GB2243969A - Electronic clapperboard for television sound-vision synchronisation - Google Patents

Abstract

Apparatus for use in synchronising sound and vision components of a television signal includes means for generating at intervals simultaneous sound and vision timing signals for incorporation into the television signal. Preferably the timing signals are generated mechanically by means 10, 12, 14, 16 disposed in the field of vision of a television camera which generates the television signal. A system for synchronising sound and vision components of a television signal may also comprise analyser means for detecting the presence of the timing signals in the television signal and for measuring the time difference between them. <IMAGE>

Description

AN ELECTRONIC CLAPPERBOARD FOR SOUND-VISION SYNCHRONISATION A perceptible time difference between the sound and vision components of a television signal is an increasingly common impairment in transmitted television signals. Transmission time difference caused by different routing of sound and vision signals over long distances is an obvious cause of such time differences, but there are other causes of sound/vision relative delay that cause trouble in everyday national television operations. Vision signal processing using frame stores (e.g.

the use of synchronisers, standards converters, or digital video effects) will delay vision by varying amounts with respect to sound. Bit rate-reduction codecs for digital sound transission delay sound with respect to vision; advanced codecs with delays of 20 to 40 ms are being studied for low bit rate transmission.

The relative delays caused by equipment can of course be predicted and equalised, but there is not always time for the necessary adjustments to be made, particularly in live news programmes. Consequently, programme items are often transmitted with annoying failures of lip-sync, that is, failures of synchronisation between the soundtrack and the image seen by the viewer.

Relative delay can also occur during post-production processing of the video signal. In this case, it should be possible to correct the timing of the original sound to the vision using the time codes normally recorded on videotape; however an audiovisual check on timing may be useful in certain circumstances.

CCIR Report 1081 (published in Volume XI of the Recommendations and Reports of the CCIR) gives the following limits for relative delays in sound-vision synchronisation: sound advanced sound delayed w.r.t. vision w.r.t. vision 'detectable' 20 ms 40 ms 'subjectively annoying' 40 ms 120 ms In the film industry a mechanical clapperboard is filmed at the start of each 'take' to establish sound and vision synchronisation. Television, as a live electronic medium, needs a repetetive electronic device to ensure sound and vision synchronisation.

Ideally, such a device would operate continuously throughout the programme. Cooper (SMPTE Journal, September 1988, pp 695-698 and US Patent 4 703 355) has proposed incorporating in the vertical blanking interval of the video signal at source a filtered and binary quantised version of the sound signal so that the true sound signal transmitted separately can subsequently be synchronised. However the timing information thus provided would be lost when the video signal passes through a synchroniser or standards converter, both of which constitute significant sources of video delay.

In accordance with the invention, there is provided apparatus for use in synchronising sound and vision components of a television signal, the apparatus comprising means for generating at intervals simultaneous sound and vision timing signals for incorporation into the television signal.

The invention further comprises a system for synchronising sound and vision components of a television signal, the system incorporating such means for generating sound and vision timing signals, and analyser means for detecting the presence of the timing signals in the television signal and measuring the time difference between them. Preferably, the analyser means detects the vision timing signal and determines the time difference between the sound and vision timing signals by interpolation or extrapolation on the detected timing signal. This permits measurement of the time difference to greater accuracy than 20 ms.

A preferred embodiment of the invention will now be described in detail , by way of example, with reference to the drawings, in which: Figure 1 is a schematic diagram of a video signal for sound/vision synchronisation in accordance with the invention; Figure 2 shows an alternative vision signal to that of Figure 1; Figures 3 a to c show means for generating the vision signal of Figure 1; Figures 4 a to c show means for generating the vision signal of Figure 2; and Figure 5 shows interpolation to detect the time reference to greater accuracy than 20ms.

An electronic clapperboard may be needed to check sound/vision timing either at the source of the audio and video signals or simply over a transmission link.

The first application requires a device which is readily portable and can be placed in front of a television camera to give a changing vision signal with a coincident sound signal at some recognisable point of the picture. For the second application a camera and microphone are not necessarily available; what is needed is an electronic test signal generator that can give a coincident sound and vision test signal and can be incorporated as part of the normal lineup routine from a mobile control room or studio area to establish correct timing of vision. and sound.

Preferably, these two devices could be designed so that they produced very similar sound signals and very similar vision waveforms, both chosen to be easily detected electronically. The sound and vision signals resulting from the use of either device could then be analysed by the same equipment at the receiving end. This analysis equipment would measure timing errors between sound and vision and give a visual readout of error; it could also give out control signals to drive the appropriate correction devices.

An additional useful feature would be to devise the vision waveform such that it could easily be keyed into other test signals such as colour bars, camera lineup chart, VT programme clock etc. In practice this means that the vision signal should take up only a small part of the active picture area.

Bearing in mind the requirements outlined above, a possible vision signal might be as shown in Figure 1. A white bar 10 is moved horizontally and a time reference point is generated when the bar 10 is coincident with white bars 12 and 14 above and below the bar 10. The background 16 to the useful part of the signal is darker than the rest to ensure that the whole signal could be added resistively to another signal whilst retaining sufficient contrast between the white bars 10, 12 and 14 and the background 16. Alternatively, the active part of the clapperboard could be keyed into another signal.

An alternative vision signal is.shown in Figure 2. In this case a white-to-black transition 20 is moved down the picture and the time reference point is generated when the white area below the transition 20 is extinguished.

As discussed below, the first of these vision signals permits the use of a simpler analyser, while the second offers a simpler generator.

Both of these proposals could easily be implemented mechanically utilising a board with a "letterbox" cut in it. In order to produce a vision signal in the form illustrated in Figure 1, a belt with a white stripe on it would be positioned behind the letterbox. A suitable arrangement for achieving this is shown in Figures 3a to c.

A board 30 has a letterbox-type slot 32 formed in it. Behind the slot 32 is a casing 34 lined with black material. Inside the casing 34 is mounted an endless band 36, a portion of which is visible through the slot 32. The band 32 is driven by means of a sprocket wheel 38 which engages in sprocket holes 39 formed along one edge of the band 36. A plurality of white stripes 33 is formed on the band 36 in the region visible through the slot 32.

The time reference signal is generated each time one of the white stripes 33 moves into alignment with white bars 37 formed on the board 30.

To generate a black-white transition as described in connection with Figure 2, a rotating cylinder 40 is mounted with its axis parallel to the letterbox slot 32 as shown in Figures 4 a to c.

A white stripe 42 is formed along its length so as to be visible through the slot 32.

In either case, the sound could be produced mechanically by some sort of hammer-and-gong or clocker actuated by the moving belt or cylinder or as an electronic "beep" triggered by a photocell momentarily illuminated through a hole in the band 36 or cylinder 40.

Both the vision signal time references described above could easily be implemented electronically. The horizontally-moving bar would involve both vertical and horizontal counters; the vertically-moving bar would involve only vertical counters. Both could be generated digitally as three-level signals requiring only a rudimentary DAC consisting of logic gates feeding a summing junction through resistors.

A television signal is temporally sampled at the field rate (50Hz in Europe) so a simple electronic analyser working on the video signal can extract timing only to the nearest 20ms. As mentioned above, the minimum time difference detectable by a viewer is believed to be also about 20 ms, so it would be desirable to be able to measure relative delay to a closer tolerance, for example,5 ms.

In the case of an electronically-generated signal, there is no difficulty; the sound signal can be locked to the TV synchronising pulses so that the sound can start at only one point in the TV field. With the mechanical generator in front of a TV camera however, the time reference point, and hence the sound, may occur at any point in the TV field and the extraction of accurate timing from the video becomes a problem. Two possible solutions are to synchronise the movement of the belt or cylinder to the TV field sync, or to analyse the speed of movement of the white bar 12 or 20 so that its exact time reference point can be estimated by interpolation (in the case of the horizontally-moving bar 12) or extrapolation (in the case of the vertically-moving bar 40).

Estimating the exact time reference by interpolation is fairly simple in the cases of the horizontally-moving bar 12. As shown in Figure 5, the locations xi and X2 of the moving bar 12 are detected at times tl and t2 separated by one sample period, that is 20ms. The exact time at which the moving bar 12 was aligned with the fixed bars 14 and 16 can then be calculated to an accuracy greater than 20ms.

However, the case of the vertically-moving bar is complicated by interlace; it would be necessary to interpolate vertically to establish the position of the bar before extrapolating temporally to establish its timing.

Claims (12)

1. Apparatus for use in synchronising sound and vision components of a television signal, the apparatus comprising means for generating at intervals simultaneous sound and vision timing signals and for incorporation into the television signal.

2. Apparatus according to claim 1 in which the timing signals are generated mechanically by means disposed in the field of vision of a television camera generating the television signal.

3. Apparatus according to claim 1 or 2 in which the timing signals are generated electronically and are inserted into the television signal during transmission or processing thereof.

4. Apparatus according to claim 3 in which the electronically-generated timing signals are of similar form to those generated by the apparatus of claim 2.

5. Apparatus according to any preceding claim in which the vision timing signal occupies only a part of the picture area so that it can be combined with other test signals.

6. Apparatus for use in synchronising sound and vision components of a television signal, the apparatus being substantially as hereinbefore described with reference to the drawings.

7. A system for synchronising sound and vision components of a television signal, the system comprising apparatus according to any preceding claim for generating sound and vision timing signals and analyser means for detecting the presence of the timing signals in the television signals and measuring the time difference between them.

8. Apparatus according claim 7 in which the analyser means is operable to generate a correction signal to cause re-synchronisation of the sound and vision timing signals.

9. Apparatus according to claim 8 in which the correction signal controls the video and/or audio delays to effect re-synchronisation.

10. Apparatus according to any of claims 7 to 9 in which the analyser means is operable to detect the vision timing signal to an accuracy greater than 20 ms.

11. Apparatus according to claim 10 in which the analyser means detects the vision timing signal and determines the time difference between the sound and vision timing signals by interpolation or extrapolation on the detected vision timing signal.

12. A system for synchronising sound and vision components of a television signal, the system being substantially as hereinbefore described.