FPS I have to play with when adding more to it.

Cheers
Stevie

[X-iDuex]

So what were your plans for making this in to a game. Also what do you think about the destructible crate idea, could you do that?

[STEVIE G]

The game engine is almost done. I just finished the animation system so need to start work on an object editor which I'll hopefully include with the game.

Breakable crates are easily possible and will be implemented.

Here's a couple of new screenies, improved graphics etc...

http://www.steviegoodwin.plus.com/images/Screenshot028.PNG
http://www.steviegoodwin.plus.com/images/Screenshot030.PNG

Stevie

[X-iDuex]

Wow man your making some serious progress, btw I checked out some of your other demo and there all really good, did you build this engine from ground up?

[STEVIE G]

Wow man your making some serious progress, btw I checked out some of your other demo and there all really good, did you build this engine from ground up?


Yeah, I've been working on it every day for a couple of weeks now and I'm really pleased that everything is just fitting together nicely. It's great fun just testing it and adding new objects. The engine was built using the best bits from previous engines .. as you can see by my older demo's I've been messing with physics for a few years now.

The idea is similar to worms in that it's turn based. You get a set amount of force per turn which you can use to throw everything at anything or charge up a shield which gives you some protection for the current and next shot. If you hit things hard enough - everything on the level can do damage. Basically you have to kill the opposition by any means necessary. You can kick and punch them, chuck them in with the Piranha's , through a meat grinder and even drown them. :) After a set time the water level will start to rise, restricting the space you work in. You'll also be able to eat some of the objects to recharge your life bar.

The great thing is that no game will ever be the same. I've plans to include a custom landscape creator and physics object creator which should allow you to build just about any kind of object and drop it into the game. You'll be able to easily mod the object graphics and background images etc..

[sakiskid]

have u realesed the updated part yet?
if so tell me plz

IT LOOKS SO COOL!:cool:

[X-iDuex]

Yeah, I've been working on it every day for a couple of weeks now and I'm really pleased that everything is just fitting together nicely. It's great fun just testing it and adding new objects. The engine was built using the best bits from previous engines .. as you can see by my older demo's I've been messing with physics for a few years now.

The idea is similar to worms in that it's turn based. You get a set amount of force per turn which you can use to throw everything at anything or charge up a shield which gives you some protection for the current and next shot. If you hit things hard enough - everything on the level can do damage. Basically you have to kill the opposition by any means necessary. You can kick and punch them, chuck them in with the Piranha's , through a meat grinder and even drown them. :) After a set time the water level will start to rise, restricting the space you work in. You'll also be able to eat some of the objects to recharge your life bar.

The great thing is that no game will ever be the same. I've plans to include a custom landscape creator and physics object creator which should allow you to build just about any kind of object and drop it into the game. You'll be able to easily mod the object graphics and background images etc..

Man that sounds awesome, the game sounds unique, challenging, and lots of fun. The modding is a great idea, is the editor going to allow you to create enemies and weapons that can deal damage and what not? And is there going to be a way to create new game modes? Are you going to implement the same sort of control scheme from the robot demo where you use keys to select a certain part of the body? And I was wondering did you get the idea from Ragdoll KungFu?

Anyways sorry for all the questions I'm just very curious about this game, I mean you've got a lot of skill and on top of that an ingenious design for a very fun game.
Oh and one last question, are you planing on releasing another demo anytime soon or when do you plan to release the actual game?

[STEVIE G]

Man that sounds awesome, the game sounds unique, challenging, and lots of fun. The modding is a great idea, is the editor going to allow you to create enemies and weapons that can deal damage and what not? And is there going to be a way to create new game modes? Are you going to implement the same sort of control scheme from the robot demo where you use keys to select a certain part of the body? And I was wondering did you get the idea from Ragdoll KungFu?

Anyways sorry for all the questions I'm just very curious about this game, I mean you've got a lot of skill and on top of that an ingenious design for a very fun game.
Oh and one last question, are you planing on releasing another demo anytime soon or when do you plan to release the actual game?

Thanks for the comments!

Probably no modable game modes appart from gravity and damage. Not sure how any other modes would fit in really. The whole idea would overcomplicate the simple and fun design and probably result in me never finishing it :( You will be able to add new weapons & obsticles.

I'm sticking with the current control scheme - I don't want it to be like RDKF although this was the inspiration for my robot demo.

I doubt I'll be releasing any new demo's until the game is complete, other than to a select few for beta testing. Maybe a few videos as I progress. I'm not working to any timescales as I'm busy with work most of the time and I'm a one man band.

[X-iDuex]

Thanks for the comments!

Probably no modable game modes appart from gravity and damage. Not sure how any other modes would fit in really. The whole idea would overcomplicate the simple and fun design and probably result in me never finishing it :( You will be able to add new weapons & obsticles.

I'm sticking with the current control scheme - I don't want it to be like RDKF although this was the inspiration for my robot demo.

I doubt I'll be releasing any new demo's until the game is complete, other than to a select few for beta testing. Maybe a few videos as I progress. I'm not working to any timescales as I'm busy with work most of the time and I'm a one man band.

I'ld be glad to help, I could beta test and I can do some artwork, I actually edited the textures in the demo.

This is a good engine and it looks like a lot of different games could be made from it, that's why I suggested the game mode modding.

Are there any plans to make any sorta multiplayer type modes? Is there going to be a sandbox mode? And since your keeping the control scheme the same way are you going to have to use the "force" or telekinesis to pick up and use weapons?

Please tell me if my questions are getting on your nerves, I really appreciate you taking the time to answer them though, thanks.

[sakiskid]

I'ld be glad to help, I could beta test and I can do some artwork, I actually edited the textures in the demo.

This is a good engine and it looks like a lot of different games could be made from it, that's why I suggested the game mode modding.

Are there any plans to make any sorta multiplayer type modes? Is there going to be a sandbox mode? And since your keeping the control scheme the same way are you going to have to use the "force" or telekinesis to pick up and use weapons?

Please tell me if my questions are getting on your nerves, I really appreciate you taking the time to answer them though, thanks.

that sounds awsome, now i know he hasnt realesed it yet....

[STEVIE G]

I'ld be glad to help, I could beta test and I can do some artwork, I actually edited the textures in the demo.

This is a good engine and it looks like a lot of different games could be made from it, that's why I suggested the game mode modding.

Are there any plans to make any sorta multiplayer type modes? Is there going to be a sandbox mode? And since your keeping the control scheme the same way are you going to have to use the "force" or telekinesis to pick up and use weapons?

Please tell me if my questions are getting on your nerves, I really appreciate you taking the time to answer them though, thanks.


I'll keep you in mind for the testing. I'm gonna take a stab at doing all the artwork. With the simple siloutte design I should be ok. Backgrounds may be a bit of a problem for me.

Sorry, no plans for network multiplayer but up to 4 teams on one machine is what I'm aiming for. Obviously there will be AI controlled teams in there too. You can fully play the game just now. Requires alot of balancing on the object damage etc.. though.

There will be a sandbox mode to allow you to test your levels and new weapons.

Yes, the whole game is centered around picking up and throwing stuff using your force / telekinesis.

No worries on the questions, I'm just glad that someone can see the same potential as me without moaning about the speed. The game idea would be unplayable if it moved any faster.

[X-iDuex]

I'll keep you in mind for the testing. I'm gonna take a stab at doing all the artwork. With the simple siloutte design I should be ok. Backgrounds may be a bit of a problem for me.

Sorry, no plans for network multiplayer but up to 4 teams on one machine is what I'm aiming for. Obviously there will be AI controlled teams in there too. You can fully play the game just now. Requires alot of balancing on the object damage etc.. though.

There will be a sandbox mode to allow you to test your levels and new weapons.

Yes, the whole game is centered around picking up and throwing stuff using your force / telekinesis.

No worries on the questions, I'm just glad that someone can see the same potential as me without moaning about the speed. The game idea would be unplayable if it moved any faster.

Gish is kinda that way, until you get really good you can't play in turbo mode.

For those people who don't like the speed would it be easier turn off the FPS limiter, instead of going in and recoding junk? Are you going to implement any sort of gamepad/joystick control?

By the way do you have a YIM screen name?

[668.5]

I dont like the zoom while you move your mouse.

[STEVIE G]

I dont like the zoom while you move your mouse.


Agreed. I was never happy with it either. I fixed it last week so that the zooming it's much smoother and only happens when necessary so that player and cursor are on screen. You can also doubleclick to move the cursor back to the center of the player.

There'll be a few options for mouse sensitivity and speed and ability to change the gravity and world drag for those who don't like the speed. Like I said though - the speed is deliberate and a slower pace suits the game and allows you to do some really fancy moves.

There will only be mouse control - analogue joypads don't work well.


By the way do you have a YIM screen name?


You what?

[Xerxes]

Can I help test it? Now that I know what the game is about, the speed feels a lot easier. It looks/sounds great!

[jojojomon]

You what?
If you are whating when he said YIM he means Yahoo Instant MEssenger

[Juggz]

Yahoo? Ha... Thats got more adverts in it than MSN

[Jokker]

Yahoo? Ha... Thats got more adverts in it than MSN

Er. No. It's got one small banner at the bottom and that's it.

Also, I've been wondering, you're using Blitz right? I didn't know you could use objects in it, do you need a plugin or something?

[STEVIE G]

Also, I've been wondering, you're using Blitz right? I didn't know you could use objects in it, do you need a plugin or something?

Yes, I'm using Blitz. Not sure what you mean by 'objects'? I'll be making the editor in Blitz which saves a physics object in my own format, which will then be available in the landscape generator as a placable object. Why would you need a plugin?

Here's a quick video update ...

http://www.youtube.com/watch?v=FCDPC8F1KI8

http://www.steviegoodwin.plus.com/images/Screenshot037.PNG

[Beakless]

Nice, I reallt like the new creatures and objects, bu the whole thing still looks like it's in space... Floating around...

This would be a great engine for a 2d shoot'em up in space! ;)

[STEVIE G]

Nice, I reallt like the new creatures and objects, bu the whole thing still looks like it's in space... Floating around...

This would be a great engine for a 2d shoot'em up in space!


Remember that fraps seems to record at half speed. Why not have a try at writing your own more stable and realistic version?!

[STEVIE G]

A shark!!

http://www.steviegoodwin.plus.com/images/Screenshot041.PNG

[Beakless]

Why not have a try at writing your own more stable and realistic version?!


WOAH! No, I wasn't saying that it's rubish, it was just constructive criticism, but if it's ment to be like that then that's cool...

Btw: "FRAPS records at 1/2 a second"? No, it really is that float-y, c'mon, you have to admit it....

[fusion1224]

The whole point of making it floaty is because it suits the gameplay, I'm pretty sure this is what Stevie is saying. And anyway I don't mind it.

You're really coming along well with this, well done. Need any help with anything?

[X-iDuex]

Do the sharks rip off limbs?? And is there going to be blood in the final version? Cause that would kick ASS!

[STEVIE G]

@ Fusion - send me an e-mail Stevie@StevieGoodwinDOTplusDOTcom and I'll fill you in with some details.

@ X-iDuex - limbs only come off when you're pan breed atm but the shark does major damage! There is already blood - albeit purple blood!

[iWinTies]

STEVIE G, so you're using a custom made physics engine? Do you plant to release it, at any time? I've been looking for a good 2D physics engine and have come up short. ODE/OPAL, Newton, Tokamak, Flatland, DyNaMo, Chipmunk, etc, they aren't (2D) user-friendly. And, I'm just not good enough with Math to create my own. :(

[STEVIE G]

STEVIE G, so you're using a custom made physics engine? Do you plant to release it, at any time? I've been looking for a good 2D physics engine and have come up short. ODE/OPAL, Newton, Tokamak, Flatland, DyNaMo, Chipmunk, etc, they aren't (2D) user-friendly. And, I'm just not good enough with Math to create my own. :(

Yes, it's my own engine. It works only on conjunction with Blitz3d though. No immediate plans to release it but maybe once I'm don with this game.

Stevie

[X-iDuex]

Why do yah use Blitz3d and non of the other Blitz programs, I'm curious because I know how to code in BASIC so it should be easy to get into.

[STEVIE G]

I used Blitz3d rather than Plus etc.. because the engine is actually 3d but displayed in 2d, the reason being is so that I can have hardware accelerated graphics and make use of the inbuilt 3d collision system.

[X-iDuex]

I used Blitz3d rather than Plus etc.. because the engine is actually 3d but displayed in 2d, the reason being is so that I can have hardware accelerated graphics and make use of the inbuilt 3d collision system.
:D I just looked at some example code and I noticed the collision command thing and i thought to myself how much easier that would make coding, and that's great it doesn't just use the CPU.

[Relys]

Great demo! Runs about 45ish fps on my really low end computer!

[fusion1224]

Ha, I didn't realise until now that Blitz3d costs money to develop with. Personally I'd rather go with the full DIY option (as in writing all the collision detection and graphics code yourself) because it allows for more customisability, but that's just me.

[STEVIE G]

Ha, I didn't realise until now that Blitz3d costs money to develop with. Personally I'd rather go with the full DIY option (as in writing all the collision detection and graphics code yourself) because it allows for more customisability, but that's just me.

I think you're misunderstanding me here. I had to implement my own single surface mesh system for handling the graphics as 2d in 3d and also my own collision system for body/body intersection and all responses etc... All blitz does is tell me when a collision with the landscape occured and prevents penetration. It would be easy for me to write my own intersection functions but why re-invent the wheel. The whole thing is very customisable so I don't get your point.

[Xerxes]

Any release date updates? I'm anxiously awaiting this game!

[STEVIE G]

I've been working on the GUI for the last couple of weeks. It's the part I always hate but need it for the editors and menu's etc... Once it's done I can get back to doing the fun stuff :)

I'm almost there ...

http://www.steviegoodwin.plus.com/images/Screenshot048.PNG

[Maltor124]

That looks great! It's obvious you're making a lot of progress. Thank god you didn't abandon this. It was fun.

[MrFalco]

This looks awsome, i hope you add online to it or maybe a website community so you can download mods and such for it. :D

[MrFalco]

sorry for the double post but any progress on the game?

[STEVIE G]

I'm still working on the world editor which won't be ready for a few weeks at least. I've just started a new job so don't have much spare time. I also find creating editors, GUI stuff etc.. the most boring part of any project. It will be fully integrated with the main game though. The good thing is that it's very simple to use and you can create a new level in a couple of minutes.

Once that's done I can continue with the fun stuff of adding new objects / creatures etc.. and tweaking the gameplay.

There won't be mutiplayer online but up to 4 players on one machine with variable AI.

Stevie

[X-iDuex]

I'm still working on the world editor which won't be ready for a few weeks at least. I've just started a new job so don't have much spare time. I also find creating editors, GUI stuff etc.. the most boring part of any project. It will be fully integrated with the main game though. The good thing is that it's very simple to use and you can create a new level in a couple of minutes.

Once that's done I can continue with the fun stuff of adding new objects / creatures etc.. and tweaking the gameplay.

There won't be mutiplayer online but up to 4 players on one machine with variable AI.

Stevie

That sounds great, I think I might have already asked this but will there be gamepad/joystick functionality so like I can use the gamepad to control my character while my friend uses the mouse to control his?

[STEVIE G]

That sounds great, I think I might have already asked this but will there be gamepad/joystick functionality so like I can use the gamepad to control my character while my friend uses the mouse to control his?

Nope. The game is turn based. Using a joypad, even if analogue would not be precise enough to be fun so mouse only I'm afraid.

[SgtBob]

I don't mean to annoy you with yet another one of these comments, but I'm getting 45-50 FPS and everything still seems slow.

It feels like it could get sped up with no harm.

I'm not asking double speed or anything, just enough so that it doesn't take almost a full minute for a capsule to fall to the water..

Other than that, good work.

Nice progress. :]

[STEVIE G]

I don't mean to annoy you with yet another one of these comments, but I'm getting 45-50 FPS and everything still seems slow.

It feels like it could get sped up with no harm.

I'm not asking double speed or anything, just enough so that it doesn't take almost a full minute for a capsule to fall to the water..

Other than that, good work.

Nice progress. :]

If you look at the above gui image then you'll see that gravity will be adjustable, even to the point that there is none. Obviously, you'll use up more 'force' throwing around objects when the gravity is higher and you'll also have less time to react to things. When the game is released you'll see what I mean.

Level Editor is almost done ...

http://www.steviegoodwin.plus.com/images/Screenshot051.PNG

[Maltor124]

Oh. My. God. I really can't wait for this game, now. I love scene editors.

[Empyreal]

I just got a new computer and tried this, and it's a little slow, but I think the speed's perfect for the method of movement. Try catapulting a ragdoll and throwing objects at even three times the speed. I don't think it'd work.

Although I would like it for the challenge. :D

Also, I see it at 50 (99 or 100) constantly, and since the first doesn't ever change, I'm thinking the second one would be the FPS. Which with it being that high would be why I think it's fast enough.

[MrFalco]

i really want to play this game now, how much percentage would you say your done with? from the screenshots it looks to be at least 50% done

[Xerxes]

Fast work on the level editor, and it looks like it'll be more fun than most of them. *cough* Lugaru map editor *cough*

So, what the above guy said. Will it be ready before the eleventh of July? That's the day I get back from my cousin's wedding.

[STEVIE G]

Will it be ready before the eleventh of July?


Not a chance. I would say that the game is around 60-70% done. While the game is fully playable now it'll be a couple of months before I integrate the GUI and add some more objects / features.

The level editor will be extremely easy to use. You can honestly create a new level in a couple of mins - with most of that taken up with the object placement.

I refuse to get into the speed thing again - it's perfect for the gameplay - end of discussion.

Stevie

[Croix]

THAT LOOKS EXTREMELY GREAT!!! I CAN'T WAIT!!! AND BY THE WAY THE SLO-MO THING IS COOL IN MY OPINION!!!





Is the games name stramash??? what does that mean???

[STEVIE G]

Thanks - another who likes the speed. I'll convert you all when you play it ;)

Stramash is actually an old scottish name for a fight or battle. I like the name "Pain" that the recent ps3 game is called .. it's quite catchy. The name is provisional atm - can anyone think of anything better?

I've finished the landscape editor now. The levels will be 2.25 x area compared to the original level. Now onto the object placement part. I'll post a few screenies once my ftp is back up. Good news is that I have a very talented artist friend who is going to help me out with a few nice backgrounds.

http://www.steviegoodwin.plus.com/images/Screenshot055.PNG


http://www.steviegoodwin.plus.com/images/Screenshot059.PNG

[kingofthespill]

I use the standard position based algorithm. For better air resistance don't use a viscous system described in Jacobsens paper - you need to model it correctly ... using DragCoef * ( Speed * Velocity ). Where Velocity is implied by new - old positions. I don't use any reflection in my collision with the landscape - this is all handled automatically by Blitz3d. I'm using 2d in 3d and the landscape is actually a mesh. Blitz collisions can be set to automatically prevent penetration so I don't even need to project out of collisions. I also have some code to prevent objects breaking by getting inverted due to a large impact. Only 3 iterations of the relaxation loop are required as a result.

I know that this collision method does not work with single particles but the idea of projecting both old and new positions onto the collision plane is how I'd do it. This is the same as what you describe.

Stevie

I really like the demo and had some thoughts on what you were saying about drag. This may seem like I am rambling, but consider the comments food for thought.:)

I recently ran into a problem when I was playing around with a home-brew physics simulation, something very primitive that simulated gravity + drag on a single 2d box. I used the less accurate Euler method to calculate position, velocity, and acceleration/force; i.e. just figure out what drag is and treat it as a constant for one time slice. The problem occurred when the object's surface area got small and the terminal velocity went way over 100m/s; then the drag force oscillated wildly, causing the object's position to "blow up" mathematically.

Even with a more accurate verlet equation I suspect the same problem will occur. Perhaps it is a problem at only some ridiculously high velocity, like if the smallest possible object collides with the 1-ton weight, both originally moving at high speeds. It seems simple to test, and perhaps keep the smallest object large enough so it never "blows up". Either that or have some other kind of drag calculation for objects with speed > 100 x terminal velocity which would work arbitrarily length time-slices. That would be a nice function to have :cool: .

I am guessing that the Euler and verlet equations are ultimately similar (and inaccurate) when forces are at their most discontinuous from time slice to time slice.

It may be academic, but I was curious if you tried using the Euler approach first before switching to verlet. Leaving out collisions, when I simulated gravity and drag (with Euler) the results seemed accurate, even using 1/10 of a second time slice.

Unrelated to the physics, I get 36 frames per second :( and it seems floaty. Perhaps you could have options for the user to control that provided their machine is faster than mine :rolleyes: .

I like the idea of this as an engine or even a rapid game development environment. There are many non-programmers who want to create their own ragdoll critters, machines, etc. I am not sure how awful of a performance hit adding textures would be, but some people might want that.

Another idea which I am not clear if you have already implemented is a goal mode, like survive against a bot or fight pvp.

[STEVIE G]

recently ran into a problem when I was playing around with a home-brew physics simulation, something very primitive that simulated gravity + drag on a single 2d box. I used the less accurate Euler method to calculate position, velocity, and acceleration/force; i.e. just figure out what drag is and treat it as a constant for one time slice. The problem occurred when the object's surface area got small and the terminal velocity went way over 100m/s; then the drag force oscillated wildly, causing the object's position to "blow up" mathematically.


This doesn't happen to me as I'm giving each body a preset drag cooeficient rather than calculating it based on surface are etc.. I also have a constant timestep. This is a must otherwise you'll run into all sorts of problems.


Even with a more accurate verlet equation I suspect the same problem will occur. Perhaps it is a problem at only some ridiculously high velocity, like if the smallest possible object collides with the 1-ton weight, both originally moving at high speeds. It seems simple to test, and perhaps keep the smallest object large enough so it never "blows up". Either that or have some other kind of drag calculation for objects with speed > 100 x terminal velocity which would work arbitrarily length time-slices. That would be a nice function to have .


Your talking mainly about collisions rather than the integration method here. I have no issues with the simulation blowing up for the reasons I mentioned above.


I am guessing that the Euler and verlet equations are ultimately similar (and inaccurate) when forces are at their most discontinuous from time slice to time slice.


They are similar but as verlet integration is position based the accumulated floating point errors aren't as bad as that of euler integration. Blitz's floats are single precision but work just fine for me.


It may be academic, but I was curious if you tried using the Euler approach first before switching to verlet. Leaving out collisions, when I simulated gravity and drag (with Euler) the results seemed accurate, even using 1/10 of a second time slice.


Nope - I've always used the verlet method , it's accurate enough / fast and rotations are automatically handled. If you're worried about the speed there will be options to increase gravity otherwise get yourself and decent PC. ;) Obviously the game is calc intensive so relies of the speed of your CPU. The speed of your GPU is pretty much irrelevant.

[Xerxes]

This thread is old. There is a new thread about this game entitled "Stramash." Please don't necro.

[Maltor124]

What the? I thought this was locked? Anyways, yes, this is an old thread. You can find the thread for the full version of the game he's working on over here (http://www.fun-motion.com/forums/showthread.php?t=741). Also, I'll move that post you made over there. :)