Abstract: In order to correct for common phase error in demodulated digital video broadcast signals which comprise data modulated on a multiplicity of spaced carrier frequencies, a demodulator includes analog to digital conversion means (20) for providing a series of digital samples of the broadcast signal, real to complex conversion means (22) for converting each digital sample to a complex number value, Fourier Transform means (24) for analysing the complex number values to provide a series of data signal values in complex number format for each carrier frequency, and signal processing means for processing the series of data signal values including phase error correcting means (30), the phase error correcting means including means for converting the data signal values from a complex number format to a phase angle format, means for determining a common phase error by assessing the phase of continual pilot signals in the broadcast signals and determining the variation in phase of the continual pilot signals between c

Abstract: Video signals are sampled at 4Fsc locked onto sub-carrier phase and frequency. The active picture lines of each field are divided into 6 blocks of 47 lines and the active line periods of those lines scrambled on a block by block basis by line order shuffling. The shuffling algorithm is generated by a line shuffling permutator driven by a PRBS generator (controls 36, 38). Active line period samples for one block are written in unscrambled form into a first memory block (32) and samples from the previous block are read out in scrambled form from a second memory block (34) for transmission. The complementary process takes place in the decoder.

Abstract: A filter is disclosed for adaptively filtering a PAL or NTSC luminance signal so as to attenuate the high frequency portion of the luminance band, wherein the chrominance information lies, when movement is detected. This reduces cross-color in the reproduced picture. Movement is detected by comparing the undelayed luminance signal with this signal delayed by an integral number of lines corresponding substantially to one field or one picture. The filter may precede a coder or follow a decoder.

Abstract: If an N.T.S.C. signal is to be demodulated using a line-locked sampling rate such as 800 times the line frequency, the ratio between the sampling period and the subcarrier period has the awkward value of 455/1600. Line-locked sampling is nevertheless desirable, e.g. in digital standards conversion. The subcarrier digital signals for digital demodulation are derived with the correct frequency from the sampling rate clock pulses CP by an adder and accumulator into which the number 582 is added modulo-2048 (the register is an 11-bit register) to generate an 11-bit number which represents the subcarrier phase angle at each sampling pulse and which addresses a ROM providing sin and cos values representing subcarrier samples. 582/2048 is not exactly equal to 455/1600 but 2048 is a desirable denominator as it is a power of 2 and implies a ROM of suitable size. ##EQU1## is exactly equal to 455/1600.

Abstract: A signal processing circuit (10) for color television signals, e.g. PAL signals, has first and second one-line delays (14, 16) connected in series to the input (12), an averager (18) connected across the two delays and a subtractor (20) connected to subtract the averager output from the output of the first delay (14). In this way the circuit provides (i) a first signal at the output of the first delay (14) which is the input signal delayed by one line period and (ii) a second signal at the output of the subtractor (20) which is the resultant of the average of the undelayed input signal and the input signal delayed by two line periods subtracted from the input signal. A band-pass filter (22) covering the chrominance band is connected to the output of the subtractor (20). A second subtractor (24) subtracts the filtered signal from the one-line delayed signal.

Abstract: Demodulated luminance (Y) and chrominance (U, V) signals are stored in field stores 1 to m each with line outputs 1 to n. Output lines are formed by summing the producs of the line outputs with weighting coefficients which are specified for each line in the output standards. Separate sets of coefficients for luminance and chrominance are stored in respective memories. The chrominance sets are arranged to reduce the vertical and temporal resolution of the chrominance information. This has the beneficial effect of reducing cross-color interference. In the implementation shown the signals are stored as time multiplexed digital samples YYYUV (i.e. with one pair of chrominance samples to three luminance samples) and the luminance and chrominance coefficient sets for the current output line are stored as a.sub.11 to a.sub.mn and b.sub.11 to b.sub.mn respectively in buffer registers. A switching waveform controls electronic switches to apply a.sub.11 to a.sub.