MinimalisticOpenGLDemo

>Psychtoolbox>PsychDemos>OpenGL4MatlabDemos

MinimalisticOpenGLDemo([multiSample][, imagingPipeline][, checkerBoardTexture][, doAccumulate=0])

This demo demonstrates use of OpenGL commands in a Matlab script to
perform some very boring 3D rendering in Psychtoolbox.

It shows a single static ball, lit with default lighting and exactly one
light source. This is meant to demonstrate the minimum amount of code to
draw anything visible with perspective projection. It also draws a static
teapot and some little box with a cone as roof.

Then it waits for a keyboard press.

After that it demonstrates how to do basic texture mapping and animation:
It loads a JPEG image of the earths surface from the filesystem, using
Matlabs imread() function, then converts the image into a Psychtoolbox
texture using Screen(‘MakeTexture’), then provides this texture as a
standard OpenGL compatible texture using Screen(‘GetOpenGLTexture’).
This standard texture is applied to a sphere using standard OpenGL commands
and finally the sphere is drawn as a rotating object in a simple animation
loop. –> You’ll see a rotating earth.

Stop the demo by pressing any key and it will finish.

The optional parameter ‘multiSample’ allows to enable anti-aliased
drawing with ‘multiSample’ samples per pixel on hardware that supports
this.

The optional parameter ‘imagingPipeline’ allows (if set to non-zero
value) to enable the PTB image processing pipeline, just to test that as
well.

The optional parameter ‘checkerBoardTexture’ allows (if set to non-zero
value) to apply a checkerboard texture to the spinning sphere, instead of
a “earth surface texture image”. This demonstrates algorithmic texture
generation and the use of trilinear mipmap filtering to improve image
quality for high frequency edges and such…

The optional parameter ‘doAccumulate’ allows to demonstrate an additional
motion blur effect by use of the accumulation buffer. If you set the
imagingPipeline flag to zero and doAccumulate to 1, then use of the -
nowadays deprecated and extremely slow - accumulation buffer is
demonstrated. If you set imagingPipeline to 1 and doAccumulate to 2 then
a new fast technique is demonstrated. Both achieve the same visual effect
with very similar code, but the latter technique is well supported on
recent hardware, much more flexible and much faster.

Notable implementation details regarding use of OpenGL:

The call InitializeMatlabOpenGL at the top of the script initializes the
Matlab-OpenGL toolbox and enables the 3D gfx support in Psychtoolbox to
allow proper interfacing between the OpenGL toolbox and Psychtoolbox.

After this call, all OpenGL functions are made available to Matlab with
the same - or a very similar - calling syntax as in the C programming
language. OpenGL constants are made available in a format that is optimized
for Matlab, where the first underscore is replaced by a dot, e.g.,
GL.DEPTH_TEST, instead of the C-style GL_DEPTH_TEST.

In order to execute OpenGL 3D drawing commands to draw 3D stims into a
Psychtoolbox Onscreen- or offscreen window, one needs to call
Screen(‘BeginOpenGL’, windowPtr). After OpenGL drawing and before
execution of standard Screen() commands, one needs to call
Screen(‘EndOpenGL’, windowPtr) to tell Psychtoolbox that 3D drawing is
finished.

Some OpenGL functions that return complex parameters to Matlab are not
yet implemented - this is work in progress. The performance will be also
lower than when coding in a compiled language like C++ or C – that’s the
Matlab tax you’ll have to pay ;-)

The toolbox checks after execution of each single OpenGL command if it
caused some error. It aborts your script with an error message, if so. If
you are happy with your code and want to disable these error checks in
order to squeeze out a bit more speed, you can call
InitializeMatlabOpenGL(0,0) instead of InitializeMatlabOpenGL at the top
of your script. This will disable automatic error-checking. You can then
use the commands gluErrorString or glGetError to perform manual error-checks
in your code if you want.

Apart from that, use of OpenGL for Matlab is the same as OpenGL for the C
programming language. If you are used to OpenGL coding in C, it should be
a zero effort transition to code in Matlab+PTB. If you don’t know OpenGL
then get yourself one of the many good books or visit one of the many
OpenGL tutorials on the internet.

The OpenGL Red Book is a great introduction and reference for OpenGL
programming. Release 1.0 is available online, later releases can be
purchased in any good book store:

http://www.opengl.org/documentation/red_book_1.0/

For more infos, code samples, tutorials, online documentation, go to:

http://www.opengl.org

The OpenGL for Matlab toolbox was developed and contributed under
GPL license by Prof. Richard F. Murray, University of York, Canada and
ported to M$-Windows and GNU/Linux and tuned by Mario Kleiner.

The earth surface JPEG-image is taken from the Linux/KDE application
kdeworldclock. kdeworldclock and its components are licensed under
GPL.

Path   Retrieve current version from GitHub | View changelog
Psychtoolbox/PsychDemos/OpenGL4MatlabDemos/MinimalisticOpenGLDemo.m