Last news

01/10/11 - Games i worked on


08/09/11 - First version of dynamic voxel space streaming in H4Engine :

A new plugins module to manage collision, update, generation of large voxel streammed space (2^20*2^20*2^20).

Up to 256 materials can be used to paint voxel space.

















02/2011 : Experiment PID Regulator for physic simulation with H4Engine

A proportional–integral–derivative controller (PID controller) is a generic control loop feedback mechanism (controller) widely used in industrial control systems – a PID is the most commonly used feedback controller. A PID controller calculates an "error" value as the difference between a measured process variable and a desired setpoint. The controller attempts to minimize the error by adjusting the process control inputs.

So what... ? what can a do with these PIDs ?

here is the full code of simulated quadrirotor flying drone  coded with H4engine and using some pid controller :

cdrone.cpp cdrone.h

This machine is flying with 4 rotors and some gyroscope, accelerometers, altitude sensor  etc.. . The main goal is to have a good stable flying machine, just  by controlling the trust of  4  motors with PID regulators... not so simple in real world.

Here, the coding is pretty simple for a very good result, cause :

    - it just use a physic rigid body (box) with forces applied ( controlled via pid regulators. ) on each upper corner to simulate the thrust of each rotor.

    - there are less parameters than in real world (no wind, linear response of electric motors, precision of sensor, etc...)

    - if use a virtual hidden force to simulate combined torque of the 4 motor to control rotation on y axis. On the real machine it's a bit more complex to control this.


Implementation of SSAO in H4ENGINE

Programmed a multicore hierachical profiler for h4engine



H4' Engine : Dual Quaternion Skinning and 3DSMax update


  • Ladislav Kavan (many thanks to him !) described dual quaternion skinning method to replace our traditionnal blended matrix skinning -> see PDF .

    This tech was used in Crysis Engine.

    I updated Maxcomponent to support this method in "3DSMAX9 32 bits" and "3DSMAX2009 32 bits" and "3DSMAX2009 64 bits".

          1. download the DLL "maxcomponents.dll" for your version of 3DSMAX :
            • 3DSMAX9 32 bits version : HERE 3DSMAX2009 32 bits version : HERE
            • 3DSMAX2009 64 bits version : If you are interested in 3DSMAX2009 64bits version, contact me at "h4engine@free dot fr" (replace the dot!) . I will release a version soon
          2. rename the current maxcomponents.DLL in the 3DSMAX.EXE's directory
          3. copy the new dll.... run and see dualquaternion skinning in action !!!!

H4 Engine use this method to render skinned meshes. This tech is very easy to implement, and the additionnal GPU cost is negligeable.


classic matrix skinning --vs-- Dual Quaternion Skinning



    This technique produce best skinning result than traditionnal linear matrix blending. It preserve shapes volume. An other good point is that you only need 8 float per bone comparing to the 12 needed for classical skinning (gpu registers are not unlimited !!! it can help ;-) , so it can avoid splitting the mesh if there are too many bones...)


THE DEEP : A prototype of third person action game  2009. Developped @ Darkworks Studio




I AM ALIVE : the last game i worked on. This game will be released for 2009.

Take a look on the trailer. This game is created and realized by Darkworks Studio and will be released for XBOX360 and PS3.




PHOTOGRAPHY : a MUST SEE new site from Severine.Levastre.free.fr. Indeed a very talented photograph...


.COLDFEAR : a Zombie game i worked on as Lead Programmer. Released for PS2 and XBOX 1. A Darkworks Studio creation, Published by UbiSoft.



LA TOILE DU DIABLE : a Demo coded when i was working at Delphine Software.

Here are some screenshots & videos:



ANTS : This little demo is inspired by the famous tutorial of Mat Buckland about neural networks and genetic algorithms. I programmed my personal evolution of his 'neural ants' sample.

These ants are just learning to go toward something on the map. For the moment, the switching between position of nearest food and nearest base is done
by manual programmation. I tried a version of NN that 'decide' to search food, and then to carry it to green base but convergence toward an efficient behaviour was really too slow and hazardous.

Many greatings to Mat Buckland

Download, launch and WAIT.



Delphine Software : i have been working several years for Delphine Software International.