>Psychtoolbox>PsychDemos

BlurredMipmapDemo - Show blurring of live video or movies via adaptive mipmapping.

This demo shows how to selectively blur regions of a live video stream
captured from an attached camera, or a live playing movie. The method
employed here uses a GLSL shader to select an individual blur level for
each output pixel. Pixels close to the mouse cursor (which simulates gaze
position, e.g., as aquired by an eyetracker) will be drawn sharp, whereas
pixels at larger radial distance will be drawn increasingly unsharp by
low-pass filtering them.

The method of blurring here is computationally efficient and fast,
because it generates and uses a image resolution pyramid to precompute
different versions of the video image at different resolutions. Then the
shader just looks up color information for each individual output pixel
from the proper “level of detail” in the pyramid of low pass filtered
images. The resolution pyramid is auto-generated in a fast, GPU
accelerated way for each image by use of OpenGL “mip-mapping”. This
method uses a fixed low-pass filter for downsampling, usually a box
filter, but the exact type of filter used is system dependent and can
vary across different graphics cards, graphics driver versions or
operating systems.

The demo requires recent graphics hardware and won’t work on old graphics
cards.

Usage:

BlurredMipmapDemo([deviceIndexOrMoviename=0][, gazeRadius=25][, customMipmap=0])

‘gazeRadius’ Some falloff radius in pixels. Controls how quickly the
resolution degrades with increasing distance from the cursorposition.
Smaller values give a smaller simulated “foveal area”.

‘deviceIndexOrMoviename’ = Optional: Either the index of the video
capture device, or the full path to a movie file. Defaults to
auto-selected default video source.

‘customMipmap’ Optional: If non-zero, use a custom shader for
downsampling during mipmap building, instead of the relatively simple
builtin filter of the GPU.

Press any key to finish the demo.