UT99.org | Unreal Tournament Community

Shadow wrote:That heat/distortion effect is possible, yes. But first I have to implement general DX/OpenGL shader support.

Nice to hear this.
You making crazy, hard and brain-burning job.

Torax wrote:

papercoffee wrote: I want to learn Z-Brush

I tried it some time ago. Rarely interesting program.
And it's not difficult to learn basics - controls are pretty easy, i spent few hours to understand it.
Here are two my tryings to make something interesting))) the head of the man was first work, the buddy with the bird head (reminds me about one of gods from Egyptian mythology) - second one.
I know that bird-head-guy looks like somebody broke his neck and rotated head around 180 degrees
But i'm not good in human anathomy cozz never was interested in it and really bad in the art of drawing

Well ...I had once Sculptris from the Z-Brush devolper.
I tried a head ...it looks rather Asian to me hehe.

Spoiler

Manga Head

Heh, if there's anything shader-related, I can help with the shader code itself.

Torax wrote:

Shadow wrote:That heat/distortion effect is possible, yes. But first I have to implement general DX/OpenGL shader support.

Nice to hear this.
You making crazy, hard and brain-burning job.

Yes. Working to get money. Studying hard shit (biology). Having a life AND living out passion for this game brings me close to my own limits.

@Saya, nice to know. Any experience with the DX (hlsl) and/or OpenGL shading languages?

I've done quite a lot of work on HLSL and GLSL in my spare time, all sorts of silly things. Lighting, color correction, bloom, depth of field, ambient occlusion, motion blur, antialiasing, film grain, cel shading... you name it. It's a shame the only engines where I'd love to put them don't have custom shading stacks.

Saya-chan wrote:I've done quite a lot of work on HLSL and GLSL in my spare time, all sorts of silly things. Lighting, color correction, bloom, depth of field, ambient occlusion, motion blur, antialiasing, film grain, cel shading... you name it. It's a shame the only engines where I'd love to put them don't have custom shading stacks.

Sounds cool.
How you made that? I mean are there any programs (graphical workplaces), where you can write code and then lauch it and watch what happening?
I was interested in about to learn GLSL a bit, but actually had no ideas where to start from..

Well, for HLSL I mostly worked with ENBseries (it's horrible, let me tell you) and eFX (which is now dead).
For GLSL I just wrote my own small OpenGL program, which is very easy when you have things like GLEW and GLFW.

This may interest a few people
https://code.google.com/p/qeffects-gl/

Dr.Flay wrote:This may interest a few people
https://code.google.com/p/qeffects-gl/

It supports Unreal? Hmm...should try it

*EDIT*
Fun thing) Graphics looks interesting, reminds a bit Borderlands at default settings, but at slow PC lags as hell
Thanks for found Dr.Flay

The graphics comparisons don't look that interesting to me. All in all it's just a collection of post processing effects. The only interesting thing to me is SSAO Support (who fucking needs fake emboss bump mapping?).

oh, qeffects-gl... I haven't heard good things from other developers about it. And you won't hear anything good from me either. For a decent shader injector I'd rather talk about gedosato (which I hope becomes api-agnostic soon).

Unreal Engine 1 models severely decimate vertex data. This makes it impossible to retain detail in high polygon meshes.
I mean check out the curves on this flack cannon...



Is there any fix to this problem?

Myth wrote: Is there any fix to this problem?

Remodeling...

papercoffee wrote: Remodeling...

Don't be so critical. It is much easier

Myth wrote: Is there any fix to this problem?

There are few possible ways fixing this problem. The good and the best.
First way is to rescale model. I'd found that decimation appears if only model too small itself - the smaller it is, the more distortion of the smooth shapes appears. Possibly, you can rescale model before importing, make it larger and this bug will be at least less visible. But you must know that overall size of model must fit 256x256x256 size cube, what means that you have free space in range from -128 to 128 points in each direction. And it must fit those borders during animation too. Otherwise, vertex positions will be flipped, i.e. - if some vertex during animation moved to some position that equals 0,0,129 (meaning X, Y and Z directions) from start of coordinates, it will be moved to position that is equal to 0,0,-128 or something like this, causing model to totally skew.

The second way is a bit harder to realise but it is the most effective because it fully saves model data about vertex positions nad other stuff i'd never thought about.
I mean porting model as skeletal mesh, via PSK - model data, and PSA - animation data.
One interesting thing also that you cant import model if there's no psa file related to it. It can be even simple "Idle" animation with one frame but it MUST exist.

Well I hope that's enough. If your model has a length of 1 unit across its major axis you still only have 1/256 resolution. Having 2 vertexes less than 1/256 away would cause them to become one.

It would be a shame cause the models are looking great.

I'm worried about these awesome details: