Abstract: A coded data stream with end-of-block (EOB) words is divided into groups each having n EOB words so that loss of an EOB word can be determined at the receiver. When such loss is determined, EOB word(s) is or are inserted locally at the receiver to prevent shift of correctly decoded blocks relative to their true position in the image to be reconstructed.

Abstract: To be able to transmit television pictures in a digital format on a medium having a limited bandwidth, the picture are subjected to a coding operation and thereby converted into a series of transmission bit blocks TRB.sub.n of variable length (for example DCT and variable length coding). Each transmission bit block is divided into a main block H.sub.n having at least the most important code words and an auxiliary block S.sub.n comprising the other code words. By carrying out a formatting operation, a group of transmission bit blocks is converted into a channel bit block consisting of a number of sub-channel bit blocks SCB.sub.n of equal length. The main blocks are accommodated at the first bit positions of the sub-channel bit blocks and the bits of the auxiliary blocks are spread over the remaining bit positions. It is thereby achieved that the most important information is always found at fixed locations so that the influence of sync losses is reduced considerably.

Abstract: Apparatus for deriving a compatible low-definition interlaced television signal and other components for restructuring the original signal from an interlaced high-definition television signal. The HDTV signal is divided into four component signals by horizontal and vertical low-pass and high-pass QMF filtering. For the vertical filtering, one field of a frame is filtered by an odd-length filter, the other field by an even-length filter. A time delay of one sample is introduced before the high-pass odd-length QMF filter. The component signals are each subsampled by a factor of two after filtering. The component signal covering the low-frequency horizontal and low-frequency vertical range is a reduced definition interlaced TV signal. To reconstruct the original HDTV signal, all four component signals are upsampled and filtered and a one sample delay is introduced following the low-frequency odd-length filter.

Abstract: In a television transmission system a digital luminance signal Dy and two digital color difference signals DCHR(1) and DCHR(2) are applied to a distribution circuit 3 at the transmitter end, which digital color difference signals are applied thereto possibly after having been subjected to a vertical decimation operation with a decimation factor R. The distribution circuit 3 partitions all these signals into groups of Q sample per group. After receiving an even number of luminance groups and an even number of chrominance groups, a first half of this number of luminance groups and a first half of this number of chrominance groups is applied to a first auxiliary picture signal output 3(4) of the distribution circuit 3, while simultaneously the other half of the number of luminance groups and the other half of the number of chrominance groups is applied to a second auxiliary picture signal output (3(5) of the distribution circuit.

Abstract: In a system for the transmission of video signals digitized by means of differential pulse code modulation, the constituent signal words are submitted to a variable-length coding operation. A further bit rate reduction is then realized in that the allocation of the code words to the signal words is periodically adapted to the local statistics of the video signal.

Abstract: In a television transmission system for digital picture signals each picture is divided into sub-pictures of N.times.N picture elements for the purpose of limiting the bit rate. Each sub-picture is subjected to a two-dimensional forward transform to determine coefficients. These coefficients are converted into a series of serially occurring coefficients in which the AC coefficients in this series are arranged in such a way that the magnitude variation of these AC coefficients in the series is monotonous. The magnitudes of the AC coefficients thus arranged are subjected to a differential encoding so that the difference between the magnitudes of two successive AC coefficients is transmitted instead of the magnitude of the AC coefficients themselves.