GB2139448A - High Definition Video Signal Transmission - Google Patents

Abstract

A high-definition video signal having 4N+1 lines per picture on two interlaced fields is transmitted on two video-bandwidth channels, as a main signal and an auxiliary signal respectively, the main signal being compatible with a standard 2N+1 lines per picture, 2:1 interlaced signal. Each line of the main signal is formed of a combination of two (or more) lines of the same field of the input signal, and each line of the auxiliary signal is also formed of a different combination of the lines of the same field of the input signal. The auxiliary signal comprises a 2N+1 DIVIDED 2 lines per picture 2:1 interlaced signal.

Description

SPECIFICATION High Definition Video Signal Transmission This invention is concerned with systems for transmitting and receiving a video signal having an increased number of independent picture elements (pixels) compared with a conventional signal. Such a video signal may be termed a HDTV (high definition teievision) signal.

It can in principle be applied to component signals or any composite signal in which the components are line simultaneous and not on a carrier (or on a line-locked carrier).

It is assumed that the HDTV source scanning format (e.g. from the camera) has a 2:1 interlace, with nominally double the number of lines compared with a conventional format. For example, if the conventional format were 625/56 2:1, the source format would be 1249/50 2:1.

There is a requirement that such a signal should be transmitted through two video-bandwidth channels. The first, main, signal would be a 625/50 2:1 compatible signal; that is to say this signal alone could be received by a conventional receiver and displayed at normal definition. On the other hand an HDTV receiver would receive in addition the second, auxiliary, signal and use this to increase the definition of the displayed signal.

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

The invention will be described by way of example, with reference to the drawings, in which: Figure 1 is a vertical-temporal diagram showing the relation between vertical position and time (i.e. field) of some of the source lines; Figure 2 illustrates the timing relationship of the output lines for a hypothetical HDTV input signals with 9 lines per field; Figure 3 is a vertical-temporal diagram for this hypothetical system; Figure 4 is a block diagram of the transmitter in a system embodying this invention; Figure 5 is one possible buffer timing diagram for the transmitter of Figure 4; Figure 6 is an alternative buffer timing diagram for the transmitter of Figure 4; and Figure 7 is a block diagram of a corresponding HDTV receiver.

A first embodiment of the invention will be described with reference to the drawings. Figure 1 is a vertical-temporal diagram showing the relationship between vertical position and time of a few of the lines generated at the source. These are denoted by the circles. The first, main, channel carries information derived from these lines and sent on a conventional format having nominally half the number of lines. This signal can be displayed on a conventional receiver without further processing. The transmitted information is derived by linear intra-field interpolation of the source lines, that is, taking a fraction a of one line and a fraction 1-a of the neighbouring line in one field, and a fraction 1/2-a of one line and a fraction 1/2+a of the neighbouring line in the other field.These lines are denoted by crosses in Figure 1 in the positions corresponding to the interpolation used.

The second auxiliary or augmentation channel carries an approximately equal number of lines derived by different linear combinations (viz. b: (1-b) and c: (1-c) on alternate fields) of the same neighbouring values. These lines are denoted by +'s in Figure 1, but their values may or may not correspond to real positions in the space. For example, the sum of the fractions weighting the neighbouring values may equal unity as noted above, or may be zero or an intermediate value.

The lines of the high definition picture are recovered by combining the lines of the incoming signals on both the main and auxiliary channels. It will readily be appreciated that this is possible because, for each pair of source lines, two independent linear combinations of them are available at the receiver. Thus the original lines can be recovered by linear combinations of the incoming lines.

For example, if A and B represent the content of two adjacent source lines then one channel could carry 3 1 A+B 4 4 and the second channel could carry 1 3 -A±B.

4 4 If these signals are denoted by C, and C2 then A and B would be given by: A=1/2(3C1-C2); B=1/2(3C2-C1) (1) The use of a combination of lines to generate the auxiliary signal has the substantial advantage that, on recovery of the HDTV signal, the noise is evenly distributed between lines, and undesirable subjective differences in the appearance of alternate lines due to different noise contributions are avoided.

The timing of the operation now requires further elaboration. This is because the source and conventional scanning formats are both assumed to be 2:1 interlaced. Thus it follows that each must have an odd number of lines per picture and since both have the same field frequency it follows that the line frequency of the higher definition format cannot be exactly twice that of the conventional format. Specifically, it is assumed that if the conventional format has 2N+1 lines per picture the higher definition format has 4N+1 lines per picture. Then the lines of the conventional format which are generated at the source, according to the spatio-temporal relationships of Figure 1, must be compressed slightly to allow 2N+1 conventional lines to occur in the same interval as 4N+1 source lines.

This timing relationship is shown in Figure 2 for the 9-line source picture shown in Figure 3, the conventional picture having 5 lines. The first two lines of the first field of the conventional picture are synthesised from successive pairs of source lines and, before compression, each lasts for exactly twice the period of a source line. This also applies to the last two lines of the second field. However, the half line at the bottom of the first field and the top of the second field contains inexact information and lasts for only one source line period. This is of no consequence as it can be arranged to occur in the field blanking interval.

Then time-compression is applied so as to shorten all the synthesised lines except for the half line which is expanded. It should be noted that the relative timings shown in Figure 2 are for clarity and do not illustrate a practical situation as they imply negative delay for some lines.

The conventional signal, so generated, has 2N+1 lines occurring at a regular rate of frequency f,, related to the line frequency of the source, fsl by the equation: f,/fs=(2N+1 )/(4N+1 ) (2) However, these lines cannot be used for timing information at the higher definition receiver because the display must use the source line frequency, fsl which cannot be derived simply from f1. This timing information must be derived from the signal in the second channel which must, therefore, have a line frequency f2 given by: f2/fs=1/2 (3) from which fs can be obtained by simple frequency doubling. Thus it follows that the second channel carries 2N+1/2 lines at exactly half the source line frequency.

Figure 4 shows a block diagram of the transmitting end of the system. A (4N+1 )-line signal is received over a line 10 and lines are distributed alternately by a switch 20 to two buffers 40 and 50. These function so as to expand the duration of the lines by a factor of exactly 2 and retime them such that they are cosynchronous. Each such buffer may comprise a line store which can be written into and simultaneously read from at half the rate. Then the timing diagram of the buffers appears as in Figure 5. Alternatively it may comprise two line stores, each being alternately written to and read from. Then the timing diagram appears as in Figure 6.

The time-expanded signals are then fed to a matrix unit 60 which forms the two linear combinations required for the main and auxiliary signals. The unit also receives a feed of field pulses over a line 70 to enable the matrix coefficients to be changed on a field-by-field basis.

The main signal is then fed to a further buffer 80 whose function is to compress the signal by a factor (4N+2)/(4N+1), as discussed above, so as to obtain the correct conventional line frequency.

Finally apprnpriate synchronising pulses and blanking are added to both signals in units 100 and 110 and the outputs are sent over lines 120 and 130.

Figure 7 shows a block diagram of the higher definition receiving end of the system. Essentially, the functions of Figure 4 are repeated in the reverse order. The main and auxiliary signals are received over lines 200 and 210 and the synchronising information is separated from them in separators 220 and 230. The conventional signal is then retimed in buffer 240 so as to expand the lines by a factor (4N+2)/(4N+1) except for the half-line as shown in Figure 2. The synchronised conventional and auxiliary lines are then matrixed in unit 250 which performs the inverse matrixing operation to that in the transmitter, changing its coefficients at field rate.

The matrixed lines are then retimed by buffers 260 and 270 so as to compress them by a factor of exactly two and stagger them in time so as to occur sequentially. A switch 280 then selects them alternately so as to produce a single signal.

Finally synchronising pulses and blanking are added in a unit 290, the signals being derived from a frequency doubler 300 fed with synchronising pulses derived from the auxiliary signal. A standard (4N+ )-line signal then appears over output line 310. The output of frequency doubler 300 is also applied to a squarewave generator 320 which operates the selector switch 280.

The particular signal combinations defined in equations (1) above are not the only combinations which can be used; indeed there may be many other sets of combinations which will be found in practice to be as good or better. For example, an alternative way of deriving the signals would be that, as before: 3 1 C1=-A±B (5) 4 4 but the second video signal is generated in a different way, namely: 3 1 C2=-B-A. (6) 4 4 With the derivations given by equations (5) and (6), the receiver can regenerate the original signals A and B by the following combinations: A=1 .2C1-0.4C2 (7) B=1.2C2+0.4C1 (8) In the methods so far described the 625-line compatible signal in the first video signal (C1) is generated between only two lines of the original HDTV signal.However, it is possible to use higher-order interpolation, for example the first and second channel signals may each be generated from four lines of the original HDTV signal, as follows: C1=a1A+b1B+c1C+d1D (9) C2=a2A+b2B+c2C+d2D (10) where A, B, C and D are successive HDTV lines and a1. . . are the coefficients applied to them.

In order to regenerate A, B, C and D at the receiver, two equations must be solved for four unknowns. This can only be done if the signals on the following two lines of the transmitted signals are used. However, these depend on the original HDTV lines C, D, E and F, and therefore there are now six unknowns to be solved using four equations. In general, therefore, a recursive method, using the results of one line to help solve the following line, is necessary. This will have a severe noise penalty.

However, the recursion can be limited if the coefficients used to generate the second-channel signal are chosen so that: a2/b2=a1/b1 (11) and c2/d2=C1/d1 (12) Now equations (9) and (10) can be expressed as: b, d1 C1=a1(A+ B)+c1(C+ D) (13) a1 c1 b1 d, C2=a2(A+ B)+c2(C+ D) (14) a1 C1 These two equations (13) and (14) can be solved to generate b, (A+ B) a, and d, (C+ D).

c1 After two further source lines b, (C+ D) a, and d, (E+ F) c, can be generated. Then the two equations: d, C+ D=f2 (C1, C2) (15) c1 and b, C+ D=f1(D1,D2) (16) a, can be solved to yield the correct values of C and D. The values of a2 and c2 can be chosen to obtain the best signal-to-noise ratio in the final HDTV picture. In this way noise from line 2 of the compatible first video signal affects only lines C, D, E and F of the reconstructed HDTV picture and has no effect elsewhere.

In general, the system just described is an example of more sophisticated arrangements using a complex matrix operation to regenerate the HDTV signal from the two transmitted signals.

Claims (8)

1. A method of transmitting a high-definition input video signal having 4N+1 lines per picture on two interlaced fields, comprising transmitting a main signal which is compatible with a standard 2N+1/2:1 signal together with an auxiliary signal, in which each output line of the main signal comprises a combination of contributions from two or more lines of the same field of the input signal, and in which each output line of the auxiliary signal comprises a different combination of contributions from two or more lines of the same field of the input signal.

2. A method according to claim 1, in which the combinations are different as between even and odd fields.

3. A method according to claim 2, in which the combinations used to form the main signal are a and 1-a in one field and 1/2-a and 1/2+a respectively in the next field.

4. A method according to claim 2, in which the magnitudes of the combinations used to form the main signal are in the proportions 31 44

5. Apparatus for use in the method of claim 1, comprising an input for receiving a 4N+ 1 lines per picture 2:1 interlaced video signal, and matrix means connected with the input for generating a main signal which is compatible with a standard 2N+1/2:1 signal, and an auxiliary signal, each output line of the main signal comprising a combination of contributions from two or more lines of the same field of the input signal, and each output line of the auxiliary signal comprising a different combination of contributions from two or more lines of the same field of the input signal.

6. A method of regenerating a high definition video signal from a main signal and an auxiliary signal generated by the method of claim 1, comprising inversely combining the lines of the main and auxiliary signals such as to regenerate the 4N+1 lines of each picture.

7. Apparatus for regenerating a high definition video signal from a main signal and an auxiliary video signal, substantially as herein described with reference to Figure 4.

9. Apparatus for regenerating a high definition video signal, substantially as herein described with reference to Figure 7.

signal generated by the method of claim 1, comprising inverse matrix means for combining the lines of the main and auxiliary signals such as to regenerate the 4N+1 lines of each picture.

8. Apparatus for transmitting a high definition