GB2115638A - Colour television coding and decoding - Google Patents

Abstract

A colour coder (Figure 6) transmits a PAL or NTSC signal comprising a luminance signal including alias components (2fsc sampler S12) and a one-dimensional chrominance signal C consisting of U+V and U-V on alternate lines, modulated on to the subcarrier fsc (sampler S22). The phasing of the Y and C signals is such that effectively they are transmitted in phase quadrature; although for transmission comb filters (F12, F22) restore the conventional UV quadrature phasing and the decoder (Figure 7) has filters (F11, F21) restoring quadrature phasing of luminance and chrominance prior to demodulation (samplers S11, S21), so that chrominance and luminance crosstalk is avoided. In order to compensate self impairments (U/V crosstalk, reduction of vertical chrominance resolution, luminance aliasing and reduction of diagonal luminance resolution due to luminance anti-aliasing pre- and post- filters (F01, F02)), luminance and chrominance compensation signals representing the lost information are extracted at the coder and transmitted separately (e.g. via a combining and separating system (S52, S62, F52, F62, F51, F61, S51, S61) analogous to the basic coder/decoder combination) and added at the decoder. The compensation signal may also be arranged to carry additional high- frequency chrominance components. <IMAGE>

Description

SPECIFICATION Colour television coding and decoding The present invention relates to PAL colour television coding and decoding.

Our UK Patent Application No 8005486 (Publication No 2044577A) describes a method of PAL coding and decoding based on a digital transmission system (see also our UK Patents Nos 1 534268-1 534270). in this system (see figure 1 of the present application) the transmitted signals are a iuminance signal Y2 sampled at twice the colour subcarrier frequency (2fisc) and a chrominance signal C1 consisting of the sum (U+V) of the chrominance signals and their difference (U-V) on alternate lines, sampled at the subcarrier frequency. The digital signal is, for convenience, termed W, and may be derived from separate luminance Y and chrominance signals U, V, or from a PAL signal, and similarly reformed into either PAL or YUV signals (indicated as PAL', Y', U', V').

The W signal cannot convey all the information inherent in the Y, U, V signals. In particular, the Y2 signals contain alias components due to the sub Nyquist sampling, whilst the C1 signal suffers from loss of vertical chrominance resolution and U/V crosstalk. Preferably the YUV signals are comb filtered before and after sampling (F01, F31, F41, Fo2 F32, F42), to minimise the impairment: thus luminance filters Fro1, Fo2 which may be identical to filters F", F,2, having (in the chrominance band) response peaks at integral multiples of fH/N (fH being the television line frequency and N an odd integer) reduce the aliasing at the expense of losing diagonal resolution.The chrominance filters, either summing or differencing over an Nline delay, are shown in figure 2. It will be noted that these filters may be based on delay elements of a line, field or picture period.

One of the advantages of the W transmission signal, is that, once the impairments have been introduced, they are not cumulative: the route W YUV-W is transparent.

This system also forms the basis of a method of PAL coding and decoding. The route YUV-W PAL' constitutes a coder and the route PAL-W YUV' constitutes a PAL decoder. In the coding process, the Y2 and C1 signals are converted to analogue form and the C1 signals modulated on subcarrier. Considering only the high frequency components, presence of alias components in the Y signal results in, effectively, double sideband suppressed carrier modulation of a carrier at fsc If the luminance and chrominance signals are appropriately phased, therefore, they can be transmitted on a single channel and separated again without crosstalk. This is entirely analogous to quadrature modulation.In order to form a PAL signal, however, the modulated chrominance is filtered to generate the 90" quadrature relationship between the U and V components, by a comb filter F22 which imparts a phase difference of 900 to components offset by -fH the spectrum of V having a qFfH offset due to the action of the PAL switch. A filter F2, at the decoder removes the U/V quadrature prior to synchronous demodulation, whilst corresponding filtering, and synchronous sampling at the decoder, of the luminance ensures that the overall filter combination does not introduce any spurious cross effects. These filters are also shown in Figure 2.

F2, comprises a halving subtractor 100, subtracting the undelayed input signal from that obtained via an N-line delay 101 followed by a chrominance bandpass filter 102. The output of F.1 is formed by subtracting (103) this output from the delayed signal, thus forming half the sum of the delayed and undelayed signals over the chrominance band, and allowing low-frequency components to pass unfiltered (with an N-line delay). Equalising delay 104 compensates the delay inherent in the bandpass filter.

Filters F,2, F22 have essentially the same responses as F", F2, prior to addition of the two components and comprise subtractors 110, 11 3, N-line delay 111, bandpass filter 11 2, equalising delay 11 4 and adder 11 5, but are reconfigured to share a single N-line delay whilst imparting no delay to the low-frequency components, so that the overall delay due to the four filters is N-lines through both Y and C paths.

It is thus a property of the system that the route W-PAL-W is transparent and thus the PAL coder followed by the PAL decoder is equivalent to the route YUV-W-YUV. In this route the Y signals never interact with the U and V signals and thus there are no luminance/chrominance cross-effects, although the self-impairments discussed above do, of course, remain. Thus we have a so-called "clean PAL" system in which luminance and chrominance can be combined, sent through a PAL "bottleneck" and separated without interaction.

Although the W signal cannot convey all the information inherent in the separate Y, U, V, signals, it can convey all the information contained in the PAL signal, and the PAL formed by the assembler circuit is identical to that entering the splitter, i.e. the route PAL-W-PAL (and hence PAL-W-YUV-W-PAL) is also transparent. It can be shown that this is true not merely for PAL signals generated by the method described, but for any signal entering the splitter.

For present purposes, it may be mentioned that alias components generated in the Y2/Y channel are cancelled in the assembler by alias components generated in the Cl/C channel.

This property of no interaction depends only on the transparency of the W-PAL-W path and thus only on the filters in that path, i.e. in the PAL assembler and splitter. As previously mentioned, these may be based on line, field or picture delay elements, but the filters F11, F21, F12, F22 must be based on the same delay element.

The filters in the YUV-W and W-YUV paths govern the form of the self-impairments in the received signals. Any combination of delay element bases may be used without destroying the 'clean' property, but clearly some combinations are preferable to others. Useful combinations are listed in our application 8005486. In addition we now believe that filters based on picture delay elements in the luminance path may be combined with filters based on line delay elements in the chrominance path to give an acceptable result.

As the digital transmission system of figure 1 considered as a PAL coder or decoder, accepts and delivers analogue signals, the routes from YUV to PAL and PAL to YUV may be realised without recourse to sampling by simulating the effect of the sampling using analogue multiplication. Then the coder and decoder may be realised using entirely analogue circuitry.

Furthermore, if the delay elements in the filters are everywhere the same the analogue circuits of the coder and decoder simplify to the forms shown in figures 3 and 4 with detail of the filter in figure 5.

In the coder, colour difference signals U, V (the iatter via a PAL switch 200) are modulated on phase-quadrature carriers at fsc (multipliers 201, 202) and added in adder 203 as in a conventional PAL coder. A subtractor 204 forms the difference of luminance Y and the combined colour signal, which, after passage through the filter L of figure 5, is subtracted in subtractor 205 from the luminance supplied via a compensating delay 206, to form the encoded PAL signal.

In the decoder of figure 4, chrominance is obtained from the PAL signal via filter L and demodulators 300, 301, low pass filters LP1, LP2 and V-axis switch 302, whilst the luminance output is obtained by subtracting (303) the output of L from the PAL signal delayed by equalising delay 304.

Although different in form, the coder and decoder of figures 3 and 4 are mathematically identical to those described above with reference to figures 1 and 2.

Such a coder and decoder enable luminance and chrominance to be combined in a PAL form of signal and separated without cross-effects. This is obtained, however, at the expense of self-effects which may or may not be acceptable, depending on the application. For example, with picturedelay based filters everywhere, the received picture suffers no impairment, if stationary, but movement portrayal may be unacceptable. On the other hand, with line-based filters everywhere, the movement portrayal is unaffected but there is a loss of spatial resolution. The exact compromise is still not yet known.

Such a coder and decoder are also compatible with the conventional PAL system. Thus the clean coded signals may be decoded by a conventional decoder to give YUV signals which suffer from reduced impairments at the expense of reintroducing some of the fine cross-effects.

Similarly, conventionally coded PAL signals may be decoded by a clean decoder to give YUV signals which suffer from the self-impairments but also from some of the coarse cross-effects.

This aspect of compatibility is very important since the change to any new form of PAL coding or decoding is bound to be gradual, so that mixed systems involving combinations of clean and conventional techniques are bound to arise.

In order that the received picture should suffer substantially no impairment (relative to a YUV standard) whatever the picture content, the present invention proposes that the clean PAL system just described be augmented so as to transmit the extra information.

The invention is defined in the claims below, to which reference should now be made.

Thus, a method is provided for deriving an augmentation signal, consisting of residual luminance and chrominance components, for a Weston clean PAL coder which, when suitably decoded, can be added, in a suitable way, to the signals obtained by clean decoding to give Y, U and V signals free of any self-impairments as well as cross-effects. Here again compatibility is an important issue since it would then be possible to have any combination of conventional, clean or augmented coders with any kind of decoder, with a corresponding hierarchy of quality. This implies that the augmentation signal is sent in a separate frequency band from the PAL signal so that it could be ignored by conventional and clean decoders.Thus, according to another aspect of the invention, a method is provided for frequency shifting the augmentation signal so that it can be combined with the clean PAL signal in such a way that the clean PAL signal can be extracted by simple low-pass filtering. According to yet another aspect of the invention, a method is provided for deriving a frequency-shifted augmentation signal with enhanced chrominance bandwidth.

Some embodiments of the present invention will now be described, by way of example, with reference to the accompanying drawings, in which: Figure 6 shows the circuit of a PAL coder based on the exchange system of Fig. 1 but with added circuitry to form an augmentation or compensation signal; Figure 7 shows the circuit of a PAL decoder, likewise based on the system of figure 1, but with added circuitry to decode the compensating signal and add it to the clean YUV signals; Figure 8 illustrates graphically the spectrum folding of the luminance compensation signal; Figures 9 and 10 are block diagrams of further embodiments of a coder and decoder according to the invention, analogous figures 3 and 4; Figure 11 is a modified version of the coder of figure 9; ; Figures 1 2 and 1 3 illustrate the portions of frequency spectrum occupied by the signals generated in the coders of figures 9 and 14; Figure 14 shows a modified coder and decoder, incorporating frequency shifting of the compensation signals; Figures 1 5 and 1 6 are diagrams similar to figures 12 and 1 3 illustrating the operation of the coder and decoder of figures 1 7 and 18; Figures 1 7 and 1 8 show modified versions of the coder and decoder of figures 9 and 10, for wideband chrominance; and Figure 1 9 shows a further modification.

In PAL coder and decoder shown in figures 6 and 7, those component parts also present in the arrangement of figures 1 and 2 are given the same reference numerals, whilst the additional circuitry is shown with dashed lines. The analogue-to-digital converter and digital-toanalogue converter combinations of figure 1 are shown simply as samplers Sil, S21, S12, S22.

The chrominance samplers S2" S22 are assumed to lack the low-pass output filter of the digital-toanalogue converter in figure 1, so that the following fs modulator is not required.

To emphasise the symmetry of the way in which the U and V signals are treated, the chrominance combiner is rearranged compared with figure 1, adders and subtractors 150, 160 forming (U+V)/and (UV)/signals before passing each through an identical filter F31,, F3," to a changeover switch X22 operating at half line frequency. The chrominance separator of the decoder is similarly arranged.

In the coder, the filter F01 filters the luminance Y over an N-line delay (subtracts diagonal luminance information). The lost information is obtained at a second output Y". This can be regarded as being produced by a filter Fro3; F01 and F03 constituting a splitter which can be identical to the PAL splitter F11, F2,.

References here and elsewhere to identical filters do not preclude them from being based on different line delays. In the drawings, the delays are marked Tc (chroma) Ty (luminance) and Tp (PAL). To obtain a transparent system, each class of delay must have the same delay period.

Similarly, the residual chrominance components lost in the filters F3,', F3," are extracted by subtractors 1 53, 1 63 which takes the difference over the N-line delays 1 51, 1 61, and a switch X52 synchronous with X22.

The residual luminance and chrominance signals (Y", C") are then sampled at appropriate relative phase in samplers S52, S62 and combined in an assembler (F52, F62) identical to the PAL assembler F,2, F22 to provide a single bandpass compensation signal.

At the decoder the combined residual luminance and chrominance signals are separated in a splitter circuit, involving filters (F51, F51), identical to those (F11, F2,) used for splitting and normal luminance and chrominance signals, and sampling (S5a S61) in the appropriate phase. Then means are provided for adding the separate luminance and chrominance signals to the normal clean signals.

To show that the coder/decoder combination is transparent, i.e. that the received YUV signals are identical to the source YUV signals (apart from the bandwidth restriction of the U and V signals implied by the bandpass filters in the assemblers and splitters) consider first the original assembler/splitter combinations F12, F22/F11, F21. It is a prime feature of this clean system that the assembler/splitter combination is transparent so that the signals A and B at the coder are identical to the signals C and D respectively at the decoder.

This is entirely analogous to the transmission of two signals on a single carrier by quadrature modulation and demodulation. In the same way, then, the compensation signals E and F at the coder are identical to the signals G and H respectively at the decoder because the compensation assembler/splitter F52, F62/F51, F6, is identical to that in the main path apart from the assembler output filter.

It will be noted that the sampling for both the residual signals is at the subcarrier frequency.

This is permissible for the residual luminance because, as fig 8 shows, the effect of sampling a bandpass signal centred on fsc at a frequency of 2fsc is identical, over the bandpass region, to that of sampling at f,,, or even Tfscl if the bandwidth is low enough. The residual chrominance is assumed to have a bandwidth of less than Tfsc so that sampling at f,, causes no aliasing.

Furthermore, as both residual signals are sampled at fsc the order in which they enter the input ports of the assembler is immaterial (provided the same order is retained at the splitter). In figure 6 the residual luminance is shown entering the differencer path of the assembler and the residual chrominance, the averager path. These could, however, be the other way round but there is an advantage in the arrangement of figure 6, as will be seen.

Now consider the Y path through the coder/decoder combination. Accepting that the signals at E and G are identical the Y path becomes identical to a splitter/assembler combination, i.e. the path PAL-W-PAL in figure 1, which is transparent. Thus the received Y signal is identical to the source Y signal (with the usual assumption that the bandpass filters are perfect).

Now consider the C path through the coder/decoder combination. Accepting that the signals at F and H are identical, the path can be considered as two independent routes, carrying normal and compensating chrominance.

Moreover, as the sampling at fsc is super Nyquist, the transparency argument can be conducted in purely analogue terms.

It will be recalled that the chrominance circuits of figures 6 and 7 are rearranged versions of figures 1 and 2 to emphasise the symmetry of the way the U and V signals are treated. The effect of the PAL V axis switch and post-filter adder in figure 1 is to form the signal (U+V)/and (U-V)/on alternate lines. In figure 6 this is made more explicit by first forming the signals (U+V)/g, and (UV)/, filtering these and then selecting them alternately using switch X22 which operates at half the line frequency. This arrangement gives an identical result to that of figure 1 provided both the filters F31, F31 are identical to F3, in figure 1.The advantage of the rearrangement is that it allows the argument to be based on the signals (Q+V)/and (U-V)/ which can be designated Q and I respectively.

(These signals are approximately equal to 0 and -I in the NTSC system.) Similarly the rearrangement of figure 7 is designed to recover the Q and I signals at the output of the crossover switch X12 which also operates at half the line frequency. Then the U and V signals are formed using the same network as that used for deriving Q and I. Thus transparency for U and V can be demonstrated by proving transparency for Q and 1.

Transparency for the Q and 7 signals can be demonstrated as follows. Switch X22 selects filtered Q and I alternately. Thus it isssufficlent to demonstrate transparency for either Q and I since they do not interact. Moreover, it is sufficient to consider the signals corresponding to picture points in a vertical line since the sampling is vertical.

Suppose switch X22 selects Q at even instants of time and I at odd instants. Then the transmitted samples of normal chrominance are given by SO=2(QO+Q N) 1=+( 11+ 11-N)' 52=T1(""02+""'02-N)' etc by virtue of the vertical filtering action of the Nline delay elements 151, 161, and averagers 152, 162. The semi-differencers 1 53, 1 63, and switch X62which operates in phase with switch X22, then form samples of compensating chrominance given by S'a = (Qc-Q-N), S'1= (I1-I1-N) S11=l I 1 11 S2'=+(2""02N), etc.

At the decoder adder 1 54 forms the quantities QO SO+SO t i 1=S1+S1, Q=S2+S2 etc.

whilst the subtractor 1 64 forms the quantities 0-N=5o5o" I 1-N=5151" 2-N=S2-S21 etc.

Thus even samples of Q, interspersed with odd samples of I, emerge from the adder and, provided N is odd, odd samples of 0, interspersed with even samples of I, emerge from the subtractor. Delay 1 55 is also of N line periods so that the delay output corresponding to thQdelay -N' 1-N' 12 N input sequence QO, l1,Q2... is I -N Q1-N Q which represents the sequence of values that is complementary to those from subtractor 1 64.

Thus, both I and 0 sequences are obtained by switching alternately between the outputs of delay 155 and subtractor 164, which is the function of the cross-over switch X,2.

It is to be noted that the delay of the whole process from coder input to decoder output, discounting the assembler/splitter delay, is N sample periods, i.e. N line periods. In practice N is 1,313 or 625 and the delay Tc is 1,313or625 line periods. If this differs from Tv the delay element in the luminance filter, then compensating delays must be placed in the appropriate path to equalise the overall processing delays of luminance and chrominance.

(In figures 6 and 7 no compensating delays are shown for simplicity).

If T, is equal to Tc then the luminance and chrominance spectra are complementary. If, in addition, these delay elements are equal to the delay elements of the assembler/splitter (the PAL filters), i.e. T,=Tc=Tp, then it is possible to collapse the coder and decoder, implicit in figure 1, to the forms of figures 3 and 4 as previously noted. In the same way it is possible to collapse the compensation signal paths of figures 6 and 7 to obtain very similar circuits. In fact if the residual luminance enters the differencer path of the assembler, as shown in figure 6, then the transfer functions for residual luminance and chrominance are the complements of those for the Weston clean luminance and chrominance. As a result the compensation circuits may share the processing elements to yield the enhanced coding and decoding circuits of figures 9 and 10.As can be seen, the enhanced complementary coder requires only an extra compensating delay 207 and adder 208 and the enhanced decoder requires only an extra compensating delay 305, an adder 306 and a subtractor 307, over and above the original versions. Moreover, the coder and decoder circuits become identical apart from the position of the chrominance modulator/ demodulator. The apparent need for two compensating delays plus the delays in L may be eased by noting from figure 5 that the filters delays the difference of luminance and chrominance. Thus it is only necessary to form the sum of luminance and chrominance (209), and delay this to obtain both delayed signals by addition (210) and subtraction (211) as shown in the alternative configuration of figure 11. This is useful where the delay elements are substantial, for example 313 or 625 line periods.

The spectrum of the compensation signal P' emerging from the enhanced coder in figure 9 is complementary to that of P, the clean PAL signal, i.e. the sum of P' and P is equal to a normal PAL signal. (This would not be true if the residual luminance entered the average path of the assembler in figure 6). If the filter L takes the form of a simple bandpass filter (i.e. not a form of Weston clean PAL coding) then P carries no highfrequency luminance and P' carries only the high frequency luminance in the chrominance band.

This arrangement is equivalent to the system proposed in our UK patent application No 8121212 except that P' also carries the high frequency chrominance sidebands that fall outside the passband of the bandpass filter. This is also true where L has the form of figure 5, and thus the collapsed form of system based on figures 9 and 10 has the capacity to deal with high-bandwidth chrominance which the expanded forms of figures 6 and 7 do not have. In this sense, figures 9 and 10 are not strictly equivalent to figures 6 and 7 but more general.

If the coder of figure 9 and the decoder of figure 10 are connected via their P and P' ports then the combination is able to transmit luminance of unrestricted bandwidth and chrominance of bandwidth equal to the subcarrier frequency. However, in the broadcasting system I the luminance bandwidth is restricted to 5.5 MHz and, for a compatible enhanced system, a single signal must be radiated. Thus the compensation signal, P', is preferably combined with the clean PAL signal P.

Consider first the case where the chrominance signal is narrowband. This can be defined as where the baseband chrominance signals have a bandwidth of 1.1 MHz so that the modulated chrominance occupies only the spectral region defined by the bandpass filter in the filters of figure 5. Figure 12 shows the spectrum of P and P' in this case.

To combine P and P' the compensation signal may be shifted spectraily by an amount (preferably, though not necessarily) equal to the subcarrier frequency and added to the normal clean PAL signal as shown in figure 13. At the receiver the compensation signal is filtered out and frequency shifted by the reverse amount to be used in the decoder of figure 10 in conjunction with the normal clean PAL signal, obtained by low-pass filtering. This is the method described in our application No 8121212.

Figure 14 shows the circuits for combining and separating P and P'. Multiplication of the compensation signal P' by the subcarrier frequency carrier produces spectral energy centred on zero frequency and 2fisc The former is rejected by the high-pass filter cutting at 3/2fsc.

At the receiver, the combination of the low-pass filter, compensating delay and subtractor forms a high-pass filter which selects the frequencyshifted compensation signal. Multiplication by the subcarrierfrequency carrier now produces spectral energy centred on fsc and 3fsc The latter is rejected by the low-pass filter cutting at 2fisc to leave the original compensation signal which can be used in the decoder, in combination with the normal PAL signal obtained by low-pass filtering.

Consider now the case where the chrominance signal is wideband. Figure 1 5 shows the spectrum of P and P' produced by the coder of figure 9 in this case. It will be noted that, due to the action of the filter L, the modulated chrominance in the normal clean PAL signal is still narrowband, the high frequency chrominance being carried in the compensation signal. Now in the broadcast system I the lower sideband of the modulated chrominance is allowed, for reasons of monochrome compatibility, to coexist with the luminance so that it extends down in frequency to about 0.5 MHz. As a result, it is impossible to eliminate completely cross-luminance from the received colour picture unless substantially all the luminance is lost or at least comb-filtered.If, then, monochrome compatibility is sacrificed by curtailing the lower sideband of the modulated chrominance, as in figure 15, an improved crossluminance performance will result.

The compensation signal with wideband chrominance cannot, however, be simply frequency-shifting as before without incurring the penalty of a further increase of bandwidth for the combined signal. On the other hand, the spectral space between the normal and frequency-shifted compensation signals is available. The solution is to allow the normal clean PAL signal to carry the upper sideband of the high frequency modulated chrominance and the compensation signal to carry the lower sideband, as shown in figure 16, giving a theoretical maximum chrominance bandwidth of wfscs Figures 1 7 and 1 8 show the circuits of modified forms of the coder and decoder of figures 9 and 10 to allow wideband chrominance to be transmitted.The filter' comprises the filter L except for the output bandpass filter 66, which is shown externally. At the coder the high-pass filter selects the upper sideband of the chrominance. In the spectral region where the bandpass filter transmits, the addition and subtraction of the high-pass filtered chrominance has no effect. Thus the circuit behaves as the coder of figure 9. Below the passband of the bandpass filter there is no extra chrominance because the high-pass filter cuts in the middle of the passband. Thus the circuit again behaves as in figure 9. Above the passband the chrominance upper sideband appears in the output of P and cancels the output of P'.

At the decoder the high-pass filter selects the upper sideband of chrominance in P. As in the coder the behaviour in, or below, the passband of the bandpass filter is identical to that of figure 10.

Above the pass band the chrominance upper sideband appears in C and cancels the output of Y.

Although the circuits of figures 1 7 and 18 can convey a chrominance bandwidth of, in principle, fsc when connected via their P and P' ports, the need to combine P and P' restricts the allowable bandwidth to 2fisc This bandwidth restriction may be done at the combination stage rather than at baseband using the combining and separating circuits of figure 19. These are the same as those of figure 14 except that the P signal must now be low-pass filtered to curtail the upper sideband of the modulated chrominance. The combination of the high-pass filter, subtractor and adder ensures a smooth transition between P and shifted P' signals at the frequency (3/2) fsc and ensures that only one filter characterises the system.

The invention is also applicable to NTSC systems. Our UK patent application No 8010814 (Publication No 2045577A) describes "clean" NTSC coding and decoding systems similar to the PAL arrangements shown in figures 1 to 5 of the present application, modified to take account of the fact that the subcarrier has a half line frequency offset (as compared with the quarter line frequency offset of the PAL system by utilising phase-perturbed waveforms fsc and 2sc The manner in which the chrominance signals 1, Q are combined also has to be different. Both these factors are fully discussed in the abovementioned patent application and will not be repeated here.

Thus the arrangements of figures 6 and 7 can be converted into an NTSC coder and decoder by (i) supplying phase-perturbed NTSC subcarrier fsc/ 2fisc to the samplers (ii) insertion at the inputs of the chrominance samplers S22, S62 at the coder and the outputs of the samplers S2" S61 and the decoder of switched inverters operating at quarter line frequency so as to invert alternate line pairs and (iii) replacement of adders 1 50, 1 56 and subtractors 160, 1 66 by appropriate I, Q matrixing. Similarly the arrangements of figures 9 and 10 will operate for NTSC provided that appropriate I, Q matrixing is provided (the V-axis switch being removed), and the multiplier 62 of figure 5 is fed with a phase perturbed 2fisc sinewave of appropriate phase.

Claims (21)

Claims

1. A colour television signal encoding apparatus comprising first processing means for spectrum folding an input luminance signal with respect to a frequency twice the colour subcarrier frequency to produce alias components in the video band, to produce a chrominance signal of limited vertical resolution modulated onto the colour subcarrier frequency and to combine the processed luminance and chrominance signals so as to produce a PAL or NTSC signal from which the said processed signals can be separated without interaction, and second processing means for extracting luminance and chrominance compensation signals representing the information omitted by the first processing means from the input luminance and chrominance signals respectively.

2. An apparatus according to claim 1 in which the second processing means includes means operable to combine the two compensation signals into a single signal by spectrum folding the luminance compensation signal with respect to a frequency twice the colour subcarrier frequency, modulating the chrominance compensation signal on the colour subcarrier and comb filtering and combining the said signals, such that they may be separated without interaction.

3. An apparatus according to claim 2 including means for frequency-shifting the compensation signal to a frequency band outside the video band.

4. An apparatus according to claim 2 in which the frequency is shifted by an amount equal to twice subcarrier frequency.

5. An apparatus according to claim 3 or 4 in which the chrominance compensation signal includes high definition chrominance components outside a bandwidth equal to the difference between the colour subcarrier frequency and the upper limit of the luminance band.

6. An apparatus according to claim 5 when dependent on claim 4, in which the high definition chrominance components are transmitted as a lower sideband of the 2f5C carrier and an upper sideband of the colour subcarrier.

7. An apparatus according to any one of claims 1 to 6, in which the first processing means comprises luminance processing circuitry adapted to cause spectrum folding of a luminance input signal with respect to a frequency twice the colour subcarrier frequency to produce alias components in the video band, and to comb filter the resultant luminance signal with a modulus sine response having peaks at integral multiples of fH/N' where fH is the television line frequency, and N, is an odd integer; chrominance processing circuitry adapted to form the U and V signals into a composite chrominance signal based on U+V and U-V on alternate lines, to modulate the chrominance signal onto the colour subcarrier frequency, and to comb filter the modulated chrominance signal with a modulus sine response having peaks at odd integral multiples of fH/2N,;; and combining means for combining the luminance and modulated chrominance signals.

8. An apparatus according to any one of claims 1 to 6, in which the first processing means comprises luminance processing circuitry adapted to cause spectrum folding of a luminance input signal with respect to a phase perturbed frequency twice the colour subcarrier frequency to produce alias components in the video band, and to comb filter the resultant luminance signal with a modulus sine response having peaks at integral multiples of fH/N, where fH is the television line frequency, and N, is an odd integer;; chrominance processing circuitry adapted to form the I and 0 signals into a composite chrominance signal based on I and 0, to modulate the chrominance signal onto the phase-perturbed colour subcarrier frequency, and to comb filter the modulated chrominance signal with a modulus sine response having peaks at odd integral multiples of fH/2N,; and combining means for combining the luminance and modulated chrominance signals.

9. An apparatus according to claim 7 or 8, including luminance prefiltering means adapted to comb filter the luminance signal with a modulus sine response having peaks at integral multiples of fH/N2 where fH is the television line frequency and N2 is an odd integer, the second processing means being arranged to extract the luminance components rejected by the prefiltering means and to subject such components to spectrum folding with respect to a frequency twice the colour subcarrier frequency to generate compensating alias components.

10. An apparatus according to claim 7, 8 or 9, including chrominance prefiltering means, the second processing means being arranged to extract the chrominance components rejected by such prefiltering means.

11. An apparatus according to claim 9 and 10 when dependent on claim 2 in which the comb filters applied to the folded luminance compensation signal and modulated chrominance compensation signal are identical to those applied to the folded luminance signal and modulated chrominance respectively or vice-versa.

1 2. An apparatus according to any one of claims 2 to 6 in which the first processing means comprises inputs for receiving Y, U and V signals, V-axis switching means, multipliers for modulating the U and V signals onto subcarrier at orthogonal phase positions, combining means for combining the quadrature modulated signals, the combined chrominance signal passing through a luminance-stop filter and the luminance signal having subtracted from it luminance components passed by the said filter, the luminance stop filter including means for spectrum folding about a frequency twice the subcarrier frequency and comb filtering means such that luminance and chrominance components can be extracted from the combined signal without interaction, and the compensation signals are generated by filtering the combined luminance signals in the said filter and subtracting from the combined chrominance signal those components thereof which are passed by the said filter.

13. An apparatus according to any one of claims 2 to 6 in which the first processing means comprises inputs for receiving Y, I and Q signals, multipliers for modulating the I and Q signals onto subcarrier at orthogonal phase positions, combining means for combining the quadrature modulated signals, the combined chrominance signal passing through a luminance-stop filter and the luminance signal having subtracted from it luminance components passed by the said filter, the luminance stop filter including means for spectrum folding about a phase-perturbed frequency twice the subcarrier frequency and comb filtering means such that luminance and chrominance components can be extracted from the combined signal without interaction, and the compensation signals are generated by filtering the combined luminance signals in the said filter and subtracting from the combined chrominance signal those components thereof which are passed by the said filter.

14. An apparatus according to claim 12 or 13 including a subtractor to form the difference of the luminance and chrominance signals, and an adder and subtractor for adding and subtracting the difference, after passage through the luminance-stop filter, to/from the luminance and combined chrominance signals to form a PAL or NTSC output signal and a combined compensation signal respectively.

1 5. A colour television decoding apparatus for use with the coding apparatus according to claim 1, comprising first processing means for separating the luminance and chrominance signals from the PAL or NTSC signal, and second processing means for receiving the compensation signals and combining them with the separated luminance and chrominance compensation signals.

1 6. An apparatus according to claim 15, in which the first processing means comprises: luminance processing circuitry adapted to comb filter the PAL signal with a modulus sine response having peaks at integral multiples of fH/N" where fH is the television line frequency and N, is an odd integer, and to cause spectrum folding of the comb filtered signal with respect to a frequency twice the colour subcarrier frequency to provide a luminance signal; and chrominance processing circuitry adapted to comb filter the PAL signal with a modulus sine response having peaks at odd integral multiples of fN,, to multiply the last-mentioned comb filtered signal by a signal of subcarrier frequency, and to separate the multiplied signal into its orthogonal components.

1 7. An apparatus according to claim 15, in which the first processing means comprises: luminance processing circuitry adapted to comb filter the NTSC signal with a modulus sine response having peaks at integral multiples of fH/Nl, where fH is the television line frequency and N, is an odd integer, and to cause spectrum folding of the comb filtered signal with respect to a phase-perturbed frequency twice the colour subcarrier frequency to provide a luminance signal; and chrominance processing circuitry adapted to comb filter the NTSC signal with a modulus sine response having peaks at odd integral multiples of fH/N" to multiply the last-mentioned comb filtered signal by a signal of subcarrierfrequency, and to separate the multiplied signal into its orthogonal components.

1 8. An apparatus according to claim 15, in which the first processing means is arranged to separate the chrominance component by passing the input signal through a luminance stop filter, and to separate the luminance component by subtracting from the input signal those components thereof passed by the filter, the filter including means for spectrum folding about a frequency twice the colour subcarrier frequency and comb filtering means, and the second processing means is arranged to add to the luminance and chrominance those components of a combined compensation signal supplied thereto which respectively do and do not pass the filter.

1 9. An apparatus according to claim 15, in which the first processing means is arranged to separate the chrominance component by passing the input signal through a luminance stop filter, and to separate the luminance component by subtracting from the input signal those components thereof passed by the filter, the filter including means for spectrum folding about a phase-perturbed frequency twice the colour subcarrier frequency and comb filtering means, and the second processing means is arranged to add to the luminance and chrominance those components of a combined compensation signal supplied thereto which respectively do and do not pass the filter.

20. An apparatus according to claim 1 8 or 1 9 including a subtractor to form the difference of the PAL or NTSC signal and combined compensation signal, and an adder and subtractor for adding and subtracting the difference, after passage through the luminance-stop filter, to/from the PAL or NTSC signal and compensation signal to form respectively output luminance and chrominance signals.

21. Colour television signal processing apparatus substantially as hereinbefore described with reference to figures 5 to 1 9 of the accompanying drawings.