The present invention relates to the transmission and reception of colour television signals in component form and more particularly to a method and apparatus for handling the chrominance components of the signals. The present invention will be described in relation to a component television signal known as a Multiplexed Analogue Component (M.A.C.) signal although this is but one application of the present invention.

The present invention provides a method of assembling chrominance components signals comprising providing a plurality of fields of video signals in which the two different types of chrominance component signals alternate within each field, and arranging the fields so as to produce a resultant chrominance structure which is non- interlaced and contains alternate chrominance component signals both within each field and between adjacent fields, whereby to permit enhanced vertical chrominance resolution. Features and advantages of the present invention will be apparent from the following description of an embodi- ment thereof when taken in conjunction with the accompanyin drawings, in which:-

Figure 1 shows diagrammatically the line structure for the chrominance components of a MAC-type signal; Figure 2 shows a graph representing the repeat spectra in vertical and temporal frequency produced by the line structure of Figure 1 ;

Figure 3 shows diagrammatically the proposed type of line structure;

Figure 4 shows a graph representing the repeat spectra in vertical and temporal frequency produced by the line structure of Figure 3;

Figure 5 shows a diagrammatic representation of a filter producing the line structure shown in Figure 3; and

Figure 6 shows the characteristic of the filter shown in Figure 5 _* MAC is defined to code the two colour difference components (U and V in Europe) on alternate lines with a frame-reset sequence. This sequence is illustrated in Figure 1.

There are sound reasons why this sequence has been selected in place of the alternative four-field sequence which is not reset each frame. The main reason is that residual alias components in the frame-reset case are much less disturbing. In addition, if such a signal were ever to be processed in the studio then its two- field sequence would be easier to handle than a four- field sequence. An implication of using this sequence however is that the vertical chrominance resolution is limited to a quarter of the luminance vertical resolution capability. This is because effectively there are only 144 (575/4) vertical samples of each colour in an active picture scan whereas there are 575 active vertical samples of luminance. In order to achieve the full vertical resolution afforded by the 575 lines, field stores are required in the receiver. Nevertheless even if this complexity is allowed for colour the vertical chrominance resolution is still limited to that offered by 144 lines.

In frequency terms this limitation is evident by considering the repeat spectra (in vertical and temporal frequency) generated by this alternate line sequence, as illustrated in Figure 2.

The open circles are those repeats arising from the 625 interlace scan. The extra repeats introduced by alternate line omission in a frame reset fashion are illustrated by the shaded circles. These repeat spectra are half-amplitude but nevertheless they impose a Nyquist- limiting vertical chrominance frequency of 72 cycles per picture height (equivalent to 1. δMHz horizontally). In practice the chrominance vertical resolution obtained afte pre- and post- filtering is equivalent to 1.1MHz horizon- tally (-3dB).

If we wish to maintain a frame-reset sequence for transmission, techniques for achieving a vertical chromina resolution greater than the Nyquist limit described above are not obvious. Nevertheless the following line of thought gives a clue to a possibility:- with the current approach there appears to be a paradox -. Although there are half as many lines of each colour difference component as there are of luminance, the available vertical resolu¬ tion is only a quarter as great. As we have seen, this is due to the sampling structure formed by the chrominance lines. If we conceive of a chrominance line structure which is analogous to the luminance structure, except tht it is halved in density vertically, then we arrive at a 312-^-line interlace structure. Two such structures (one for each colour difference component) are illustrated interleaved in Figure 3«

In this case there are 2δS (575/2) vertical samples of each colour difference component, just as there are 575 for the luminance. In frequency terms, the repeat spectra appropriate to such a line-structure are shown in Figure 4« In this case the "Nyquist-limit" vertically is 144c/p (equivalent to 3«7MHz horizontally). As with the luminance field store processing is required in order to achieve this increased resolution. The relationship between the line-structure of

Figure 3 and the MAC frame-reset structure of Figure 1 is a simple one. A vertical shift upwards of alternate field in the interlace structure (Figure 3) by a "frame-line" (i.e. 1/575 of a picture height) yields the frame-reset structure (Figure l). It is thus possible for a

312- line interlaced colour structure (Figure 3) to be transmitted in a 625-interlace alternate-line frame-reset format (Figure 1 ) . A "conventional receiver" would process these colour lines as a true frame-reset structure (thus alternate fields would have a minor vertical shift) while a higher definition receiver would treat the lines as being part of the 312 line interlaced structure and thus an increased vertical resolution.

A higher definition receiver would receive the Figure 3 chrominance structure and undertake scan conver¬ sion to derive a 31 - line non-linterlaced structure. This provides a greater basis of information for subsequent vertical interpolation to provide a 625 line non-interlaced chrominance structure for combination with the 625 line non-interlaced luminance structure such a receiver would also produce.

If the source of chrominance signals is a special source such as a special camera, it is possible for the chrominance structure as shown in Figure 3 to be available directly from the source if the source scans with the basic 625 line inter-laced structure then the lines in alternate fields of the colour structure of Figure 3 are not available. In this case, therefore, values need to be interpolated. This interpolator can be the prefilter which shapes the vertical-temporal frequency spectrum of the colour components. As an example such a prefilter has been designed which would offer a -3dB vertical band¬ width equivalent to 2.14 MHz (on stationary or horizontally moving pictures). The filter coefficients are defined on a grid of 625-sequential-scan lines as illustrated in

Figure 5a, although in practice it will mean that "real lines" are filtered by the coefficients of Figure 5b and "interpolated lines" are provided by using the coefficient of Figure 5c. The vertical-temporal frequency response of such a filter is illustrated in Figure 6. If a similar post-filter is used prior to sequential i.e. non-interlace display in a higher definition receiver then the combined filter responses give a vertical resolution for stationary or horizontally moving scenes equivalent to 1.9MHz (-3dB). For vertically moving pictures this is reduced and in the worst case (having a 25Hz temporal frequency component) the combined response gives a resolution equivalent to 0.6MHz (-3dB). This is of course just an example. A different filter could be used or indeed one or both of the filters could be adaptive.

The effect on a conventional receiver would be primarily that of less severe pre-filtering. The results obtained would be more like those obtained with a 1-1 pre-filter rather than with a typical alias-rejecting seven-tap pre-filter.

By using the presently proposed invention, a 625 line MAC system can provide a vertical chrominance resolution that approaches, if not equals that of a 1125 line H DTV system in its transmission format.

. O