Television transmission system using transform coding

Claim 2

A television transmission system for the transmission of a picture represented by a digital picture signal from an encoding station to a decoding station, the encoding station having means for partitioning the picture into sub-pictures each of N×N pixels, and having a picture transform circuit adapted to perform a two-dimensional forward discrete cosine transform on each of the sub-pictures for generating an array of N×N basis picture weighting factors, said picture transform circuit being constituted by a cascade arrangement of two one-dimensional forward discrete cosine transformers each adapted to perform a forward transform in which an array of 1×N data elements d o to d N-1 applied to the transformer is transformed into an array of 1×N product elements p o to p N-1 , characterized in that the forward transform comprises the following steps: (a) the iterative determination of two sum elements and N-2 difference, elements by performing Q iteration steps, starting from the N data elements d o to d N-1 , in which in the q-th iteration step, the q-th order sum elements u q ,r and the q-th order difference elements v q ,r are obtained for which it holds that: u.sub.q,r =u.sub.q-1,2 -q+1.sub.N-r-1 v.sub.q,r =u.sub.q-1,r -u.sub.q-1,2 -q+1.sub.N-r-1 r=0, 1, 2, . . . 2 -q N-1 q=1, 2, 3, . . . Q u o ,r =dr u o ,N-r =d N-r Q is the largest integer which is smaller than or equal to -1+ 2 logN; (b) the iterative determination of a number of auxiliary sum elements by performing a number of further iteration steps on the difference elements obtained in the q-th iteration step in accordance with step a), while in the j-th further iteration step, the j-th order auxiliary sum elements w q ,j,m are obtained for which it holds that: w.sub.q,j,m =w.sub.q,j-1,m +w.sub.q,j-1,2 -q-j+1.sub.N-1-m m=0, 1, 2, . . . 2 -q-j N-1 j=1, 2, . . . 2 log(2 -q N) w q ,o,m =v q ,m w q ,o,2 -q N-1-m =v q ,2 -q N-1-m (c) generating sub-product elements by multiplication of each of the two sum elements u Q ,o and u Q ,1 obtained in step (a), as well as the difference elements v q ,r and the auxiliary sum elements w q ,j,m with the associated multiplication factors from a collection of multiplication factors, each multiplication factor being constituted by a linear combination of elements of the discrete cosine transform matrix; and (d) generating the product elements of the array of 1×N product elements P o to P N-1 by making linear combinations of said sub-product elements.

Claim 3

A television transmission system as claimed in claim 1, in which the decoding station is provided with an inverse picture transform circuit adapted to perform a two-dimensional inverse discrete cosine transform on a received array of N×N basis picture weighting factors for generating a sub-picture of N×N pixels of a transmitted picture, said inverse picture transform circuit being constituted by a cascade arrangement of two one-dimensional inverse discrete cosine transformers which are each adapted to perform an inverse transform in which an array of 1×N product elements p o to p N- 1 applied to the transformer is converted into an array of 1×N data elements, characterized in that the inverse transform is the transposed version of the forward transform.

Claim 4

An encoding station for use in a television system as claimed in claim 1, which, for the purpose of transmission of a picture represented by a digital picture signal, is provided with means for partitioning said picture into sub-pictures of N×N pixels each, and a picture transform circuit which is adapted to perform a two-dimensional forward discrete cosine transform on each sub-picture for generating an array of N×N basis picture weighting factors, said picture transform circuit being constituted by a cascade arrangement of two one-dimensional forward discrete cosine transformers which are each adapted to perform a forward transform in which an array of 1×N data elements d o to d N- 1 applied to the transformer is transformed into an array of 1×N product elements p o to p N- 1, characterized in that the forward transform comprises the following steps: (a) the iterative determination of two sum elements and N-2 difference elements by performing Q iteration steps, starting from the N data elements d o to d N- 1 in which in the q-th iteration step, the q-th order sum elements u q ,r and the q-th order difference elements v q ,r are obtained for which it holds that: u.sub.q,r =u.sub.q-1,r +u.sub.q-1,2 -q+1.sub.N-r-1 v.sub.q,r =u.sub.q-1,r -u.sub.q-1,2 -q+1.sub.N-r-1 r=0, 1, 2, . . . 2 -q N-1 q=1, 2, 3, . . . Q u o ,r =d r u o ,N-r =d N-r Q is the largest integer which is smaller than or equal to -1+ 2 logN; (b) the iterative determination of a number of auxiliary sum elements by performing a number of further iteration steps on the difference elements obtained in the q-th iteration step in accordance with step (a), while in the j-th further iteration step, the j-th order auxiliary sum elements w q ,j,m are obtained for which it holds that: w.sub.q,j,m =w.sub.q,j-1,m +w.sub.q,j-1,2 -q-j+1.sub.N-1-m m=0, 1, 2, . . . 2 -q-j N-1 j=1, 2, . . . 2 log(2 -q N) w q ,o,m =v q ,m w q ,o,2 -q N-1-m =v q ,2 -q N-1-m (c) generating sub-product elements by multiplication of each of the two sum elements u Q ,o and u Q ,1 obtained in step (a), as well as the difference elements v q ,r and the auxiliary sum elements w q ,j,m with associated multiplication factors from a collection of multiplication factors, each multiplication factor being constituted by a linear combination of elements of the discrete cosine transform matrix; and (d) generating the product elements of the array of 1×N product elements P o to P N-1 by making linear combinations of said sub-product elements.

Claim 5

A decoding station for use in a television system as claimed in claim 3, said decoding station being adapted to receive basis picture weighting factors which are transmitted by the encoding station, and being provided with an inverse picture transform circuit adapted to perform a two-dimensional inverse discrete cosine transform on a received array of N×N basis picture weighting factors for generating a sub-picture of N×N pixels of a transmitted picture, said inverse picture transform circuit being constituted by a cascade arrangement of two one-dimensional inverse discrete cosine transformers which are each adapted to perform an inverse transform in which an array of 1×N product elements p o to p N-1 applied to the transformer is converted into an array of 1×N data elements, characterized in that the inverse transform is the transposed version of the forward transform.