CA1129074A - Television signal coder - Google Patents

Abstract

Abstract of the Disclosure A television signal coder for encoding a television signal using correlations to compress the amount of information transmitted is disclosed.
n fields are selected from every m fields of the signal and a predicted signal level for each picture element in the n fields is represented by an interpolated value of the respective spatially corresponding picture elements in the preceding and succeeding fields. The difference between this inter-polated value and each corresponding picture element, i.e., the prediction error, is coded and transmitted and the signal for the remaining (m-n) fields is coded and transmitted according to any one of the coding systems based on frame-to-frame difference, field-to-field difference and inter-frame difference.

Description

1~,9~)74 The present invention relates to a television signal coder for en- ~
coding a television signal making use of correlations to compress the amount ~ -of information to be transmitted.
An example of a prior art system for the transmission of a tele-vision signal using reduced amounts of information based on the frame-to-frame correlation is shown in an article by J. C. Candy et al entitled "Transmitting Television as Clusters of Frame-to-Frame Differences" published in the "THE BELL SYSTEM TECHNICAL JOURNAL" July-August, 1971, pp 1189-1917.
In this transmission system, the frame-to-frame difference, i.e., the difference in signal level between a picture element of one frame and that of the immediately preceding frame is derived and transmit~ed only when it is significant compared with a predetermined threshold value. Because of the high frame-to-frame correlation inherent in a television picture signal, this system makes it possible to reduce the amount of information to be transmitted significantly. However, when the signal is picked up from a fast changing subject, the frame-to-frame correlation is lowered, resulting in an increase in the amount of information to be transmitted. This may be overcome by quantizing the frame-to-frame differences and making the quan-tizing levels variable depending on the rate of change in the image to be picked up.
However, the broadening of the quantization width results in the degradation of the quality of the reproduced picture with a granular noise appearing over the entire picture. This problem is avoided only by increasing the number of quantizing levels.
Therefore, an object of the present invention is to provide a television signal coder capable of reducing the amount of information to be transmitted without causing degradation of picture quality.
In the television signal coder according to the present invention, _ fields are selected for every _ fields (m, n being positive integers;
_>_~, and a predicted signal level for each picture element in the _ fields *

,9q3~

is represented by an interpolated value of the respective spatially corresponding picture elements in the preceding and succeeding fields. The difference between the interpolated value and each corresponding picture element, i.e. the predic-tion error, is coded and transmitted as a generated information among the _ fields. The signal for the remaining (m-_) fields is coded and transmitted according to any one of the coding systems based on frame-to-frame difference, field-to-field difference and inter-frame difference. ~ith regard to the coding system to be employed, no limitation is imposed.
According to the present invention, for the selected n fields, a prediction value is derived from the preceding and succeeding fields by interpolation. Compared with the conven-tional frame-to-frame prediction based on the signal in the past, the prediction error is reduced and the coding efficiency is enhanced. In addition, the coding system according to the present invention finds a broad practical application because the invention is applicable regardless of whether the input television signal is a monochromatic signal or an NTSC (Natural Television System Committee)color signal.
Broadly stated, the present invention provides a television signal coder for coding a television signal by making use of correlation, said television signal consisting of a series of fields each representing a number of two-dimensionally arranged picture elements of an optical image to be transmitted, said coder comprising: means for selecting n fields of said television signal among every m fields of said television signal, said m and n being positive integers/ said n being less than said m; means for deriving an array of interpolated values corresponding to the interpolation between said picture elements of preceding and succeeding fields adjacent to said selected n fields; first coding means for coding the difference between

-2-.. ..

Z9q~

the television signal level in said selected n fields and said interpolated values; second coding means for coding the tele-vision signal in the (m-n) fields consisting of said m fields excluding said selected n fields; and means for code-converting the outputs of said first and second coding means.
The present invention will now be described in greater detail with reference to the accompanying drawings, in which:
Figure 1 shows the principle of the technique for deriving one field from two adjacent fields by interpolation;
Figure 2 shows field-to-field and line-to-line phase relations of a color subcarrier of a television picture signal for the NTSC system;
Figure 3 is a block diagram of one preferred embodiment of the present invention;
Figure 4 shows a mode of switching between frame-to-frame coding (system-A) and coding employing an average value between two fields (system-B);
Figure 5 shows a first modification of the preferred embodiment of Figure 3; and -2a-: ' 11~9~)74 Figure 6 shows another modification of the first preferred embodi-ment adapted to an NTSC television picture signal.
For simplicity of explanation, in Figure 1 it is assumed that m=2 and n=l, where for every other field an interpolated value is used as the prediction value. As a prediction value for a picture element bjk (indicat-ing that the element is the k-th picture element in the j-th line) in the (i+l)-th field Fi+l, ~ajk + cjk)/2 is used, and after quantization has been executed on the difference bjk-(ajk + cjk)/2 as a prediction error and code-conversion has been achieved, transmission is carried out while adding positional and synchronizing information, for example, to every line. Since the phase of the color sub-carrier for an NTSC television picture signal is reversed for every two successive frames and for every two neighboring horizontal lines as shown in Figure 2, two horizontal lines having their line numbers deviated by 1 in the field Fi and Fi+2, respectively, such as the ~j-l)-th line in the field Fi and the j-th line in the fied Fi+2, the j-th line in the field Fi and the (j+l)-th line in the field Fi+2, ... are employed to derive prediction values for the lines in the field Fi+l where the phase of the respective color sub-carrier is coincident with that on first-mentioned two horizontal lines. Accordingly, a prediction value for bjk is equal to (ajk + c(j+l)k)/2. It is a matter of course that this pre-diction value of ~ajk + c~j+l)k)/2 with regard to the NTSC color signal is equally applicable to a block and white television signal. It is to be noted in Figure 2 that the symbol o indicates that the phase af the color sub-carrier is a while, the symbol indicates that the phase is ~0 + 180).
Now a first preferred embodiment of the present invention assuming m=2 and n=l will be explained with reference to Figure 3.
A digitized television signal is applied to a field memory 102 ~FM2) which is capable of storing at least one field of television signal, and to side-b of a switch 112. The output of the memory 102 is applied to side-a of the switch 112 and to subtractor 107 via lines 212 and 207, re-1~296~74 spectively. Side-a inputs of switches 112, 113 and 114 are selected upon frame-to-frame coding which makes use of a field memory 101 (FMl), whereas side-b inputs thereof are selected upon the coding which makes use of a pre diction value derivecl from two adjacent fields by interpolation.
The output of the switch 112 is applied to a subtractor 103, where the difference between said output and the output of memory 101 fed through a line 1103 is derived. The output of the subtractor 103 is fed to a quantizer 104 to be quantized, and the quantizing characteristics are appro-priately varied in accordance with the command signal supplied from a buffer memory monitor 111 through a line 1114. The output of the quantizer 104 is applied to side-a of the switch 114 and to an adder 105 via lines 414 and 405, respectively. In the switch 114, either the signal applied to the side-b after the quantization of the output of the subtractor 107 by the quantizer 108 or the signal applied to the side-a through the line 414 is selected depending upon which one of the frame-to-frame coding and the coding employing interpolated values between two fields is executed.
The adder 105 produces a locally decoded signal in the frame-to-frame coding by taking the sum of the signals applied through lines 405 and 1105, and the output is applied to the side-a of the switch 113 and to an interpolator circuit 106 through lines 513 and 506, respectively. In the switch 113, the side-a input is selected for the frame-to-frame coding, while the side-b input is selected for the coding employing interpolated values between two fields. The output of the switch 113 is fed to the memory 101. In the interpolator 106, interpolated values are calculated from the signals of two fields fed through lines 1106 and 506, respectively, and the result is applied to the subtractor 107.
The description of the switching between the frame-to-frame coding ~hereinafter referred to as system-A) and the coding employing interpolated values between two fields ~hereinafter referred to as system-B) will now be given with reference to Figure 4.

9~

When the input signal to the memory 102 is for the i-th field Fi (hereinafter abbreviated simply as Fi), the signal at a point Y in Figure 3 is a signal Fi which is derived by executing the coding of system-A between the output Fî 2 of the memory 101 (the symbol '~ ' representing a signal processed and thereafter locally decoded) and Fi and thereafter locally decoding the same, so that the interpolator circuit 106 is supplied with Fi and Fi 2. At this point in time, i.e., when the system-B coding is to be executed, all the switches 112, 113 and 114 select the side-b. Subsequent-ly, when the input signal is turned to Fi+l, the output of the memory 102 is Fi, then the system-A coding is executed between this output Fi and the output Fi 2 of the memory 101, and the switch 113 is controlled so that the locally decoded signal ~i at this point in time may be applied to the memory 101. In other words, upon execution of the system-A coding, all the switches 112, 113 and 114 select the side-a. Thereafter, the same operation is re-peated for the input signals Fi+2, Fi+3, ... .
Referring again to Figure 3, the output of the switch 114 is fed to a code converter lO9 for coding the generated information. The output of the code-converter 109 is fed to a buffer memory 110 for the purpose of transmission speed matching with a transmission path. The state of the buffer memory 110 is monitored by the buffer memory monitor 111, and the quantization characteristics of the quantizers 104 and 108 are modified in accordance with the state of the buffer memory. The modification command is fed to the respective quantizers 104 and 108 through lines 1114 and 1108, respectively. In addition, information representing quantization character-istics then used is fed to the code-converter 109 through line 1109. It is to be noted that the quantization characteristics of the quantizers 104 and 108 may be of the same type. The construction of the coder may be simplified accordingly. More particularly, as shown in Figure 5, the outputs of the subtractors 103 and 107 are applied to the side-a and side-b, respectively, of the switch 114. The output of the switch 114 is applied to the quantizer 11~9~

104, where output is fed to the adder 105 and the code-converter 109. By such modification, the quantizer 108 in Figure 3 can be omitted.
Referring again to Figure 3, a switching signal for selecting either the frame-to-frame coding or the coding making use of interpolated values between two fields in the switches 112, 113 and 114, is fed from a switching signal generator 115 to these switches through lines 1512, 1513 and 1514, respectively. At the same time, the same signal is fed to the code-converter 109 via a line 1509 to designate whether the used coding is of system-A or of system-B.
10In the above description of the first preferred embodiment, frame-to-frame coding is employed for the coding of the (_-_) fields. However, inter-frame coding can be used also.
Now description will be made of a second preferred embodiment of the present invention corresponding to the case where m=4 and _=2. With re&ard to the coding in which frame-to-frame coding and coding making use of interpolated values between adjacent two frames are executed alternately -in each of the successive frames, the above description is also applicable by merely substituting the term "frame" or "frame memory" for the term "field" or "field memory" in the description of the first preferred embodi-ment and in Figure 3~ and by substituting the term "between-every-second-frame" for the term "frame-to-frame" in Figure 4.
Next, description will be made of a third preferred embodiment of the present invention corresponding to the case where the input signal in the first preferred embodiment is an NTSC color signal. In this case, since the phase of the color signal subcarrier is reversed with respect to two neighbouring lines and between successive frames as shown in Figure 2, the j-th line in Fi, and the (j+l)-th line in Fi+2 are employed in the coding of system B upon calculating interpolated values for the j-th line in Fi+l.
In Pigure 2, the markings o and represent the lines on which the phase of the color signal sub-carrier is 0 and ~0~180 ), respectively. More particu-larly, it is only necessary to use the j-th line in Fi and the ~j+l)-th line in Fi+2, and so, in this case, the addition of one circuit element as shown in Figure 6 is effected to the circuit shown in Figure 3. More par-ticularly, at the left of the point X in Figure 3 and Figure 6 there is additionally provided a one-line delay circuit 200. With regard to the coding of system-A according to Figure 2, since the (j-l)-th line in Fi which has been delayed by the delay circuit 200 at the point X corresponds to the j-th line in Fi+2, the phase of the color signal sub-carrier on these lines coincide with each other, whereby the frame-to-frame coding can be executed. The same applies to the subsequent lines.

. . ~ .,

Claims (6)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A television signal coder for coding a television signal by making use of correlation, said television signal consisting of a series of fields each representing a number of two-dimensionally arranged picture elements of an optical image to be transmitted, said coder comprising: means for selecting n fields of said television signal among every m fields of said television signal, said m and n being positive integers, said n being less than said m; means for deriving an array of inter-polated values corresponding to the interpolation between said picture elements of preceding and succeeding fields adjacent to said selected n fields; first coding means for coding the difference between the television signal level in said selected n fields and said interpolated values; second coding means for coding the television signal in the (m-n) fields consisting of said m fields excluding said selected n fields; and means for code-converting the outputs of said first and second coding means.

2. The television signal coder recited in claim 1, further comprising first and second field memories each capable of storing at least one field of television signal, said first field memory storing in succession said selected n fields and said second field memory storing in succession at least said (m-n) fields and decoding means connected between the output of said second coding means and the input of said first field memory for periodically updating the content of said first field memory, said means for deriving interpolated values receiving as inputs the outputs of said first field memory and said decoding means, and said first coding means receiving as inputs the outputs of said second field memory and said means for deriving interpolated values.

3. The television signal coder recited in claim 2, where-in said first and second coding means comprise first and second subtractors, respectively, and wherein said selecting means includes a first switch for selectively applying either the input or the output of said second field memory to one input of said second subtractor, the other input of said second sub-tractor being supplied by the output of said first field memory.

4. The television signal coder recited in claim 3, where-in said first and second coding means further comprise first and second quantizers, respectively, and wherein said selecting means further includes a second switch for selectively coupling the outputs of either of said first or second quantizers to said code-converting means.

5. The television signal coder recited in claim 3, wherein said first and second coding means further comprise a common quantizer and wherein said means for selecting further includes a second switch for selectively coupling the outputs of either of said first or second subtractors to said common quantizer.

6. The television signal coder recited in claim 2, further comprising delay means connected between the output of said first field memory and one input of said means for deriving an interpolated value, said delay means providing a one line delay.