Well it's pretty certain we'll license torque. Its capabilities look nice, as well as the actual economics of it.
But I still like to look at screens and information to help convince myself its the right idea.

I also know Oelund is using torque at the moment, hopefully he can provide some info after a while of use.

Here are some nice torque screens i've found from a game called "afterworld"

While they're not like, drop-dead gorgeous, they look to me like, with some nicer attention to textures and lighting, and a little postprocessing, this kind of engine would be able to do everything we need it to.



Here's a shot i found taken from the torque model viewer... It's hard to say bt it looks like advanced materials are going on here..


And this shot in particular was a massive relief for me... Most the shots i've looked at seem to have stupidly low resolution terrain textures... This dispelled the idea it had put in my head

looks like a boy

Age of Decadence is being made with Torque.

I've recently started on a project on the Torque engine.
We are still in early stages, so this will be more like a first impression report.

The material and shader effects appear to be quite nice, and fairly easy to edit.
The world editor can be a bit anoying from time to time, in that you have to start a game, open a map, place your newly made model into this map, and then open the material editor.
But the plus side to this is that you edit the material while seeing exactly what the model looks like in the game with the different lights in the world.

The world editor strikes me as very annoying.. I havn't played around with it too much yet, but it appears to be very hard to place and rotate stuff accuratly.

Making models in 3D max is also a bit complicated. You need to create A LOT of dummy objects of various names in order for the exporter to pick up the objects you want to export.
All objects in the scene needs to be linked in a speciar hierarchy in order to work.
This is just a matter of getting used to that system.

There are a few issues with the engine that makes me very skeptical about it. I don't know if any of these can be done somehow, but as it is now, it doesn't look like it:

1) There is no bone merging.
It appears to be impossible to for example, have a playermodel with multiple head models that are interchangeable if you want the neck to be deformable.
You can attach a seperate head model to an attachment node in the head, but making it deform between the neck and head appears to be impossible.

2) The only form of animation blending is additive blending.
This means no 9 way blending (which I am a big fan of working with)
Sure, additive animations can make you use much fewer animations than 9way blending, but the quality of the animations are pretty much arbitrary.
You'll have absolutely no control over how a reload animation will look when the player is looking upward. You'd most likely see him grap the magazine nowhere near where the magazine actually is, then move the magazine out infront of himself and reinsert it, all while poking the weapon through his head.

I'm trying to make a custom rig to combat the majority of blend errors that happens when using additive blending (basically I move the problem from the hands to the shoulders, where it'll be less noticeable). All this would be A LOT easier if the engine suppored bone merging.

3) it appears that most of the game is done by scripting.. which pretty much means its very easy to put your own gameplay on top of their game. but at the same time its a lot harder to make your own stuff from scratch.

4) It appears there are very weird limitations to the objects you place in the engine.
I was making a flag model, and thought it'd be a nice feature if I did multiple wave animations for different wind velocities.
I suggested to the coder that we made it so that the artist could select which animation it should play when he placed the flagobject in the world (and thus take in account that the flag might be in shelter from the wind). But this was apparently impossible.
I was told that he could not add a simple variable to the model, that gave the artist the ability to select what animation the object should play. It would however be possible to hardcode the flag object to set its direction and animation to the windsetting on the specifig map.

5) not so much a problem, but rather a funny observation.
All the documentation of the game mentions that the game engine is set in meters. and it is important to remember to set the 3D Max units to meters, and keep in mind that everything is in meters.. yet everything in the game is set to players being 2.4 units high.
I was watching this tutorial where he pretty much said "remember 1 unit in max is one meter in torque.. thats why it is important to scale your player to 2.4 units"
It is of course no problem to scale your own content to what a meter ACTUALLY is, but I do find it funny that they claim everything is in meters and then scale everything after a completely arbitrary unit of messurement.

I'll let you know if I find some fixes for some of these issues.
The terrain part of the engine is not bad (although not super impressive).
It also supports deformable terrain.. so you could probably dig trenches in real time in the game (I imagine doing it by lowering the terrain and placing a model of a trench in the gab)
And have artillery craters.

i've been getting really mixed reports about how torque preforms on different rigs. Some people claim it runs at like 10fps and stuff until you turn all settings to low. There are multiple claims of this and I wouldn't want any games we made to be hit-and-miss performance wise.

about deformable terrain... UDK claims to be able to do that too, though I can't speak for how well it does it.


with UDK we'd only have access to the scripting engine. While it's a little lame, i've seen some pretty wild stuff done in UDK.

I really had my heart set on torque, but at $1000 a piece for full source, and $100 just for the scripting, I really don't feel like spending money on it before i know if it works.

At least with UDK we have the option of developing it without spending a cent, then once we're ready, buying our developer seat.
A massive concern to me is spending $1000 on torque and then hitting some obstacle which kills the project. $1000 down the drain.
While the $100 scripting version seems like a solution to this, it puts us at about the same level as UDK... which would be free to develop on.


The downside to UDK is they take 25% after you make $5000 annually from your game.
This seems steep but honestly, I think it's not a huge deal. We could sell it for $15 instead of $10 if it really bothers people. And besides, the game actually working just how we want it, for me, is infinitely more important than that 25%

A bit about the shaders in the Torque engine:
The shader can take a series of different textures.
Major:
Diffuse
Normalmap
Specularmap

Minor:
Overlaymap
DetailMap
LightMap
ToneMap

The Diffuse is pretty straight forward, although it is quite strange to work with transparent textures.
If you want a semi-transparent texture you have to save your texture as a .png without background.
Simply having the transparency as the alpha chan does not work (or at least I could not get it to work last time I tried.. could be another thing that caused it though)
You can however use the alpha chan if you need the texture to have an alpha mask. (where its either 100% visible or 100% invisible)
There are some different blend modes for transparency, including additive and multiplying. which is quite nice.
The material can be tagged for emmisive (basically full-bright) or glow (putting a pretty nice glow around it, if you are in to that sortof thing (looks very sci-fi))

Normalmaps works quite well, and can be set to parallax, using the alpha chan of the normalmap for the heightmap.
The parallax effect looks quite nice, but I think its the old fashioned basic parallax effect, and not the more accurate steep parallax effect.

The specularmap can be in full RGB (as far as I know)
However there is only a single scale for the specular(/phong). So you don't have control over both the exponent and power.
I'm not entirely sure what the scale is meant to represent.. Its sortof like an inversed exponent, where 0 is tight and 255 is wide.
You can set a colour effect for the specular.

Shadows are alright.. a brit griddy now and then, but I've seen worse.
I've seen a few problems with the shadows being cast on the collision mesh of objects rather than the actual objects. but I don't know if that is an error on our part.

So far I've got mixed emotions about this engine.
It is especially the animation system as well as the lack of bone merging that turns me off.
Visually, I don't see any problems with the engine.
Lighting looks alright, and works well with the shadows.

i didn't know parallax already worked - that's a good thing.
your post addressed pretty much all the specific things i wanted to know about the materials, very awesome job at predicting what i wanted to hear about

now. We don't have SSS, which will make skin look not as good as i would like, and we dont have fresnel/BRDF, which will make cloth look not as good as i would like.
SSS i honestly wouldnt expect of an engine by default, but almost every modern engine supports fresnel :/

When you say lightmap and tonemap, are these related to the terrain? lightmap sounds like just a regular lighting/shadow for environments, and tonemap sounds like it could be a "megatexture" colouration for terrain (like in arma/VBS)
but i'm kind of crossing my fingers and hoping that "tonemap" is a tonal ramp and with that we could fake fresnel and SSS and generally play around with the lighting model.
If we can't do at least 1-dimensional tone ramps, i'd be pretty disappointed.

btw
http://www.laurenscorijn.com/udk-comparision.html
this is the post which got me interested in udk again.
The update seems to have a totally revised lighting model, which was the one thing i hated about unreal engine graphics

ok, hernan has been fucking around with UDK tonight. I have no idea what specifically he's been doing but he sounded fairly impressed