Hydra History #2 and an interesting article

On the path to the new engine, I faced the challange of the resource manager design and this turned in an interesting discussion on gamedev. This is the link:

Engine design and resources

Talking about Hydra, the first release was a great success for me. It was a graphic engine wrote from scratch (int main...) and during its lifetime it teached me much things.

This the second shot of my engine. At the time it was only in wireframe, and I was developing an importer from the 3ds format, due to my previous work with 3d studio max.








After that, I've began coding shaders but using asm shaders...this was actually a pain. The renderer has both Direct3D and OpenGL support, so I had to create both shaders two times.
To solve this problem, I've began coding in Cg and then converting to asm every time. This process was really slow and the shaders I've implemented were not optimized and much time was spent to fine-tune them (code, export in asm, modify gpu registry bindings...).
To support some sort of scene hierarchies, I've implemented Bounding Volume Hierarchy:

Those two models were under the same node, so the resulting volume must contain both the volume of the models.
This screenshot used OpenGL renderer.





Having a simple scene hierarchy system, here comes the time to develop something more interesting.
I have always love water, both in the reality and in 3d rendering (even using max), so the natural choise was to study how to achieve the water.
I've seen what's missing in my renderer, so I create functions for both OpenGL and DirectX to support culling planes and render to texture. I created a tube under max and added some noise to create a cave, exported and then I've tried to create the water.
This was my first attempt:

This image seems like coming from ten years ago, but it was a success at the time. The reflection was done via render to texture and culling plane, and the result for debug purpose can be seen on the lower left.
The refraction was poor, and there were no noise and no light.


After this preliminary water, it comes the time to face another subject I love: shadows. This was the REAL pain. I've passed nights and many posts on gamedev to achieve a simple shadowmap, due to the texture format, due to the poor pipeline: cg->asm->hydra. It takes long time to see the shadow map working, I don't know even how much!

Finally the long waited shadow map! Ugly but hey man...I've learnt many other things!



















And then I wanted to give another touch...using normal maps.
I've noticed a family of graphic effect that derive from normal maps like parallax occlusion mapping, relief mapping, cone step mapping and stuff like that.
So I've began studying them, creating shader for normal, simple parallax and relief with fake soft shadows using FxComposer. It ended up with the overused rockwall:

Another problem here was the lack of informations like Tangent and Bitangents.
So I create a function to fill the vertex buffer with those informations...using the math explained in this article: Tangent Space Calculation



The end of the story will be in part #3...stay tuned!


Demiurge

Hydra history #1

A years ago, I've began coding a graphic engine called Hydra. It was very important for me, because it learned me much things about different fields, like graphics, engines, directx/opengl, cg, hlsl, windows and linux.
After some months of nighttime development, I've created a demo and a video about it:

http://it.youtube.com/watch?v=Bx_WeMigxRA&feature=related

Even if it's not an impressive demo, it is based totally on the renderer made by me. Started from scratch. And this leaded me to continue to study and research about graphic and engines.

And leaded me to finally be a game programmer!

Long time...

It is a long time since my last post.
Many things have changed, and many were improved!
First of all...I've done it! I'm working in the game industry! In september I've began working @ Gd-Mind, a small game company located here in Rome. The peoples are great, Tiziano Lena (Lead Prog) and Gabriele Barlocci (Senior Prog) are quite great masters...and day after day they teach me many new things.
The Hydra Engine finished with a small demo, but during the entire process it teached me many stuff about engines and now I'm thinking of starting a new engine from scratch, the old architecture is quite too complex. The engine included NVidia PhysX for physics and lua for scripting, because of the will to expand to a game engine and not a mere graphic engine. But this lead me to think about a new and more simple architecture, and choose to support only a graphic API.
Stay tuned for more info about the next steps...

The power of design...

The step is done. Friday I've done the design of the new architecture, that will allow a multi-platform/multi-render development of my engine, all based on libraries. So I wrote down all the subsystems, the new engine tree, and the code came up from my fingers during friday night.
The result is an SDK, useful to create demos and applications focusing on the apps and not on the renderer or the OS.
I feel very satisfied about it, it is modular and let me concentrate on the next steps: developing an OpenGL renderer and a OpenGL/Linux application, to let the engine be developed with my two main OS.
I'm working also on an embedded version of the engine in WxWidgets, so I can create some useful tools to build shaders, scenes, and stuff like that.

Casper: today he will become Software Engineer. With the max possible rank. Congratulations!


P.S. I'm Demiurge. The second is Casper...and the next?

First milestone!

The time has come to close the first milestone of my little engine.
Some basic features, but still important (like a full math library, geometry and texture importer from 3Ds files,
render to texture and basic shader, basic camera and scene managment).
And now a new need.
The need to move to multiplatform/multirenderer architecture, the need to see it in every way possible.
I'm still confused about the solution to this problem, cause the windows development was fast and easy with visual studio 2005 and DirectX9, but more satisfaction is yet to come.

It's time to design - how to build over different platforms.
It's time to design - how to let my brother work on our dream.

Welcome Caspar!

Render to texture

Finally the times come.
The times to develop one of the most useful feature of an engine: render to texture!

Here is the link to my simple scene...rendered 2 times: one to create the texture and another to
create the real scene!
http://i110.imagethrust.com/images/
2YVi/view-image/hydrarendertotexture.html

I like it...it begins to take shape...

BREAKING NEWS!

It's time to grow up...I'm pleased to announce that Casper, my best friend and brother, joined me in this adventure.
Say Hi to Casper!

Bad Fool Prototype

Gentlemen, start you engines!

Be prepared...
This is only a trial...
To share some notion,
to learn new stuff.
To put answers and
find new questions.

A solid backup of

ideas and solutions.