When I make some progress, I will be sure to let you know. Thanks again.

What you could do is something like that :

TheoraVideoManager * mgr = NULL;

// bind all your functions here (see wiki)

void StartPlugin(void)
{
    MEngine * engine = MEngine::getInstance();
    MScriptContext* script = engine->getScriptContext();

    // create the video manager
    mgr = new TheoraVideoManager();

    // register the script functions
    script->addFunction("loadVideoClip", loadVideoClip);
    script->addFunction("updateVideoClip", updateVideoClip);
    script->addFunction("playClip", playClip);
    script->addFunction("pauseClip", pauseClip);
    script->addFunction("stopClip", stopClip);
    script->addFunction("destroyVideoClip", destroyVideoClip);
    //etc...
}

void EndPlugin(void)
{
    // delete the video manager
    SAFE_DELETE(mgr);
}

First, here's how you will send the video data to a texture in C++ :

MEngine * engine = MEngine::getInstance();
MSystemContext * system = engine->getSystemContext();
MRenderingContext * render = engine->getRenderingContext();
MLevel * level = engine->getLevel();

// get a texture reference that will receive the video stream
char globalFilename[256];
getGlobalFilename(globalFilename, system->getWorkingDirectory(), name); // "name" if a local path like "maps/dummy.png"
MTextureRef * texture = level->loadTexture(globalFilename, 0, 0);

// send the video clip data to the texture
TheoraVideoFrame * f = clip->getNextFrame();
if(f)
{
    unsigned int textureId = texture->getTextureId();
    if(textureId == 0)
    {
        render->createTexture(&textureId);
        texture->setTextureId(textureId);
    }

    render->bindTexture(textureId);
    render->setTextureFilterMode(M_TEX_FILTER_LINEAR, M_TEX_FILTER_LINEAR);
    render->setTextureUWrapMode(M_WRAP_CLAMP);
    render->setTextureVWrapMode(M_WRAP_CLAMP);
    render->texImage(0, clip->getWidth(), clip->getHeight(), M_UBYTE, M_RGB, f->getBuffer());
}

I was looking over the documentation, and I'm still trying to figure out how to map the OGG Theora video data as a texture onto a game object passed through Lua into the C++ interface. Mind giving me a point in the right direction, Anael?

Thanks.

Is there a way of being able to manually map the texture of a game object through the engine's Lua interface? Please let me know.

It has a BSD license. Will that be more permissive for commercial purposes, Anael?

EDIT:

More info here: http://libtheoraplayer.cateia.com/wiki/ … =Main_Page

Sample class here: http://libtheoraplayer.cateia.com/wiki/ … e=Tutorial

If it turns out that this is something I'd have to manually build using Theora/Vorbis directly, I'll probably put it on the shelf until I get used to the Maratis plugin interface.

EDIT:

Also, is it currently possible to call game object textures from Lua? My understanding is that that's the only way to get the data from the video onto a plane using a plugin interface. Please correct me if I'm wrong.

A Mac version of the library will have to be compiled, but I'm guessing we have some people here who can probably help out with that.

I could see this working as a plugin IF you're able to map the data streamed from the video as a texture while writing your game logic. Otherwise, I think it would be best if video-playback were a core engine feature, and you would access it through the visual editor.

zester wrote:

If you decided this is a task you want to take on, use OpenCV instead of LibTheora or FFMPEG. Because ...

1. OpenCV already uses those librarys for video playback
2. Getting OpenCV to play video on  texture is only a couple lines of code.
3. You get the added benefit of its builtin motion tracking, in a first person type game this feature could be used
to actualy look round a corner by slightly tilting and turning your head. Or controling a game in the sameway as Kinetic.

It would probable take 5min to get this feature implemented in the Maratis Engine and an hour for the Lua bindings. I could do it but everytime I even breath on those scon's build scripts they blow up in my face.

Thanks i'll consider using OpenCV specially if it's easier to implement. I don't know about scons or even lua that much i started learning it for maratis.

I have nothing to offer in the way of scons but for C/C++ and Lua this might help give you a better understand on how to tie C and Lua

http://code.google.com/p/zester/wiki/Lua_C

Scripting in Lua is very simular to Javascript except lighter and faster unless your comparing lua's speed and ease of use to Qt5 and the V8 Javascript Engine in that case Qt5/V8 win hands down.

If you decided this is a task you want to take on, use OpenCV instead of LibTheora or FFMPEG. Because ...

1. OpenCV already uses those librarys for video playback
2. Getting OpenCV to play video on  texture is only a couple lines of code.
3. You get the added benefit of its builtin motion tracking, in a first person type game this feature could be used
to actualy look round a corner by slightly tilting and turning your head. Or controling a game in the sameway as Kinetic.

It would probable take 5min to get this feature implemented in the Maratis Engine and an hour for the Lua bindings. I could do it but everytime I even breath on those scon's build scripts they blow up in my face.

Thanks i'll consider using OpenCV specially if it's easier to implement. I don't know about scons or even lua that much i started learning it for maratis.

1. OpenCV already uses those librarys for video playback
2. Getting OpenCV to play video on  texture is only a couple lines of code.
3. You get the added benefit of its builtin motion tracking, in a first person type game this feature could be used
to actualy look round a corner by slightly tilting and turning your head. Or controling a game in the sameway as Kinetic.

It would probable take 5min to get this feature implemented in the Maratis Engine and an hour for the Lua bindings. I could do it but everytime I even breath on those scon's build scripts they blow up in my face.

Hi,

the engine doesn't have a integrated feature for reading video.

But you can use a 3rdparty library, like libtheora (to read ogg theora video) : http://www.theora.org/
I did it for a project some years ago, unfortunately I don't have the code anymore.

You read the stream and send the result to a texture using MRenderingContext

Or using FFMpeg maybe, but I never tried : http://ffmpeg.org/

Well, thanks for the fast reply. And that was really helpfull thanks.