The Tank 2.0 - Evolution of Tank Modelling?

View previous topic View next topic Go down

In a future Tank sim I'd like to see?

33% 33% 
[ 1 ]
33% 33% 
[ 1 ]
33% 33% 
[ 1 ]
0% 0% 
[ 0 ]
0% 0% 
[ 0 ]
 
Total Votes : 3

The Tank 2.0 - Evolution of Tank Modelling?

Post  Blackwatch on Fri Apr 01, 2011 11:07 am

As my Tank Modelling Mood is constantly increasing, I'd like to start a discussion here with all you tanksim fans how the tanks should evolve - or more direct, what features you'd like to see in a next gen Tanksim / Mod and how the creation process could be improved... and then maybe craft a proof of concept.

1. The Base:
I've came to this when I was looking at my Tiger Model (and saw all the errors I did) but it will serve as a basis for development discussion. So that all of you know what I'm talking about here are two pics:

The Specs:
Polycount: 10.914 Tris
Materials: 3 (Diffuse & Normal) - A 2048² for Hull and Turret (and the other things attached to them), a 2048² for Wheels and Sprocket (and some small leftover things), finally a 1024² for the Tracks.

Assessment:
While the Polycount is really low (Animated hatches aren't possible), the Materialcount and Textureresolution could be considered as a standard. The Object distribution on the Textures is not optimal imho as it does not allow for easy creation of Variants e.g. an early / mid Tiger Model or an Assault Tiger without either adding further Materials or restarting from scratch (I'll get back to this later).

The current situation in Games
T34 vs Tiger
OK Tvt is quite old now, but as we're modding for it we have quite precise data for it. The Tiger has around 25k Polys for the exterior model and around 50k Polys for the interiour. Texture wise its 2x 2048² and 1x 512².

Red Orchestra 2
RO2 currently serves me as the 'State of the Art' Tankgame. As the game is not yet released, there aren't any proven facts about the Tank models available yet. Judging the available images and videos I'll go for 30 to 35k for the exteriour models and around 150k for the interiour. Giving an estimation about the Textures isn't possible but I have doubts that they've gone higher than 2048² in Resolution.

The Future
Imho the increase of Polycount is less a problem as modern engines like the UDK can handle up to 30m Tris per scene (not taking the Performance of the computers and the fact that you need a game around into consideration) so sticking to the estimated values or just slightly increasing them (35 to 40k and 175 to 200k) might be a good path to go esp when considering that there are still normal maps and realtime tesselation is coming.

The texture aspect is more difficult though, as increasing the texture resolution to 4096² exponentially increases the memory consumption and increasing the amount of Textures results in an increase of draw calls but considering that one (roughly) can store 4 2048² Textures for one 4096² thats imo the better path of progress.

But the whole progress thingy leads straight into a dilemma, esp for us modders as every step results in an increased amount of production time - just remember that RO2 is shipped with 'just' two tanks and two more announced to be added with a later patch. So it is imperative to carefully plan the development of Tanks (or assets in general) to maximize the use of synergy effects thus increasing the asset output.

Solution?
My thought on this is to create a set of things common to all tanks (of a fraction) e.g. Headlights, Column Lights, Tools, Tow cables and so on with their own material. While looking at the tanks itself its mainly the tracks / road wheels and / or the superstructure that changes more or less intensivly so an approach may be to make one Material for the lower hull (including the tracks and wheels) and one for the upper hull

2. Enhancing features
Using the increasing polycount to make round wheels even rounder is imho not the correct way of using the available budget as there are several options to improve tank models and gameplay. Some of those ideas I have put into the poll:

Physics enhancements
- Individual track links, like in this example here. I made a quick calculation for my Tiger model and came up that a low poly track - and the correct amount of links - in that case should be possible for around 10k polys although the Tiger track is not the most complicated (guess its the early type King Tiger track)

- Detachable Parts, actually being able to shoot off parts of the tank, like turret bins, tools, hatches up - not just see a flying debris particle effect - up to a fatal ammo explosion

Optical enhancements
- More war like Machines - Tanks in current sims look like they were coming directly from the factory line and not warlike so this means adding personal gear of the crew, adding camo, prepared tow cables etc

- See the interiour from the outside - Being able to view parts of the interour of a tank through open hatches etc like in the video here. This may be a neccessity for the Detachable Parts feature.

- Historical Diversity - Almost every Unit throughout the ware modified their vehicle in a different way to expect their wished like adding extra track links, spare wheels, stowage bins etc. So for specific battles (and participating units) there should be historical correct models.

These are my current thoughts which I now open up for discussion so give me your thoughts and add further ideas if you have Smile

Blackwatch

Posts : 27
Join date : 2010-04-16
Age : 40
Location : Munich / Germany

View user profile

Back to top Go down

Re: The Tank 2.0 - Evolution of Tank Modelling?

Post  brit44aldo on Wed Apr 06, 2011 6:14 am

pick a model and do your best. move on to the next model, because you are a limited resorce.

brit44aldo

Posts : 10
Join date : 2010-02-02

View user profile

Back to top Go down

View previous topic View next topic Back to top

- Similar topics

 
Permissions in this forum:
You cannot reply to topics in this forum