Experimental evidence from neurophysiological recordings in the middle temporal (MT) area of the macaque monkey suggests that motion-selective cells can use disparity information to separate motion signals that originate from different depths. This finding of a cross-talk between different visual channels has implications for the understanding of the processing of motion in the primate visual system and especially for behavioral tasks requiring the determination of global motion. In this paper, the consequences for the analysis of optic flow fields are explored. A network model is presented that effectively uses the disparity sensitivity of MT-like neurons for the reduction of noise in optic flow fields. Simulations reproduce the recent psychophysical finding that the robustness of the human optic flow processing system is improved by stereoscopic depth information, but that the use of this information depends on the structure of the visual environment.