Abstract: An image data recording and transmission system is described in which a data compressor 76 decorrelates input image data into sub band component data, a data recorder 78 stores the sub band data, a data decompressor 80 decompresses data read from the data recorder and a transmission signal generator 82 produces a bandwidth limited transmission signal from the decompressed data. The combined action of a data sequencer 18, a quantiser 14 and an entropy encoder 20 within the data compressor act to remove from the data stream that information corresponding to frequencies not transmittable with the bandwidth limited transmission signal (e.g. PAL or NTSC) subsequently produced by a transmission signal generator 82. Accordingly, the data recorder need not use storage capacity recording data which cannot be used by the transmission signal generator.

Abstract: A signal processing system includes a compression encoder 16 for compressing an input block of data samples including a plurality of sub-blocks, each sub-block including an array of data samples, into a block of compressed data. A sequencer 14' accesses the data samples of the input block to be passed to the compression encoder 16 with the sub-blocks being accessed in a predetermined first order and the data samples within each sub-block being accessed in a predetermined pseudo-random order such that successively accessed locations in the array of samples for a sub-block are scattered throughout that sub-block. The pseudo random accessing sequence enables occupancy of a buffer 73 for the compressed data block to be linearised. The sequencer can include a counter 62, the output of which addresses a look-up table 65 for generating the pseudo-random sequence in a repeatable manner.

Abstract: During shuttle mode operation in a system for the reproduction and storage of frequency separated video signals, the read head intersects a plurality of tracks corresponding to different fields of the image rather than following complete tracks. Accordingly, the full information for any one field is not recovered. To provide an image during shuttle mode, the system reads the dc luminance information of the tracks it intersects at the part where this information is stored. As the information content of the dc luminance information is relatively high, a recognisable image can be formed from this information alone. The dc luminance information is stored in the centre of the tracks with the other components arranged in order of increasing frequency towards the edge of the tape. The data which is most important to generation of a high quality image is stored in a place where it is less likely to be degraded.

Abstract: In deriving a key signal for a digital mixer from an input video signal, most values of the key signal are derived by a clip and gain circuit. However, in the special case where the normalized product values for adjacent pixels lie on opposite sides of the upper and lower bounds, the following equations are used to determine the key values (K.sub.i and K.sub.h) for the two adjacent pixels:K.sub.i =0.5 (Ub-Lb)*[P.sub.b -0.5 (Ub-Lb)]/(P.sub.h -P.sub.i),K.sub.h =K.sub.i +0.5 (Ub-Lb),where Ub and Lb are the upper and lower bound values respectively, and P.sub.i and P.sub.h are respectively the lower and higher product values of said two adjacent pixels.

Abstract: A digital video tape recording/playback system having a plurality of recording heads (A, B, C, D) on a rotating head mechanism 94 for recording slanting tracks 108 which extend diagonally across a tape 82 provides processing of video signals in a plurality of data processing channels with selective connection of the data processing channels to the recording heads such that data from a plurality of data processing channels are stored in respective portions of a slanting track (A+/A-; B+/B-; C+/C-; D+/D-). The data in each data processing channel and recorded in each track portion is sampled with a substantially even distribution over a field of video signals to provide for effective concealment of recording/playback errors. Preferably, two data processing channels are recorded on each track, the respective portions being located in upper and lower portions of the tape, respectively.

Abstract: An image data processing apparatus generates output data representing an image in a spatial domain from a plurality of sets of input data, each representing in a spatial frequency domain a sub sampled version of the image in the spatial domain. An error detector detects any erroneous parts of the plurality of sets of input data in the spatial frequency domain. Each of the plurality of sets of input data is transformed from the spatial frequency domain into a sub sampled version of the image in the spatial domain. Any parts of the sub sampled versions that are dependent upon an erroneous part of a set of input data are interpolated from one or more spatially adjacent parts of one or more other sub sampled versions. The sub sampled versions are combined to form the output data.

Abstract: A sampling frequency converter having an oversampler for operating on input data supplied to the converter, a signal processor for controlling operation of the sampling frequency converter and for generating oversampled output data from the oversampled input data, and a downsampler for generating output data at the output sampling frequency from the oversampled output data. The input and output of data by the signal processor are controlled by respective input and output interrupt signals, the output interrupt signal having a lower interrupt priority than the input interrupt signal.

Abstract: During shuttle mode operation in a system for the reproduction and storage of frequency separated video signals, the read head intersects a plurality of tracks corresponding to different fields of the image rather than following complete tracks. Accordingly, the full information for any one field is not recovered. To provide an image during shuttle mode, the system reads the dc luminance information of the tracks it intersects at the part where this information is stored. As the information content of the dc luminance information is relatively high, a recognisable image can be formed from this information alone. The dc luminance information is stored in the center of the tracks with the other components arranged in order of increasing frequency towards the edge of the tape. The data which is most important to generation of a high quality image is stored in a place where it is less likely to be degraded.

Abstract: A sub-band filter system in which the frequency separated data is multiplexed together at the output of each filter stage. The stream of multiplexed data produced can have a block-by-block format or a sample-by-sample format. In the latter case, the filtering can be effected by a finite impulse response filter with switchable tap coefficients for effecting both decimation and high and low pass filtering.

Abstract: A data compression system in which decorrelated data is fed to an entropy encoder (20) in batches having an internal monotonic decrease in average information content from start to end. This results in longer zero runs and more rapid attainment of end of block codes. A data sequencing circuit (18) comprising nested inter-sub band (60) and intra-sub band (72) counters for use with sub band filtered video data is also described.

Abstract: Apparatus for storing a video signal includes a signal compression arrangement constituted by a spatial two-dimensional sub-band filtering arrangement that filters a digital video signal to form data sets constituting respective sub-bands of the two-dimensional spatial frequency domain, a quantizer that quantizes the data sets in accordance with respective values which are such that the amount of quantization of one of the data sets constituting a sub-band to which dc luminance information of the signal is at least predominantly confined is less than the average of the amounts of quantization of the remaining data sets, and an entropy encoder that selectively encodes at least some of the quantized data sets so that the quantized data sets, as selectively entropy encoded, form a compressed video signal. The signal compression arrangement is followed by a storage arrangement for storing the compressed video signal.

Abstract: A digital video signal is compressed by spatial sub-band filtering to form data sets constituting respective sub-bands of the two-dimensional spatial frequency domain. The data sets for a field or frame are stored. A first sequencer controls writing, in accordance with a desired sequence, of the stored data to a quantizer in which they are quantized in accordance with respective values, those values being such that the amount of quantization of at least a data set constituting a sub-band to which dc luminance information of the signal is at least predominantly confined is less than the average of the amounts of quantization of the remaining data sets. The quantized data sets are then encoded in an entropy encoder which has a first coding portion for coding quantized data representative of dc luminance information and a second coding portion for coding quantized data representative of ac luminance information.

Abstract: To enable a digital composite color video signal (NTSC or PAL) to be compressed directly, that is without dividing it into its components and compressing the components separately, the digital composite color video signal is decorrelated, for example by spatial sub-band filtering, to form data sets, for example sub-bands, representing dc luminance information, ac luminance information and dc chrominance information. The data sets are then quantized in accordance with respective quantization values from a quantization matrix, those quantization values being such that the amount of quantization of each of some of the data sets to which the dc luminance information and the dc chrominance information is at least predominantly confined is less than the average of the amounts of quantization of the remaining data sets. At least some of the quantized data sets are then entropy encoded.

Abstract: A filter for filtering analog or digital samples using two sections, each including a multiplier stage having a plurality of multiplier sections, each with a respective output but a common input for applying a plurality of coefficients to a sample input to the multiplier stage, and a delay and summing network comprising a plurality of delay sections and alternate summing networks connected in series, the output of one of the multiplier sections being connected to the first delay sections in the series with each output of the other multiplier sections being connected to a respective summing network, the delay and summing network of one section being interleaved in series with the delay and summing network of the other section, and further comprising means for omitting alternate samples.