| 摘要 |
The invention relates to a method for processing a data signal which is transformed into an n-dimensional matrix form and which is to be processed by an n-dimensional quantizing matrix. In the invention the n-dimensional quantizing matrix is processed by an n-dimensional transfer function matrix before processing the data signal. In order to form the transfer function matrix at least one transfer function is selected, and it is given a fixed point in the n-dimensional transfer matrix. At least a part of the elements of the transfer function matrix are processed by at least one transfer function, so that elements at a predetermined distance from the fixed point are processed by a processing quantity dependent on the transfer function. <IMAGE> |
