2009-Oct : Characters Revised

[smith1215]

Damn son!!! really nice work.

[Niklas Elling]

Thanks for the comments.

Pwn: They were done in maya, zbrush and photoshop.

[Sgt So and So]

Hot DAMN.

[Baracca Chewbama]

Originally Posted by Fozzy The Bear

Rendered in Marmoset Engine

Rendered in Marmoset Engine? Does this mean 1944 D-Day Operation Overlord has changed engines again?

[Ekoms]

Originally Posted by Baracca Chewbama

Rendered in Marmoset Engine? Does this mean 1944 D-Day Operation Overlord has changed engines again?

Doubt it. Marmoset, specifically the toolbag, is just a nice and quick way to view 3d models in a real time engine. Takes all of 30 seconds to import and get the textures applied, plus you can adjust lighting on the fly. A very cool tool.

Marmoset's not nearly as advanced an engine as this game would require anyway.

[DavidUpton]

My only critique is that it all looks too good. The uniforms look like they are fresh out of the factories - clean, uncreased, undamaged. I think it would be good to see some dirt.

[napalmi]

Well he is still in development so he hasn't yet participated in any battle

But yes, I concur. But I think damage and dirt should add up as the time passes. Would be cool when the new recruits with their clean and tidy uniforms would look ragged and filthy after a few firefights.

[Katana]

Like the new improvement into the new Axis model. *Thumbs Up*

[btraill]

Jesus... For indy developers, this stuff is top notch. Better then what I've seen in most recent name studio games.

[Garret Rowe]

Yeah, hopefully in game we won't lag. That would just be bad. But really nice pictures, looks so realistic!

[Brasse]

Woah, that German model blew my mind.

[b0x]

Originally Posted by Fozzy The Bear

oh wow

[DavidUpton]

Two things that I have only just noticed:

1. Where is the insignia on his cap?
2. His y-straps are wrong.

[Pwn]

absolutley fantastic! Will the polygon count and high-resolution textures portrayed in these screenshots be available during real-time rendering in the game?

[JonReb]

Wow looks like things have been coming along beautifully. I think I'll need to buy a $10k machine for this beast.

[napalmi]

Originally Posted by JonReb

I think I'll need to buy a $10k machine for this beast.

Don't worry, you won't need an expensive computer to play this.

[ncthorn]

Originally Posted by napalmi

Don't worry, you won't need an expensive computer to play this.

Although I understand that the game is targeted to run on a wide range of machines, I do hope that there are still plans to take advantage of some of the features found in newer/high end rigs. Im really hoping for some impressive effects such as smoke and explosions, as well as pretty visuals as I feel that these things do help with the immersion of the game. This model is certainly headed in the right direction and if it is relatively low poly to boot, thats all the better.

[Bunglo]

We're planning ahead with all of our art to help ensure the game doesn't look like garbage in the next years to come. All the while trying to make sure our system requirements are as low as possible by optimizing and using texture sizes that get the job done but still look good.

We can do this by creating textures that are 4-8 times larger than the intended medium/high range texture dimensions.

A brick texture that's intended to be used on buildings or anything else it can be applied to is created at 2048 and then automatically scaled down to 512 upon exportation to ensure realistic resource usage with out suffering on quality all that much.

If your rig can handle it, you'll be able to upscale to high texture resolutions (1024, 2048, etc) as the texture has been created at the 2048 limit.

The same is being done for the models by creating the maximum amount of LODs until it no longer makes a difference if you have 20 tris or 300 tris.

At the moment it's looking like the average building will use 3-4 LODs along with occulder models that will cull anything the player isn't able to see. Each LOD appears in relation to the players position to the model.

For example, if a player is 50m away from a building with 3 LODs, the player will see LOD1 rather than LOD0 which is the model in it's entirety. Stand 100m away from a building, LOD1 is replaced with LOD2. Stand 150m and LOD2 is replaced with LOD3 which is the lowest tri count the model will have as going any lower will make no difference performance wise.

As said above, depending on how many tris a given model has, it could have 1 LOD or it could have 5, we'll make as many LODs as necessary until the tri-count is no longer a factor.

[Guardian]

So if you look at a building for 500m with binoculars, will it show it as the higher LOD or the lower one?

[Bunglo]

It will show what ever LOD is programmed to show at the distance the binoculars are zoomed into.

If you're 500m away, staring at a building with out equipment to help you zoom in on an object (so just using your "eyes"), you're probably going to see the lowest LOD model.

If you use a scope or binoculars to zoom in on an object, what ever that piece of hardware zooms you to, let's say 200m, you'll see the appropriate LOD model that displays at 200m.

Basically it will be as if you were standing 200m from the object, rather than the 500m.

This won't be a hard set rule though, just a generalization of how optimization along with identification will be done on things such as buildings.

Actual player models will be handled in a similar manner but we need to ensure that those zooming in are able to distinguish if the player in question is carrying a rifle, rocket launcher, mg, etc.

That really shouldn't be a problem as long as the LOD models have good silhouettes, allowing for "easy" identification.

One game that didn't do very well on this was BF2. It's an aging game but zoom in on a player from so many meters away and you'll see a player model without his helmet and more importantly, the weapon he's carrying.

Character quality largely depends on textures (as most art does), larger texture = sharper, clearer materials = prettier graphics.

As long as we keep creating textures at dimensions that are large enough for upscaling but not so large they look like crap on the downscale we shouldn't run into any problems for those who have higher-end machines.

And while tri-count is an issue, textures take up a lot more resources than polys.

[Guardian]

Thanks for the answer, and that is what I was hoping it would be, the games I play now keeps things looking low detail if you are far away and doesn't recognize scopes.

[ncthorn]

Thanks for the response bunglo. Sounds like you guys have a solid plan. One more thing though. How are you planning to handle the way grass draws at a distance? As you probably know, ARMA II draws the actual blades of grass only over a very short distance and then partially obscures items over a distance to make it look as though they are buried down in the grass. Personally I think it looks somewhat goofy, especially with the grass only drawing over a very short distance around the player. Are there plans to make low poly LOD type blades of grass to be drawn at long distances? I think it would make for an absolutely beautiful game world to have lush grass drawn at long distances.

[Bunglo]

As someone who finds it extremely important for foliage to look as close to the real thing as possible, this is something I'm going to work hard to get right.

I, along with probably every other gamer who cares, want to see large, open fields of tall grass that isn't growing in a second as I approach it. I also don't want to see it instantly disappear as I walk away.

At the moment I can't really comment on how this will be handled but when the time comes I can assure you this is something I find important for any game. Hopefully we'll be able to come up with a decent solution.

[Pwn]

Originally Posted by Bunglo

As someone who finds it extremely important for foliage to look as close to the real thing as possible, this is something I'm going to work hard to get right.

I, along with probably every other gamer who cares, want to see large, open fields of tall grass that isn't growing in a second as I approach it. I also don't want to see it instantly disappear as I walk away.

At the moment I can't really comment on how this will be handled but when the time comes I can assure you this is something I find important for any game. Hopefully we'll be able to come up with a decent solution.

A good distant LOD terrain texture that blends well with the foilage could help to make the "appearance" and "disappearance" of the foilage almost invisible as the player navigates. However, this is a disadvantage to distant players hiding within the foilage and the strategy of the game becomes unrealistic.

[DurrutiColumn]

To be honest the German has the 1941/Hollywood look with jackboots. By 1944 laced ankle-boots were the norm for front line troops along with plenty of camo.

[Lucan]

Wow, this looks awesome.

I like the expression on the German's face too.
DAT ASS

[DavidUpton]

Originally Posted by DurrutiColumn

To be honest the German has the 1941/Hollywood look with jackboots. By 1944 laced ankle-boots were the norm for front line troops along with plenty of camo.

The part about the boots is correct - usually jackboots were mainly worn by NCO's and officers. However, for standard infantry divisions, camo was not used extensively. A soldier would usually not receive more than one or two pieces of camouflaged items.

[Fozzy The Bear]

Originally Posted by DavidUpton

The part about the boots is correct - usually jackboots were mainly worn by NCO's and officers. However, for standard infantry divisions, camo was not used extensively. A soldier would usually not receive more than one or two pieces of camouflaged items.

Depending if you were in an elite Heer unit (Splinter Pattern) or Waffen SS (dot 44 peas).

[RonanHayes]

All Characters are constructed as follows:

1. Head
2. Hands
3. Jackets
4. Trousers
5. Boots
6. Gear (Straps, Belts, Hats, Helmets, Bags etc.)

[MajorMike]

Amazing job guys! I've been keeping tab on this project since April of 2007 and even though there aren't as many screenshots up and about the forums, it's still good to see Ronan answering all the possible questions that are still out there (Including ones that have been asked time and time again).
I just wanted to say I admire you and your teams persistence, Ronan, and we all can't wait for the next "Work In Progress" to oogle at for a number of months!

Looking forward to some fantastic models in the future!