GB1222022A - Detecting registration errors between television signals - Google Patents

Abstract

1,222,022. Television. BRITISH BROADCASTING CORP. 7 July, 1969 [23 Aug., 1968]. No. 40508/68. Heading H4F. To detect a registration error between two television signals representative of the same scene, e.g. between any two channels of a colour television transmission, an error signal is generated having a predetermined duration and an amplitude dependent upon the time elapsing between the occurrence of a predetermined waveform in one of the television signals and the occurrence of a corresponding waveform in the other television signal. The error signal may be displayed on a black and white monitor, the screen controls of which may be set to give a mid-grey over the screen when there is no registration error, and when there is an error light and dark bands on the screen to the right of the lines in the picture giving rise to the error signal, e.g. a registration chart comprising a grid of white lines on a black background. For complete camera registration both horizontal and vertical registration error signals are derived and applied to respective scan displacement controls. For vertical registration the chart must include a diagonal line. The error signal generating arrangement employs one or two "timer" circuits which have "start", "stop" and "reset" inputs and an output. If a pulse is applied to the "start" input and subsequently a pulse is applied to the "stop" input, a pulse appears at the output, the amplitude of which is proportional to the magnitude of the time interval between the pulses applied to the "start" and "stop" inputs. The duration of the output pulse is dependent upon the time interval between the reception of the "stop" input and a pulse at the "reset" input. The timer does not produce an output unless the pulses are received in the sequence start-stop-reset-start &c. In Fig. 2 two such timers 10A, 10B are used and if a pulse in channel A arrives before the corresponding pulse in channel B an output pulse is produced by timer 10A only, the amplitude of which is proportional to the time interval between the pulses. The pulses applied to timer 10B are in the wrong sequence and thus there is no output. If the pulse in channel B arrives first an output pulse is only formed by timer 10B, and this pulse is inverted in order that the sense of the registration error may be indicated. The timers in Fig. 2 may each comprise a ramp generator followed by a switch and hold (store) circuit as shown in Fig. 8, the reset pulses which are applied to the hold circuits being derived via monostable circuit 64, differentiator 66 and clipper 67. Blanking circuits 60A and 60B are included in each channel and operated by a delayed pulse from monostable circuit 64 to prevent any pulses occurring after the maximum timing error being detected by the arrangement. In Fig. 4 only one timer 10C is used and channel A is connected to the start input, and channel B to the stop input via delay 40 (T2) and the reset input via delays 40 and 22. The delay period of delay circuit 40 is arranged to be just greater than the maximum mis-registration to be detected and thus a pulse applied to input 20B always reaches the timer after the pulse applied to input 20A. The magnitude of the output pulse from the timer is proportional to T2 plus the amount by which the pulse in channel A is ahead of the pulse in channel B. If the pulses in channels A and B are coincident the output pulse has a reference magnitude. A compensating pulse of the opposite magnitude may be generated 41 so that the polarity of the error signal indicates the sense of the registration error. The registration error signals may alternatively be derived by arranging for the input pulses to initiate linear ramp waveforms of opposite polarity which are added together, or to initiate oscillations the phase of which are compared in a phase discriminator.