摘要 |
According to the invention, acoustic signals (snd) are produced from a video data flow containing a temporal sequence of individual images (img). A displacement vector field (vfd) is determined from the image data of the individual images (img), for example by means of an MPEG algorithm; at least one characteristic variable (chs) is then derived from the displacement vector field (vfd), e.g. a dominant displacement vector, in an analysis module (AAN); and an acoustic signal (snd) is produced on the basis of said variables (chs) by means of a synthesis module (SYN). |