Consequence Interceptor Concept

Thinking about improving the Artwork in Vega Strike, or making your own Mod? Submit your question and ideas in this forum.

Moderator: pyramid

travists
Expert Mercenary
Expert Mercenary
Posts: 893
Joined: Thu Jul 08, 2010 11:43 pm
Location: Sol III North American Continent

Re: Consequence Interceptor Concept

Post by travists »

Deus Siddis wrote:This also ties back into the discussion of weapon systems being externally differentiable, since maneuvering thrusters will tend to be as big as or bigger than guns, if you can hide and protect them inside a hull section the same should be possible with at least non-turreted weapons.
Which is not to say that such weapons would not be mounted such that they are visible for access/aesthetics reasons. What is with the If one ship has visible weapons all should and if one does not none can anyway? Some ships would, by nature, have highly visible changes with upgrades others not.

Let's take the Ox. Given it's current in-game appearance I would expect it to show nearly every upgrade. Cargo expansions, crew quarters, guns all would be hung off the skeletal frame. Anything that size would have massive RCS thrusters, and given the general layout, likely on booms well outside the core of the ship. The mule just screams internal and compact, So guns in bumps on the hull and only ports for RCS thrust makes sense.

I do caution you, Deus, an artist can be his worst critic. What you see as glaring defects may be seen by others as amusing, even endearing, quirks. You know how it is supposed to look and every thing that is wrong, that does not mean that everyone does. Then there is the dodo.....
klauss
Elite
Elite
Posts: 7243
Joined: Mon Apr 18, 2005 2:40 pm
Location: LS87, Buenos Aires, República Argentina

Re: Consequence Interceptor Concept

Post by klauss »

Deus Siddis wrote: Due to the asymmetry of the weapon placements on the derivative, its AO bake isn't really accurate and the situation would get worse with further greebling.
I see. There are better ways to do the unwrapping with overlapping islands. As you noted, by respecting symmetry. But even in this broken case, the baking procedure can be made to output the lightest pixel among the overlapping islands, resulting in a conservative AO bake. It works, not as well as a properly unwrapped model, but it works.
Deus Siddis wrote: About when is PRT mapping going to become important?
We don't have the tools to do the bakes yet outside of windows. ATM, the only tools for PRT baking are Blender (with a highly convoluted procedure I wouldn't recommend), and xNormal (which only runs on Windows).

When that is fixed (ie, by getting an xNormal for linux or by coding our own PRT baking tools), the techniques are already there to use so we could then work on PRT maps for all units.
Deus Siddis wrote:
klauss wrote: There's nothing impossible on the physics side. A navigational computer can easily coordinate the irregular thrusters to get the required thrust vectors. The Derivative does have retro thrusters quite prominent, and turning on all axis can be handled with coordinated action of lateral ones. The engine can right now show that, it was coded for the B5 mod which never prospered.
What I'm thinking about though is that the derivative doesn't have thrusters oriented to make it strafe directly left or right efficiently.
+45° thruster and -45° thruster both at the same time equal 0° thruster with 1.414 thrust. Less efficient, but workable. And nobody said the thrusters have to be placed at 90° of the hull, they can be slanted and thus have no issue.
Deus Siddis wrote:
klauss wrote: Maneuvering thrusters don't need geometry. Look at the space shuttle. They're quite inconspicuous there. Maybe a hint of them in textures.
The difference is power. The shuttle's RCS output in miniscule. In VS lateral thrusters can push your craft from 0.1 to 10 gs left, right, up and down. The space shuttle can only push itself at 3gs going forward during liftoff with the help of hot burning liquid chemical rockets supplied by a massive disposable external fuel tank and two massive solid fuel boosters.

So if we go back to TB's theory of the graphics reflecting the simulation, maneuvering thrusters must be massive and thus at least somewhat prominent.
That's nitpicking. The stats are there and unless you want to change them, you have to live with high-power RCS thrusters. It makes a whole lot more sens to change the stats to reflect their actual power than to change the model to make them massive. There's no point in having massive orthogonal thrusters, when you can just turn the ship and point your big thruster in the proper direction. The only massive thrusters I'd include on a model are forward and retro.
Oíd mortales, el grito sagrado...
Call me "Menes, lord of Cats"
Wing Commander Universe
Gungnir
Mercenary
Mercenary
Posts: 98
Joined: Thu May 06, 2010 5:57 am

Re: Consequence Interceptor Concept

Post by Gungnir »

travists wrote: Theoretically, we should already be there "10-20k". I think the question to be asked is what ships do not meet the existing standard be fore we try to up the standards.
Theoretically being the key term. I haven't played 0.5.1 or SVN (can't due to driver issues, stupid Mesa), but in 0.5.0 there were maybe a couple capships that met the fighter polycount guidelines, no fighters got anywhere close (not that I could tell at least, they looked ~2000 at best). While I agree that some models need to be redone more than others, I won't argue if someone wants to get modeling and improve something. Also, if we're going to up the standards, it would be best to do it now and start using a standard that will work and look decent several years from now, before we start the massive project of remodeling almost every asset in the game. As I said before, it's easier to scale a 100k model down to 15k than to remodel a 2k up to 15k.

On the subject of thrusters, I feel that the size/visibility should be decided by the art style and ship class; for instance, an Ox's thrusters might be smaller and weaker than an interceptor's, proportionately (small, slow adjustments vs. rapidly changing directions and dodging during dogfights), and I would suspect that factions like the Highborn would try a lot harder to hide the thrusters than poor factions like the Forsaken and ULN would.
~Gungnir
segfault wrote:if I was actually in space I'd totally be throwing on autopilot and relaxing in the back during the trip, sipping space wine and listening (rlaan?) jazz.
Rig: i5 2500k @ 5ghz, 2x OCZ Agility 3 120gb SSD boot drives, AMD Radeon HD 7950 @ 1100/1575 (Catalyst 12.1 Linux and 12.3 Windows), dual-boot Fedora 16 KDE and Windows 7 Pro
travists
Expert Mercenary
Expert Mercenary
Posts: 893
Joined: Thu Jul 08, 2010 11:43 pm
Location: Sol III North American Continent

Re: Consequence Interceptor Concept

Post by travists »

Fair enough
1: Are the current standards sufficient, or do they need to be reworked?
1a: If they need to be reworked, how high can we realistically go?
1b: If they are sufficient, how many ships are up to standards?

2: Does displacement add enough to be worth the effort to include it?

3: Ogre is batted around a lot, and seems to be making progress. Is it better to focus on the game side of the engine and outsource the graphics engine or should the in-house graphics remain?
3a: If we switch, does Ogre add enough features to make new models secondary until the switch is complete?

Once we know these, what ships need what will be somewhat easier.
klauss
Elite
Elite
Posts: 7243
Joined: Mon Apr 18, 2005 2:40 pm
Location: LS87, Buenos Aires, República Argentina

Re: Consequence Interceptor Concept

Post by klauss »

travists wrote:Fair enough
1: Are the current standards sufficient, or do they need to be reworked?
1a: If they need to be reworked, how high can we realistically go?
1b: If they are sufficient, how many ships are up to standards?
Well, things may be a bit misleading in this regard. Take the Ox:
claudiofreire@klausstop:~/svn/vegastrike/trunk/data/units/vessels/Ox> ~/svn/vegastrike/trunk/vegastrike/build/objconv/mesh_tool ox.obj oxtmp.bfxm obc
0_0: 3960 faces, 14376 vertices, 0 lines, 560 tris, 3400 quads, 15280 indices
1_0: 372 faces, 1265 vertices, 0 lines, 144 tris, 228 quads, 1344 indices
2_0: 1379 faces, 4939 vertices, 0 lines, 294 tris, 1085 quads, 5222 indices
3_0: 3096 faces, 9511 vertices, 0 lines, 1432 tris, 1664 quads, 10952 indices
4_0: 72 faces, 288 vertices, 0 lines, 0 tris, 72 quads, 288 indices
5_0: 418 faces, 1530 vertices, 0 lines, 76 tris, 342 quads, 1596 indices
6_0: 72 faces, 288 vertices, 0 lines, 0 tris, 72 quads, 288 indices
7_0: 672 faces, 2560 vertices, 0 lines, 128 tris, 544 quads, 2560 indices
Total faces: 10041
Total indices: 37530
It performs horribly slow on intel hardware, it has quite a high polycount, but doesn't look like it at all.
So, does this mean polycounts have to be increased? Nope. Not IMHO. See, it's quite an art to make each poly count, and I don't think it's been done on the Ox. It has way too many polies for the amount of visible detail (though it's not a bad model, just inefficient).

IMHO, the standards are Ok. Since they've been discussed and put on the wiki, none (or a handful) models have been added in accordance to them. The standards don't need fixing, they need to be followed.
travists wrote: 2: Does displacement add enough to be worth the effort to include it?
I don't think so. We haven't even fully exploited normal maps yet, and you're thinking of even more complex tecniques. Exploit normalmaps first, parallax mapping, and before going the displacement way (which is a sure way towards really onerous system requirements), be sure you've exhausted the simpler techniques first.
travists wrote: 3: Ogre is batted around a lot, and seems to be making progress. Is it better to focus on the game side of the engine and outsource the graphics engine or should the in-house graphics remain?
3a: If we switch, does Ogre add enough features to make new models secondary until the switch is complete?
Ogre doesn't provide art. So no. It does not make new/better models secondary. Think about it, it's like saying that since you got a new, really nice and professional brush, you no longer need to paint.
Oíd mortales, el grito sagrado...
Call me "Menes, lord of Cats"
Wing Commander Universe
klauss
Elite
Elite
Posts: 7243
Joined: Mon Apr 18, 2005 2:40 pm
Location: LS87, Buenos Aires, República Argentina

Re: Consequence Interceptor Concept

Post by klauss »

Gungnir wrote:As I said before, it's easier to scale a 100k model down to 15k than to remodel a 2k up to 15k.
I disagree. It's in fact not easy to go either up or down.

Take the Ox. It needs LODs for our friends with onboard LPUs (lame processing units or intel). The model has 10k polies (mostly quads, so proably 20k tris). Get us a LOD with 5000 tris, and another with 1000. And, if you're up for it, a superlod with 60k tris, to make the cruiser budget.

Then tell me whether it was easy or hard. The good thing, you'll both learn and contribute in the process :D
Oíd mortales, el grito sagrado...
Call me "Menes, lord of Cats"
Wing Commander Universe
Deus Siddis
Elite
Elite
Posts: 1363
Joined: Sat Aug 04, 2007 3:42 pm

Re: Consequence Interceptor Concept

Post by Deus Siddis »

klauss wrote: I see. There are better ways to do the unwrapping with overlapping islands. As you noted, by respecting symmetry. But even in this broken case, the baking procedure can be made to output the lightest pixel among the overlapping islands, resulting in a conservative AO bake. It works, not as well as a properly unwrapped model, but it works.
I used that second technique also on the Derivative. I manually broke the model into pieces and only saved the bake from the pieces that had non-unique (conservative) shadowing.
klauss wrote:When that is fixed (ie, by getting an xNormal for linux or by coding our own PRT baking tools), the techniques are already there to use so we could then work on PRT maps for all units.
Wow, then you are much closer than I thought.

I could never understand why xNormal remains an closed source project. Someone probably would have done a linux port by now if it wasn't for this.
klauss wrote: +45° thruster and -45° thruster both at the same time equal 0° thruster with 1.414 thrust. Less efficient, but workable. And nobody said the thrusters have to be placed at 90° of the hull, they can be slanted and thus have no issue.
But it inefficiently uses eight thrusters to strafe left or right.
klauss wrote: That's nitpicking. The stats are there and unless you want to change them, you have to live with high-power RCS thrusters. It makes a whole lot more sens to change the stats to reflect their actual power than to change the model to make them massive. There's no point in having massive orthogonal thrusters, when you can just turn the ship and point your big thruster in the proper direction. The only massive thrusters I'd include on a model are forward and retro.
Well the same argument can be applied to retro thrusters, which could instead be pointed backward for a better top acceleration. Then you end up with ships which only usefully accelerate forwards (outside of precise docking maneuvers).
klauss
Elite
Elite
Posts: 7243
Joined: Mon Apr 18, 2005 2:40 pm
Location: LS87, Buenos Aires, República Argentina

Re: Consequence Interceptor Concept

Post by klauss »

Deus Siddis wrote: I used that second technique also on the Derivative. I manually broke the model into pieces and only saved the bake from the pieces that had non-unique (conservative) shadowing.
It's easier to put all bakes piled up in gimp as layers and set them to "lighter" (or max or whatever it's called) mode ;)
Deus Siddis wrote:
klauss wrote: That's nitpicking. The stats are there and unless you want to change them, you have to live with high-power RCS thrusters. It makes a whole lot more sens to change the stats to reflect their actual power than to change the model to make them massive. There's no point in having massive orthogonal thrusters, when you can just turn the ship and point your big thruster in the proper direction. The only massive thrusters I'd include on a model are forward and retro.
Well the same argument can be applied to retro thrusters, which could instead be pointed backward for a better top acceleration. Then you end up with ships which only usefully accelerate forwards (outside of precise docking maneuvers).
Yes, but you could say that breaking is important enough and may be required with little enough warning to warrant its own set of thrusters ;)

But yes, only the forward thruster really makes sense.
Oíd mortales, el grito sagrado...
Call me "Menes, lord of Cats"
Wing Commander Universe
Deus Siddis
Elite
Elite
Posts: 1363
Joined: Sat Aug 04, 2007 3:42 pm

Re: Consequence Interceptor Concept

Post by Deus Siddis »

klauss wrote: But yes, only the forward thruster really makes sense.
It might be cool if one of the major factions mostly agreed with this statement in all its shipbuilding, like the Aera.

But as far as practicality goes, for ships that rely mainly on fixed weapon systems the ability to accelerate in one direction while shooting in another seems like it makes some sense.
Gungnir
Mercenary
Mercenary
Posts: 98
Joined: Thu May 06, 2010 5:57 am

Re: Consequence Interceptor Concept

Post by Gungnir »

klauss wrote:
Gungnir wrote:As I said before, it's easier to scale a 100k model down to 15k than to remodel a 2k up to 15k.
I disagree. It's in fact not easy to go either up or down.

Take the Ox. It needs LODs for our friends with onboard LPUs (lame processing units or intel). The model has 10k polies (mostly quads, so proably 20k tris). Get us a LOD with 5000 tris, and another with 1000. And, if you're up for it, a superlod with 60k tris, to make the cruiser budget.

Then tell me whether it was easy or hard. The good thing, you'll both learn and contribute in the process :D
Do you have an obj or blend you can shoot my way? I'd love to take a crack at it. And while yes, it's not necessarily easy to change polycounts at all, I've found it easier in my personal experience to reduce poly- and vertex-counts than to add random greebles which may or may not mess up the art style.

Also, I must disagree with you on the issue of tessellation. While we haven't fully exploited other map types yet, there's no sense in not having it. It wouldn't hurt systems too weak to run it, and it could increase visual quality dramatically for those that can. The only negative to it is the effort required to add OpenGL 4.0+ support to the engine (I would lobby for 4.2 while we're at it, partially resident textures could help with textures on planets and other large objects).
~Gungnir
segfault wrote:if I was actually in space I'd totally be throwing on autopilot and relaxing in the back during the trip, sipping space wine and listening (rlaan?) jazz.
Rig: i5 2500k @ 5ghz, 2x OCZ Agility 3 120gb SSD boot drives, AMD Radeon HD 7950 @ 1100/1575 (Catalyst 12.1 Linux and 12.3 Windows), dual-boot Fedora 16 KDE and Windows 7 Pro
travists
Expert Mercenary
Expert Mercenary
Posts: 893
Joined: Thu Jul 08, 2010 11:43 pm
Location: Sol III North American Continent

Re: Consequence Interceptor Concept

Post by travists »

klauss wrote:Ogre doesn't provide art. So no. It does not make new/better models secondary. Think about it, it's like saying that since you got a new, really nice and professional brush, you no longer need to paint.
I was thinking more "I'll get my better paint, canvases, and brushes in a month, I think I'll take a break till they arrive and clean up my studio."
klauss
Elite
Elite
Posts: 7243
Joined: Mon Apr 18, 2005 2:40 pm
Location: LS87, Buenos Aires, República Argentina

Re: Consequence Interceptor Concept

Post by klauss »

travists wrote:
klauss wrote:Ogre doesn't provide art. So no. It does not make new/better models secondary. Think about it, it's like saying that since you got a new, really nice and professional brush, you no longer need to paint.
I was thinking more "I'll get my better paint, canvases, and brushes in a month, I think I'll take a break till they arrive and clean up my studio."
In contrast to paints and brushes, the model you create in blender or any other modeller tool isn't changed by the rendering engine used in-game to show it. So, again, there's no point in waiting for Ogre. Ogre will have normal maps, it will have diffuse maps, it will have everything VS already has, and nothing new.

Ogre does not support displacement maps, HDR, animated meshes or anything out-of-the-box, effort has to be put into it to make it work. Though it's certainly easier with Ogre than coding any of those from scratch, VS's own in-house engine isn't that far behind as most think. It's its artwork which is at fault. In fact, pretty much the same shaders would be used, with perhaps minor modifications to make them compatible.

The only point where Ogre really does provide, out-of-the-box, a great improvement, is in text rendering. Which has nothing to do with ships.
Last edited by klauss on Tue Jan 24, 2012 5:43 am, edited 1 time in total.
Reason: Clarified shaders
Oíd mortales, el grito sagrado...
Call me "Menes, lord of Cats"
Wing Commander Universe
klauss
Elite
Elite
Posts: 7243
Joined: Mon Apr 18, 2005 2:40 pm
Location: LS87, Buenos Aires, República Argentina

Re: Consequence Interceptor Concept

Post by klauss »

Gungnir wrote: Do you have an obj or blend you can shoot my way? I'd love to take a crack at it.
Check the masters
Oíd mortales, el grito sagrado...
Call me "Menes, lord of Cats"
Wing Commander Universe
Deus Siddis
Elite
Elite
Posts: 1363
Joined: Sat Aug 04, 2007 3:42 pm

Re: Consequence Interceptor Concept

Post by Deus Siddis »

travists wrote: I was thinking more "I'll get my better paint, canvases, and brushes in a month, I think I'll take a break till they arrive and clean up my studio."
But the point is moot unless your skilled graphics programmers and skilled content creators happen to be the same people. I think only chuck_starchaser has ever fit that description.

If you tell pure content creators to learn C++ backwards and forwards, learn OGRE's API, learn VS' total lack of an API and then come back and help surgically replace the game's graphics code with OGRE, all in their free time... well not much will come from that.

You are better off thinking of a project's resources as separate and nonexchangeable between code and content (with Python falling between the two IMHO). Content creators model, texture and when applicable animate content that works with the present and future graphics engine(s) that the graphics programmers are at the same time improving. One doesn't need to wait for the other. It's like parallelism for human teamwork instead of multi-core processors, for a perversely techno-centric metaphor.
klauss
Elite
Elite
Posts: 7243
Joined: Mon Apr 18, 2005 2:40 pm
Location: LS87, Buenos Aires, República Argentina

Re: Consequence Interceptor Concept

Post by klauss »

@DeusSiddis: couldn't have said it better myself
Oíd mortales, el grito sagrado...
Call me "Menes, lord of Cats"
Wing Commander Universe
travists
Expert Mercenary
Expert Mercenary
Posts: 893
Joined: Thu Jul 08, 2010 11:43 pm
Location: Sol III North American Continent

Re: Consequence Interceptor Concept

Post by travists »

Got it, for all of it's fans, Ogre is not the finger-paints to Master Series Art Set jump that it has been touted to be. Displacement has it's place, but is not necessary. Current poly guidelines are fine if they are followed efficiently.


Anybody have any idea how to rate existing ships on: fitting the concept for that ship, poly usage, poly count, over all appearance, etc.? I'm sure that there are some obscure ones that anything would be better and others that while in need of some improvement are fine for now. A triage if you will. One of the newer members, Warlord, is fairly impressed with the visuals as they stand. As for modeling, I do ok on some smaller stuff, but it is not my bailiwick and anything I make would likely need to have someone else grebel up.
Gungnir
Mercenary
Mercenary
Posts: 98
Joined: Thu May 06, 2010 5:57 am

Re: Consequence Interceptor Concept

Post by Gungnir »

klauss wrote:Check the masters
Thanks, getting started on it now. Oh my gosh this thing is ugly. Layers upon layers of invisible polys, plus almost 35,000 verts... No wonder Intels have trouble with it. I think I might get it down to ~5000 just from a decent clean-up, no loss of detail needed :P

On a related note, I could use any concept art/references/design docs you can give for greebles once I get to that stage. Maybe I should start a different topic for this project?
~Gungnir
segfault wrote:if I was actually in space I'd totally be throwing on autopilot and relaxing in the back during the trip, sipping space wine and listening (rlaan?) jazz.
Rig: i5 2500k @ 5ghz, 2x OCZ Agility 3 120gb SSD boot drives, AMD Radeon HD 7950 @ 1100/1575 (Catalyst 12.1 Linux and 12.3 Windows), dual-boot Fedora 16 KDE and Windows 7 Pro
Blackbeard
Merchant
Merchant
Posts: 38
Joined: Mon Dec 05, 2011 4:51 am
Location: Great Southern region, Western Australia

Re: Consequence Interceptor Concept

Post by Blackbeard »

Just a question, how do you submit a concept, and who too?

Because reading this, I can't help feeling I've jumped the queue with my models, I've already got three at various stages of completion, but I completely forgot to ask if VS wanted them in the first place. :oops: :oops:

Btw, just to throw my five cents in, I really like the new models for the admonisher and consequence. :D
I haven't seen the new derivative model yet, but I'm sure it will be great. :)
klauss
Elite
Elite
Posts: 7243
Joined: Mon Apr 18, 2005 2:40 pm
Location: LS87, Buenos Aires, República Argentina

Re: Consequence Interceptor Concept

Post by klauss »

Post them drawings in the content vetting forum. Our official canon guy is quite mia, but it's quite possible someone else will help evaluate the idea. If agreed, it will then go to the wiki
Oíd mortales, el grito sagrado...
Call me "Menes, lord of Cats"
Wing Commander Universe
Post Reply