Hello everyone. Don't know if it's the right place to ask but here's my problem.

I have a skecth in wich I create a graphic buffer to be rendered on demand with P5Sunflow (thanks to antiplastik)
That means that when i am satisfied with my 3d lsystem growth and angle of observation i can render the frame.
Here's the code, it is executed when the 's' key is pressed.
The sketch uses the peasycam for cam controls but despite my efforts i cannot spot how to retrieve the cam angles on the 3 axis.
cam.getRotations() and its value stored in a float array don't work as i expect.
Could you help me?
Thanks in advance.
Here's the code for a on demand render of 1 frame.

Code:

     case 's': 
	noLoop();
	recorder = createGraphics(width, height, "hipstersinc.P5Sunflow");
	recorder.beginDraw();
//	scale(40.0);
//	rotateX(???);
//	rotateY(???);
	draw();
	recorder.endDraw();
	recorder.save((int) random(10000)+".png");
	recorder = null;
	loop();
     break; 
 

Quark wrote on Apr 20^th, 2010, 3:57am:



Hi, did a quick test yesterday.
Don't know, but when i rotateX, Y, or Z using these values, rotations are not correct:

float[] rotations = cam.getRotations();
rotateX(rotations[0]);
rotateY(rotations[1]);
rotateZ(rotations[2]);

I was trying this other solution:

float[] camPos = cam.getPosition();

PVector c = new PVector(camPos[0]-0,camPos[1]-0,camPos[2]-0);
float r = sqrt(pow(c.x,2)+pow(c.y,2)+pow(c.z,2));
theta = atan2(c.y,c.x);
phi = acos(c.z/r);


then, on the the buffer:
recorder.beginDraw();
translate(width/2, height/2, 0);
rotateZ(theta);
rotateY(phi);
rotateX(-HALF_PI);
scale(16.0);
draw();
recorder.endDraw();

It's still not working as I expected....  


GC

Rereading your post, you could probably just capture the camera matrix and apply it to the buffer.

PGraphics3D g3 = (PGraphics3D)g;

recorder.setMatrix(g3.camera);

I believe anything that is set as the lookAt is positioned at 0,0,0.  I guess the getRotations and getDistance may tell you how it is looking at it.

Quote:

Hi Quark, and thank you for the reply.
That thing is driving me crazy. You are right, recorder.stuff may be better but is not required to affect the transformation.. at least for this situation.
Anyhow, that not solves the problem.

Filling the variables phi and theta with the value extracted while drawing/rotating does not work. When rendered the angles does not macth.
Here is the result. The shape has been rotated (or better, the cam) to the left on the y axis and to the bottom on the x axis.



As you can see, the orientation is completely different.
It is a shame, because i would like to render big images generated with an l-system but i would like to do it by first rotating it to see wich part is more interesting.