| 摘要 |
<p>Encoding circuit for transforming a picture signal into blocks of, for example, 8*8 coefficients, in which each block of coefficients is read motion-adaptively. In the case of motion within a sub-picture, the block of coefficients is read in such an order that the obtained series of coefficients comprises, as it were, two interleaved sub-series. The first series starts with a dc component. In a first embodiment the second series starts with the most relevant motion coefficient. In a second embodiment two interlaced sub-fields are separately transformed and the second series also starts with a dc coefficient. As a result the coefficients are transmitted as much as possible in their order of significance. This particularly produces the largest possible clusters of zero value coefficients. Such clusters are transmitted as one compact run-length code so that an effective bit rate reduction is achieved, also for moving pictures. <IMAGE></p> |
