In the case of neuroscience experiments, trough psychopy for example, graphics are rendered "on line", so there is no data to be loaded as a texture.
Additionally the down-sampling process is realised by pseudo-retina cells, which are implemented on the software side and should be independent of the video stream source. Thanks, Nicolas 2010/11/15 Greg Ewing <greg.ew...@canterbury.ac.nz>: > Nicolas wrote: > >> I understood what you explain to me about GPU and CPU sides, but I do >> not have the choice about the input type and it is not possible to >> move my processes onto the GPU side. > > Have you considered just moving the downsampling step onto > the GPU? You could do it by loading the image as a texture > and then rendering a small quad from it using GL_NEAREST > sampling. > > -- > Greg >
