How good is the engine?

#1 Also check out the very good orange and blue engine effects. How hard would it be to create those(they have to turn on and off smoothly)?

#2 those are billboards

#3 Chris, buy the game.

Ask yourself this, how would you create those scenes? What assests would you use? What is the minimum requirements for your base systems?

Now compare this against the FAQ and the game itself.

#4 I'm saving up every penny I can for something big... It won't last long, but its worth it either way, and yes, every little cent counts.

>Ask yourself this, how would you create those scenes?

There are many ways to create them, the best way depends on the engines features. It could really slow down the game if you had to use a sphere with a starmap on it compared to some of the other advanced systems that can save rendering time.

Minimum requirements by the time this game is done(2-4 years) should be a lot higher than todays standard systems. My guess is at least a Pentium3 900 MHZ in two years(then they should easily have the 4.5 GHZ processors out, and many of us using things above 1.5 GHZ).

I'm just asking if this type of massive game IS or ISN'T feasible due to technical engine features, and maybe speed, a simple yes or no would be fine, no reason to go into such detail over it.

I think the engine can do it, and I probably shouldn't be posting this, but it is good to know if there would be any problems, maybe someone would know of something I didn't know. Most of the things in these scenes are simply done, the effects shouldn't be any problem for the engine, right? Also, are there any technical programming problems you all think the team might be up against?

Its best to do some research before jumping into anything, so I'm doing research.

#5 Well, for starters, there is the starfield itself -- obviously, it uses a different method of creating that starfield than is in use in the V12, which just from the screens you can tell uses a 'flat sky' approach.

This is something that could be changed with really a minimum of effort -- you'd just have to know how.

I *think* that Tribes 2 uses multiple LOD models which it swaps in and out as required. Also, the collision detection on the ships would most likely have to be recoded, since the ordinary vehicle physics use a pretty standard bounding box/concave hull approach.

There's a lot more, but look at it this way: the engine you're comparing it to is an engine built from the ground up to accomodate a space sim world, and is built using technology that won't be commonplace for 2 years.

As for what you're asking -- frankly, all of the important stuff could really be coded *fairly* easily, things like a more space-oriented LOD system, a better starfield, removal of the terrain system, and a better-suited object system.

The collision detection kind of gives me a pause, as it's something I've never even played with, but obviously there are people who know it a hell of a lot better than me.

#6 I've had a pretty good ammount of collision detection work, mostly in the general logic, creating my own system, and the applying of different methods to use. Overall, it more depends on what the engine can do built in with collision detection, to a certain point, and if it can do what I want it to, then it shouldn't be a problem.

#7 fred, did you mean the engine effects are billboards?

If so, could you explain that, I don't see how they could be.

Luc, I think someone said the sky in V12 was a semisphere not flat, correct me if I'm wrong.

I think the simple answer to your question Chris is:

yes or no :)

It depends how much work you want to do on the engine (we are talking about V12?), you get the source after all. Also who knows where V12 (V14, V16, etc) will be in two years anyway. Making a space sim was discussed in the V12 forum, with input from GGs.

[sarcasm] In the vastness of space the chance that two objects will collide is so small I don't think you need to worry about CD at all [/sarcasm] :)

But seriously, I think the great thing about doing a future sim is that no-one can say your spaceships aren't realistic. So CD can be simplified dramatically by placing a few minor limitations on spaceship design. I would go for heirarchical AABBs or spheres (or both) and make sure the SS designers know how to optimise for this and give them limitations (eg up to an x-level heirarchy with no more than y BVs per level). If that is done it should be an easy task to add a suitable collision handler to V12. I'm getting off topic now so I'll leave it there.

BTW there is now a discussion forum in the V12 section so I guess thats where V12 discussion should go. [disregard if you weren't talking about V12]

#8 no, i mean the steam trail(or whatever it called) from those spaceship's engines

#9 I think we mean the same thing fred, the blue or orange glow behind the engines?

Surely they can't be camera facing billboards, they have to be viewed from every direction. Even if they were like rogue spear trees I think it would be too obvious.

#10 heh, i though you talked about the whole engine..heh.

regarding the billboard, why not? camera facing billboard always facing camera, (thats why they called "camera facing"). as long as it's always facing camera, it can achieve the glow effect you want.

#11 fred is, I believe, correct that the engine exhaust plumes and their 'lens flares/detector saturation' effects can simply be done using camera facing bilboards. Each engine appears to have one or more plumes, each plume having a central axis about which they are symmetric. Thus when you move around the scene they appear consistent, unlike a tree billboard (which isn' symmetric about it's trunk). I have programmed such, and they are both fast to draw and simple to program.

As to the sky, I would guess a simple skybox would be sufficient. I have tried high resolution versions of these and they look great. Tribes/V12 uses a hemisphere and so would have to be reprogrammed, but skyboxes are extremely simple. It's the artwork which would take the time.

Space sims are, however, fundamentally different beasts to tribes-like FPS games. V12 may not have full freedom of orientation for objects, so you'll need to add an ortho-normal coordinate system or quarternions. Collision detection is relatively straight forward for like size objects, but what about a small fighter approaching a massive cruiser? Having said that, V12 will offer a host of features which will make the game easier to program than starting from scratch.

Good luck, your project idea is an ambituous one if you search for Eve-like detail (in artwork more than programming).

Doug. EnkiSoftware Limited

#12 hehehe funny

I think youll find that would be upto the artist, not the engine

#13 Doug, I thought that might be right. ie not completely camera facing but constrained to one axis.

But, wouldn't you see an edge where the plume meets the engine, and how would it look when you are nearly directly behind the SS, wouldn't you see that it has no depth? If that is what they have done, they've done it very well. Don't get me wrong I think you're probably right but I would like to know how they avoid these problems.

#14 "and how would it look when you are nearly directly behind the SS"

I was wondering the same thing...

#15 for a fast moving object, like these space ships, the chance that a player will see the trail directly from behind is very very small, if it happened, it will suck, but mostly for a short period of time, which no one really notice.

if look directly from behind, you will notice a clean cut by the near clip-plane on the steam trail, while the engine 'glow' will still look alright as it still facing the camera.

#16 For a more in-depth view of this technique (possibly not the one used), discussing your concerns, it works roughly like this:

The plume has an axis running from inside the exhaust to it's end, and a width somewhat smaller than the exhaust. The game draws a rectangle constrained to lie along this axis, but facing the camera as much as possible. By drawing from inside the exhaust geometry, the 'cut off' is not just a straight line.

In addition, another transparent polygon is drawn in the exhaust system at the end of the plume and perpendicular to it. This is so that when viewed down the axis of the plume (or close to the axis), an appropriate glow effect is observed. Care must be taken that parts of this polygon which overlap the previous one are removed, so that the blending does not add them together and show the straight lines where they intersect on the screen. This is calculated by the CPU as the depthbuffer is not used to store the depth of transparent polygons such as this (although a modified technique could use the depth buffer if coded correctly).

Halo's/lens flare etc can then be drawn as required if the exhaust is visible as completely camera facing polygons (triangles or quads depending on whether you are concerned about fill rate or vertex rate).

Hope this helps!

Doug EnkiSoftware Limited

#17 If any of you own Homeworld, play a mission and zoom into the resource collector or any fighter-sized craft as it is moving and pay close attention to the exhaust trail. Rotate around the ship, viewing it at various distances. It looks like what is essentially a solid object that has a triangular cross section, which seems to be animated to rotate along its length.

Jay daSilva
Thera Studios

#18 >for a fast moving object, like these space ships, the chance that a player will see the trail directly from behind is very very small, if it happened, it will suck, but mostly for a short period of time, which no one really notice

Ships can follow eachother for long ammounts of time in space sims, so you would be able to see this a lot.

The lens flare can be handled pretty easily in the artwork sense, and programming shouldn't be that hard either, other than that the lens flare has to smoothly grow in color shape and size to look real, which could cause problems with it being outside artwork and not more of a programming effect.

#19 Lens flare, engine plumes, etc. are the least of your worries in terms of making something that looks close to the screenshots you linked to - there's alot of work there.

Detailed, original, and good-looking spaceships are going to take a serious amount of work to create. The approach used in I-War 2 for example, was to create a cinematographic quality high poly model, then use this to create the texture of the low poly model used in the game. Such an approach makes the textures look real, rather than just paint. However, it's alot of work both in artists time and tool building time.

High quality backgrounds (skyboxes) also require alot of artistic time and talent, in addition to a quality 3D package to develop the images required.

This isn't to say that it's not possible, especially if you go with the start small and build things up approach.

One detail which V12 may/may not be suited for is lighting. This can make or break space games. I would use true vertex lighting, and look for ways to handle self shadowing (which will make the large ships look magnificent). But apart from all the tech and artwork, the make or break issue is getting a realisable fun design.

Good luck!

Doug EnkiSoftware Limited

#20 I don't think we would have a problem making unique ships....