WO1990006038A1 - High definition b-mac television signal transmission system - Google Patents

Abstract

Apparatus for encoding a high definition multiplexed analog components television signal by decimating predetermined samples for transmission comprises separate processing paths (Figs. 17 and 18) for luminance and chrominance component signals respectively. The separately processed luminance and chrominance components signals are multiplexed into a composite signal for transmission via a skew-symmetric filter portion (1710) shared by both processing paths. A complementary skew-symmetric filter portion (1901) to the filter portion of the encoder apparatus is associated with a decoder coupled to a high definition television receiver. Predetermined samples are diagonally filtered and horizontal resolution information folded into the signal. Vertical resolution is improved through scan conversion at the receiver.

Description

- i -

HIGH DEFINITION B-MAC TELEVISION SIGNAL TRANSMISSION SYSTEM

BACKGROUND OF THE INVENTION

1. Technical Field

The invention relates to the field of television signal transmis¬ sion systems and, in particular, to a television signal transmission system for transmitting a signal providing a higher resolution image than is transmitted under standard resolution National Television Sub¬ committee (NTSC) or European formats.

2. Description of the Relevant Art

There is a growing interest in the transmission of television signals which increase picture definition in both the horizontal and vertical dimensions. In the vertical dimension, such a signal may have as many as twice the number of lines in comparison with exist¬ ing standards while in the horizontal dimension, the number of picture elements per line is likewise increased. As a result of providing stan¬ dard horizontal and vertical resolution, there are adverse effects rom providing wide screen displays of a transmitted signal. A viewer of a standard signal may complain of the fuzzy or unclear quality of the image if viewed from a relatively close proximity. These adverse effects are overcome by a higher resolution image but existent trans¬ mission systems are not readily adaptable to transmitting a high reso¬ lution image.

One solution to the problem of providing a high resolution image without increasing the required bandwidth for signal transmis¬ sion is described in U.S. patent application Serial No. 092,305 filed September 2, 1987. A high resolution signal is diagonally filtered and alternate samples decimated on alternate lines according to digital sampling techniques to leave a figure-of-five or quincunx pattern of picture samples. Odd line samples are added to even line samples forming a line summation signal. By means of skew-symmetric filter¬ ing techniques, high resolution horizontal information regions of the signal which normally carry diagonal information of marginal value. The loss of diagonal information from a transmitted signal does not cause perceptible impairment to the original high resolution image input at the transmitter. Only video line stores are required at the transmitter. No field stores are required. Furthermore, the digital filtering required at a receiver is relatively inexpensive in comparison with prior art interpolation techniques.

Another solution to the problem of transmitting a high resolu¬ tion image is to transmit a standard television signal and to create and transmit a so-called augmentation channel. In accordance with the first solution described above, a new receiver is required for pro¬ cessing the received signal of standard bandwidth. In accordance with this solution, no change in receiver circuitry is required for receiving and displaying a standard resolution image. However, sepa¬ rate adapter circuitry is required for receiving the augmentation channel containing high resolution data and for reinstituting the high resolution data into the standard resolution image to provide a high resolution image.

A method and apparatus for increasing the definition of an NTSC video signal using an augmentation channel is described in U.S. patent application Serial No. 228,274 filed August 4, 1988. According to the approach taken in that application both a line summation signal and a line difference signal are formed from a high resolution televi¬ sion signal. By reverse alternate sampling of the line summation sig¬ nal at mid video signal bandwidth, high resolution horizontal informa¬ tion may be translated to baseband and added to high resolution verti¬ cal information from the line difference signal and transmitted together as the required augmentation channel.

Neither of these approaches, however, provides a solution to the problem of transmitting a high definition multiplexed analog com¬ ponents (MAC) television signal with increased horizontal and vertical resolution as compared with standard resolution television signals but which does not require modification of existing MAC receiver/decod¬ ers for receiving a standard resolution MAC composite signal.

During the 1970's sub-nyquist sampling techniques were applied in the art of digital television by the present inventor and colleagues at the Independent Broadcast Authority of the United Kingdom. Digi¬ tal techniques for the elimination of aliasing are described in an arti¬ cle entitled "An Introduction to Sub-Nyquist Sampling" by K. H. Barratt and the present inventor appearing in the LB. A. Technical Review at pages 3-15. In a companion article at pages 21-26, entitled "Digital Sub-Nyquist Filters" by J. H. Taylor, comb filters are described for down conversion and up-conversion of video data appli¬ cable in a PAL television signal environment for optimizing .signal sampling. These anti-aliasing digital sampling techniques provide a developmental basis for improving horizontal resolution; however, there still remains a requirement in the art for a method and appara¬ tus for improving both horizontal and vertical resolution of a trans¬ mitted multiplexed analog components television signal. SUMMARY OF THE INVENTION The present invention relates to a method and apparatus for transmitting and receiving a high definition multiplexed analog com¬ ponents (MAC) television signal. According to the B-type MAC trans¬ mission format for composite signal transmission, the video signal is carried within an active line period while all other signals comprising at least audio, control data, utility data and teletext are transmitted during a line blanking period or a longer field blanking period. Sepa¬ rate luminance and chrominance signals are digitally sampled, com¬ pressed and transmitted during separate portions of a video line sig¬ nal. In accordance with the well-known B-MAC format, luminance samples are compressed for transmission at a ratio of 3:2 while chrominance is compressed at a ratio of 3:1. Chrominance informa¬ tion is translated into U and V components, each component being transmitted every other line.

To accomplish a transmission of high resolution video informa¬ tion within the boundaries of multiplexed analog component formats generally and in accordance with the method of the present invention a folding of high horizontal resolution information is accomplished into the high frequency diagonal components of the sampled video signal. At baseband frequencies below 5 MHz (7 MHz when time- compressed 3:2 according to the format), the spectrum is unmodified. Thus, for example, since standard B-MAC decoders are typically equipped with 6.3 MHz passband lowpass filters at their input, the folded high resolution information does not affect reception. The additional transmitted information at high frequencies is simply blocked and ignored.

In particular, according to the present transmission method, a high definition analog television signal is first orthogonally sampled at 28 MHz (a rate of eight times the color subcarrier of 3.58 MHz or 8 Fsc). As a result a two dimensional sample spectrum is achieved which is then passed through a diagonal digital filter which decreases the diagonal frequency response but which decrease is practically imperceptible to a viewer.

The diagonally filtered data is then decimated by_discarding alternate samples on alternate lines. As a result, a figure-of-five or quincunx pattern of samples remains. As a result of the decimation of alternate samples, the baseband spectrum remains unchanged but high resolution repeat spectrums comprising horizontal and vertical resolution components exist at half the sampling frequency and at the sampling frequency. The repeat spectrums serve to fold additional resolution into the baseband signal.

The samples of the folded signal may then be converted to ana¬ log form and passed through a low pass skew-symmetric filter cen¬ tered at seven megahertz or similarly digitally filtered. Accordingly, high resolution information related to the horizontal dimension is folded about a diagonal axis at seven megahertz into the approxi¬ mately five to seven megahertz or high frequency portion of the passed baseband signal. Effectively, the high resolution information is traded for the diagonal information.

According to transmission apparatus of the present invention, the digital diagonal filter of the encoder may comprise separable hori¬ zontal and vertical filters. The vertical filter at the transmitter may be very simple provided the horizontal filter is sufficiently complex to achieve a 40db rejection in the stop band. For example, the hori¬ zontal filter at the transmitter (permitting a 0-5 MHz passband) may be at a complexity on the level of sixteen coefficients. Besides per¬ mitting a much simpler vertical filter at the transmitter to achieve a diagonal filter, it has also been found that as a result of employing a complex horizontal filter at the transmitter, a much less expensive and simpler diagonal filtering arrangement may be employed at the receiver. In particular, a 5-MHz low-pass filter at the receiver need comprise only eight coefficients. While the invention is described in terms of improving luminance horizontal detail, the technique and apparatus may be adapted for improving chrominance horizontal detail.

An embodiment for improving chrominance horizontal detail applies similar principles to those applied for improving luminance horizontal detail. However, an intentional increase in horizontal chrominance detail has been found to be unnecessary for a B-MAC signal, allowing a simpler technique for chrominance transmission.

A vertical filter interpolator receives at its input a 525 line, 1:1 non-interlaced signal, a 1050 line, 2:1 interlaced signal or an 1125 line 2:1 interlaced signal. The signal is appropriately filtered in a vertical direction and an output provided to a 4:1 line decimation cir¬ cuit. The output of the line decimation circuit is filtered in a hori¬ zontal dimension about a center frequency of 2.5 megahertz and the high frequency output samples provided to a multiplexer. The multi¬ plexer combines the chrominance, luminance and any data/audio sig¬ nals for transmission. However, prior to transmission, the combined high definition B-MAC signal may be passed through a skew-symmet¬ ric filter portion centered at 10.7 megahertz which in combination with a complimentary filter portion at a decoder eliminates aliasing in a decoded signal. Consequently, the same skew-symmetric filter portion of the encoder may be shared for both luminance and chrominance processing. BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is a graphical depiction of vertical versus horizontal definition for a high definition television signal.

Figure 2 is a representation of an orthogonal sampling grid for sampling the signal of Figure 1 at 28.6 megahertz (8 Fsc after prefiltering at 9.0 megahertz).

Figure 3 is graphical depiction of vertical versus horizontal definition as a result of application of the orthogonal sampling grid of Figure 2 such that a baseband spectrum results as well as a repeat spectrum centered at the sampling frequency.

Figure 4 is a graphical depiction of vertical versus horizontal definition with diagonal information, a block of data, for example, between five and nine megahertz, having been filtered from the orthogonally sampled spectrums of Figure 3.

Figure 5 is a representation of a sampling grid at 14 MHz gen¬ erated by discarding alternate samples on alternate lines to achieve a figure-of-five or quincunx sample pattern.

Figure 6 is a graphical depiction of the result of the decimation of alternate samples where, besides the repeat spectrum at 28 MHz, two repeat spectrums at 14 MHz, half the initial sampling rate are introduced.

Figure 7a is a graphical depiction of vertical versus horizontal resolution for showing the process of filtering about a center fre¬ quency of approximately 7 MHz, the filter having a skew-symmetric low-pass response and Figure 7b the characteristic amplitude versus frequency response.

Figure 8 is a graphical depiction of a first step of processes accomplished at a receiver. By applying a resampling at fourteen megahertz using the figure-of-five pattern shown in Figure 5, the folded high resolution horizontal dimension information is returned to high frequency. Figure 8a represents a first graphical depiction of vertical versus horizontal definition and Figure 8b represents a second graphical depiction of amplitude versus frequency showing how aliasing is eliminated and information close to 7 MHz is regenerated by the skew-symmetric characteristic of the filter. Figure 9 is a graphical depiction of vertical versus horizontal definition showing the result of upconverting rom 14 megahertz to 28 megahertz. Upon diagonal filtering, a null for high diagonal frequen¬ cies in the range of 5-9 megahertz is produced leaving a signal having high horizontal resolution but an imperceptible sacrifice in diagonal information.

Figure 10 is a schematic diagram of apparatus of a transmitter for encoding a high definition B-MAC television signal.

Figure 11 is a schematic diagram of apparatus of a receiver for decoding a high definition B-MAC television signal.

Figure 12 is a graphical representation of the characteristic response of the sixteen coefficient horizontal low-pass filter, includ¬ ing coefficient data, shown in Figure 10.

Figure 13 is a graphical representation of the characteristic response of the eight coefficient horizontal low-pass filter, including coefficient data, shown in Figure 11.

Figure 14 is a graphical representation of vertical resolution in lines per picture height versus horizontal resolution in lines per pic¬ ture width showing characteristics of the application of the present invention in combination with conventional but proprietary scan con¬ version line doubling techniques in B-MAC versus results of more expensive multiple field store techniques used in a 1125 line MUSE signal transmission system.

Figure 15 is a graphical representation of amplitude versus frequency and horizontal versus vertical resolution for the three fil¬ ters applied in the present technique: diagonal filtering, - pref iltering and skew-symmetric filtering, Figure 15a being of amplitude versus frequency and Figure 15b being of vertical versus horizontal resolution.

Figure 16 is a graphical depiction of the two dimensional con¬ tour response of the transmitted signal comparable to Figure 14, hori¬ zontal resolution being traded for diagonal resolution in a system according to the present invention. Figure 17 is a block schematic diagram of luminance processing circuits at the location of a high definition B-MAC encoder in accor¬ dance with the present invention.

Figure 18 is a block schematic diagram of chrominance pro¬ cessing circuits at the location of a high definition B-MAC encoder in accordance with the present invention.

Figure 19 is a block schematic diagram of processing at the location of a high definition B-MAC decoder in accordance with the present invention.

Figure 20 is a block schematic diagram of a high definition television receiver for processing the output of the high definition B-MAC decoder of Figure 19.

Figure 21 is a block schematic diagram of a high definition television receiver having multiple applications in a high definition or standard resolution television environment.

DETAILED DESCRIPTION OF THE DRAWING

Referring to Figure 1 there is shown a high definition televi¬ sion signal graphically depicted in terms of vertical versus horizontal resolution.

According to Figure 1, the present method is assumed applica¬ ble to a high definition 16:9 aspect-ratio picture scanned sequentially using 525 lines, the horizontal resolution being at least 945 lines at 9 megahertz. A sequential scan signal of this type supports a vertical definition of 480 lines.

With this signal as an input, this description refers by way of example to a new type of B-MAC signal which carries increased reso¬ lution at the imperceptible expense of diagonal resolution. However, it may be likewise applied in other multiplexed analog component video signal transmission systems. The increased resolution is folded into the high video frequencies. At baseband frequencies below 5 MHz (7 MHz when time-compressed in MAC) the spectrum is unmodified. As known B-MAC decoders have low-pass input filters having a pass band limited at 6.3 MHz, decoder operation is unaf¬ fected by the additional transmitted information. On the other hand, a B-MAC decoder in accordance with the present invention retrieves and decodes the folded horizontal detail information and causes a high resolution image to be displayed by a receiver.

A standard resolution 525-line 2:1 interlace video signal con¬ sists of two fields each containing 240 active lines. Lines of every other (odd) field are spatially offset relative to lines of even fields so that all 480 active lines are regularly spaced on the display screen.

In principle, this line structure can carry a vertical resolution equal to 480 lines for static pictures. However, a normal interlaced display does not achieve this value. The reason for this lies in the fact that only 240 lines are displayed in each field, and the human eye/brain is expected to sum the two fields and perceive all 480 lines. It cannot do this perfectly. The intensity of the first field perceived by the eye/brain has decreased to 50% of its initial value by the time that the second field arrives (1/60 seconds later). This has two consequences:

(i) Line structure becomes visible in the display.

(ii) Vertical frequencies exceeding 240 lines are partially aliased in the display.

The net result is that the perceived vertical resolution of a standard resolution 525 line interlaced display lies somewhere between 480 lines and 240 lines. The reduction from 480 lines is con¬ ventionally described by introducing the concept of a so-called "Kell Factor":

Perceived Resolution = 480 x 0.66 (Kell Factor)

= 320 Lines

For static pictures, Kell Factor may be entirely eliminated and resolution restored to 480 lines by displaying all 480 lines (from both odd and even fields) in each 1/60 second field period. This technique is known as scan conversion. Application of the technique involves use of a field store memory storing all 240 active lines to move infor¬ mation between fields, and a display at twice the normal line fre¬ quency. However, the method can be directly applied only to static parts of the picture, since significant motion can occur between fields. Consequently, a motion detector is also required so that inter- field interpolation can be used for stationary objects, while line-interpolation is used for moving objects. Sean conversion tech¬ niques employing adaptive field-store line doubling techniques thus achieve 480 lines of vertical resolution for static pictures and approx¬ imately 320 lines in moving dynamic areas of an image.

Field store line-doubling is gaining acceptance as a standard method for increasing vertical definition by TV-set manufacturers. Its main advantage is that it eliminates line structure and signifi¬ cantly improves picture quality without requiring any additional information transmitted. Several manufacturers of TV-sets and pro¬ jectors are using proprietary line-doubling techniques, including Philips, Hitachi, Sony, Ikegami, etc. Its obvious advantages are:

(i) The technique applies equally to component signals (luminance, chrominance) or NTSC signals received from any source including S-VHS VCRs. (ii) The technique is applied in the television receiver and therefore permits retention of a 525 interlace connec¬ tion to the TV-set (NTSC or wideband Y/C). (iii) The field-store in the TV set image projector can be used for other consumer features such as picture-in-pic- ture and noise reduction, (iv) The technique requires no additional transmitted infor¬ mation and allows any high definition television (HDTV) format to concentrate on the problem of increasing hor¬ izontal definition. In U.S. patent application Serial No. 255,328, filed October 11, 1988 of Christopher Birch entitled "Method and Apparatus for Improv¬ ing Vertical Definition of a Television Signal by Scan Conversion", there is described a technique for improving the vertical resolution of a television signal by developing a plurality of alternative interpolated values for display and selectively choosing a particular value in accor¬ dance with tests of the video signal for shading, movement and verti¬ cal edge transitions.

In order to solve the problem of increasing horizontal resolu¬ tion, the present high definition MAC system employs sub-Nyquist sampling (spectrum olding) to trade diagonal resolution for increased horizontal resolution. The process will be described in connection with Figures 2-9 and the apparatus at a transmitter or receiver will be described in connection with Figures 10-11. All frequencies quoted (bandwidths and sampling frequencies) will be referred to the uncompressed luminance signal of a B-MAC signal. Equivalent band- widths and sample frequencies in the time-compressed (MAC) domain must be increased by a actor of 1.5 (3:2).

The 525-line 2:1 interlaced luminance signal is first bandlimited to 9 MHz using a low-pass analog filter. Referring briefly to Fig. 10, the filter is described as an 8-7 MHz pre-filter. This bandwidth (BW) is sufficiently broad to achieve a horizontal resolution of 945 lines per picture width (PW), calculated as follows: ines/PW -^SBW x Active Line x 2 = 945 Lines/PW

Total Line Line Freq

This signal is initially sampled at 28.6 MHz (8 Fsc) using an orthogonal sampling grid as shown in Figure 2. As a result of orthogo¬ nal sampling and in accordance with Figure 3, a baseband spectrum as well as a repeat spectrum centered at the sampling frequency results. The baseband spectrum comprises high horizontal resolution compo¬ nents at as high as 9 MHz or 945 lines as calculated above.

A diagonal digital filter is then applied which decreases the diagonal frequency response (Step 2). Separable horizontal and verti¬ cal filters are employed for simplicity as will be further described in connection with the discussion of Figures 10 and 11. Referring to Figure 4, it may be seen that blocks of diagonal (horizontal versus vertical) information are removed at horizontal frequencies between five and nine megahertz in the baseband spectrum as well as the repeat spectrum.

Now, alternate samples are discarded leaving a "quincunx" (figure-of-five) sample pattern at approximately 14 MHz (4 fsc) with¬ out causing aliasing. According to Figure 5, alternate samples on alternate lines are removed.

The result of Step 3 is a sequence of digital samples which now have the compacted 2-dimensional spectrum shown in Figure 6 wherein repeat spectrums exist at the fourteen megahertz sampling frequency of the alternate sample decimation step.

Before transmission, the horizontal resolution improvement information is folded into the signal for transmission. For example, the samples may be converted to analog form and passed through a transmission filter with specific characteristics. The analog trans¬ mission filter may have skew-symmetric low-pass response which is -6dB at 7MHz. The result is shown in Figure 7a where high resolution information at 7-9 megahertz is translated to fill the void formed from diagonal filtering. Alternatively, a digital non-recursive filter may be applied as will be further described herein alleviating a requirement for digital to analog conversion or upconversion to 28 megahertz sampling.

Effectively, horizontal resolution between 7 MHz and 9 MHz has been folded around 7 MHz and replaces diagonal resolution between 5 MHz and 7 MHz. According to Figure 7b, the filter charac¬ teristic response is shown at 6dB attenuation at 7 MHz. Note that in multiplexed analog component (MAC) transmission, the transmission filter will be skew-symmetric about 1.5 x 7 MHz due to the time com¬ pression factor (3:2). The signal is now ready for transmission, which may be considered step 5.

When the signal is received, it is resampled at 14 MHz using the alternate line quincunx (figure-of-five) sampling pattern (step 6). By resampling at fourteen megahertz, a translation of the high frequency information occurs to 7-9 MHz according to Figure 8a.

The resampling process thus regenerates the horizontal energy between 7 MHz and 9 MHz. Referring to Figure 8b, losses in the transmission filter around 7 MHz are also precisely compensated by the alias term A, provided the transmitting filter in combination with the decoder process has a skew-symmetric response: A + (1-A) = 1. In other words, the alias term introduced is necessarily cancelled no matter where the amplitude A is measured in the overlapping dashed line, solid line area proximate to 7 MHz.

The spectrum of Figure 8 applies to the 14 MHz sequence of quincunx samples. What may remain is to upconvert to 28 MHz, introduce a 2-dimensional filter to remove the remaining energy on the diagonal, and to bandlimit the signal to 9 MHz.

A diagonal filter (which produces a null for high vertical fre¬ quencies in the range of 5-9 MHz) cannot be implemented at the 14 MHz sample rate. Therefore, the up-conversion of sample rate to 28 MHz occurs as a part of the digital filtering process. According to Figure 9, the upconversion and diagonal filtering results in a null at diagonal frequencies and improved horizontal resolution.

Step 7 results in a sequence of samples at 28 MHz with an orthogonal sampling grid. They carry a spectrum with horizontal res¬ olution of 9 MHz, with no aliasing. These samples are available for direct conversion to analog form. After bandlimiting to 9 MHz, the analog signal may be displayed on a high definition receiver.

Alternatively, the samples may be passed directly to a known but proprietary field-store scan converter for line-doubling to increase the vertical definition such as the apparatus described by U.S. application Serial No. 255,238, entitled "Method and Apparatus for Improving Vertical Definition of a Television Signal by Scan Con¬ version" of Christopher Birch filed October 11, 1988 and incorporated herein by reference. The result of the scan conversion is a 525 sequential-scan signal sampled at 56 MHz (2 x 28 MHz) and carrying horizontal luminance resolution up to 18 MHz. (The line doubling pro¬ cess halves the active line period and doubles both the sampling fre¬ quency and the video bandwidth).

The described process provides a very cost effective imple¬ mentation in a multiplexed analog component signal transmission sys¬ tems as will be described in connection with block schematic dia¬ grams Figures 10-11 of the encoder and the decoder respectively of the present apparatus.

Transmitter apparatus according to Figure 10 performs all steps of the above described process but for transmission: step 1, 28 MHz orthogonal sampling after application of an 8.7 MHz pre-filter; step 2, digital diagonal filtering; step 3, 28 MHz - 14 MHz alternate sample decimation; and step 4, digital to analog conversion and skew-symmetric trans¬ mission filtering.

Although it may appear that the diagonal filter must be imple¬ mented at a sample rate of 28 MHz, the fact that samples are to be discarded at step 3 allows a simplification. Figure 10 shows a configu¬ ration in which the main elements of the digital filter can be imple¬ mented at a 14 MHz sample rate.

The diagonal filter is implemented as two separable (horizontal and vertical) filters. It has been found that the vertical filter can be very simple i.e., a line store 101 and adder 102 provided the horizontal filter 103 achieves approximately a 40 dB rejection in the stop band. To accomplish such a rejection, the horizontal filter 103 (5 MHz low- pass) employs 16 coefficients at 28 MHz. The design of the horizontal filter 103 is a symmetrical non-recursive filter which has been opti¬ mized with 9-bit coefficient values. The response and the coeffi¬ cients are presented in Figure 12.

According to Fig. 10, switch 100 switches alternate samples of a pre-filtered 28 MHz sampling signal into two 14 MHz paths. According to the upper path, the samples are vertically filtered and horizontally filtered. The lower path is substrated from the upper path at adder 102 while the upper path is added to the lower path at adder 104. At the output of filter 103 is shown a low pass combing characteristic with zero energy at zero frequency while at the output of adder 104 is shown a low frequency combing characteristic (solid line) with energy at zero frequency extending to 9 MHz and a high frequency aliased characteristic (dotted line) extending from 5 MHz up. The result at the output of adder 105 is a signal for transmission with horizontal resolution folded into the 5-7 MHz region. For each al, bl sample pair, one sample remains from quincunx sampling on alternate lines, separated by 14 MHz as shown. Consequently, a 28 megahertz process is accomplished at 14 MHz because of the alter¬ nate sample decimation.

Figure 11 shows the decoder, in which it is also possible to implement the digital filter at 14 MHz. In this case, the 5 MHz low- pass filter ill contains only 8 coefficients, the characteristic response and coefficient data are shown in Figure 13.

According to Fig. 11, a 14 MHz input signal is sampled accord¬ ing to the figure-of-five sampling pattern shown at 14 MHz. The receiver decoder further comprises adder 112 and line store 113 in an upper path. The lower path further comprises interpolation circuit 114 including single element delay D and averager (divide-by-two) circuits for restoring missing data to the sample pattern. The upper and lower paths are switchably upconverted at 28 MHz and added at adder 115.

Figure 14 shows the idealized two-dimensional frequency response achieved by the system in comparison with the known high definition 1125-line MUSE system developed by Japan Broadcasting Corp. (NHK). The MUSE system involves a plurality of field stores and thus is considerably more expensive to implement than the present invention including scan conversion apparatus involving one field store. Nevertheless, for dynamic or static images, the horizon¬ tal resolution is either equivalent or clearly superior to the MUSE system according to the present invention and with scan conversion is _almost comparable in vertical definition.

Figure 15 shows the actual response achieved using the filters which have been described. According to Fig. 15a, the position of the folded energy (hatched area) is also shown in one-dimension. In either Fig. 15a or 15b, A relates to the encoder diagonal filter, B to the pre-filter and C to the skew-symmetric filter. The skew-symmetric transmission filter may be implemented in analog form, according to the above-described method. According to the above-described appa¬ ratus, a symmetrical non-recursive filter may be used alternately which produces an ideal linear phase characteristic. If a digital filter is used to create the skew-symmetric response, it is unnecessary to up-convert to 28 MHz sampling. The required digital filter will have zeros in all alternate coefficients except for the central term and so leads to a 14 MHz implementation. (A 28 MHz implementation is not required when alternate samples are automatically dropped with alternate zero coefficients. The central term is dealt with separately). Figure 15 assumes use of a 16 coefficient non-recursive filter.

Figure 16 shows the actual 2-dimensional response achieved by the system. From Figure 16 in comparison with Fig. 14 shows improved horizontal resolution out to 945 (950) lines at the cost of diagonal information. In accordance with the above-described embod¬ iment and method, luminance horizontal resolution is improved and with scan conversion vertical resolution improved as well. The present invention may be adapted to provide improved chrominance horizontal resolution recognizing the 3:1 compression of U/V chrominance data in alternate transmitted lines. Scan conversion has -already been adapted for obtaining chrominance vertical resolution improvement. However, in B-MAC, to intentionally improve chrominance resolution may not be required.

While detailed schematic diagram of a high definition B-MAC transmitter and receiver has been described in reference to Figures 10 and 11 respectively, luminance and chrominance processing will be further described with reference to Figures 17-19.

Referring to Figure 17, there is shown a block diagram of cir¬ cuitry for luminance processing at an encoder location which works in concert with circuitry for chrominance processing at the same encoder location. According to Figures 17 and 18, the separate pro¬ cessing paths share the same skew-symmetric filter portion 1710.

A 525 line 1:1 non-sequential scan, a 1050 line 2:1 interlaced or a 1125 line 2:1 interlaced video luminance signal is provided to a ver¬ tical filter/interpolator circuit 1701 in accordance with the invention resulting in a 480 line static vertical resolution. As already described, alternate lines are decimated at deeimator 1702 to result in a 2:1 interlaced 525 line signal and provided to a diagonal filter 1703. The 28 megahertz sampling output of the diagonal filter is then provided to a decimation circuit 1704 for decimating alternate samples on alternate lines leaving a quincunx sample structure at a 14 megahertz sampling rate. The resulting samples are stored in memory and read out at a different rate effectuating a 3:2 sample compression in the time domain at time compression circuit 1705. The time compressed samples are then mixed with chrominance and any data/audio signals for transmission at multiplexer 1706. Skew-symmetric filter portion 1710 then is shared by the processed luminance and chrominance signals.

Referring to Figure 18, the chrominance processing is similar to that for luminance processing but recognizes that U/V component signals are transmitted every other line, and chrominance is normally compressed, in accordance with the B-MAC transmission format, at a ratio of 3:1. Consequently, a chrominance input of the same number of lines in either interlaced or. non-interlaced format as the luminance signal is provided to a vertical filter/interpolator 1801 to achieve a 120 line dynamic or 180 line static chrominance resolution in picture height. In the horizontal dimension, the achieved chrominance reso¬ lution can be up to 300 lines of picture width. Because every other line is processed, a line decimation circuit 1802 for U and V compo¬ nents is operated at a 4:1 ratio. The output of the line decimation circuit is provided to a horizontal filter 1803 centered at 2.5 mega¬ hertz. High frequency color samples output from the horizontal filter are then time compressed at 3:1 according to B-MAC transmission standards at time compression circuit 1804, the color signals are then mixed for transmission, that is, U and V components are mixed with luminance signals at multiplexer 1706 and transmitted every other line as in standard B-MAC with audio or data channels also input to multiplexer 1706. Again, however, it is indicated in Figure 18 that the same skew-symmetric filter portion 1710 may be applied to the combined signal prior to transmission.

Referring now to Figure 19, one embodiment for processing of a high definition B-MAC signal comprising both luminance and chrominance information is shown. In accordance with the present invention, a complimentary skew-symmetric filter portion 1901 at 10.7 megahertz receives the HMB-MAC signal. Consequently, aliasing of the composite signal is eliminated. The output of the complimen¬ tary skew-symmetric filter portion 1901 is sampled at a 21 megahertz rate and is provided to a luminance processing path and a chrominance processing path. The result is a figure of five or quincunx sample structure. Luminance samples are time-expanded 3:2 at time expansion circuit 1902. The output is then provided to a diagonal filter/interpolator stage 1903 which provides a digital output with missing alternate samples replaced in alternate lines.

According to the chrominance path, chrominance samples are time expanded by 3:1 at time expansion circuit 1904 and the time- expanded samples at 7 megahertz are provided to a vertical filter/ interpolator 1905 for providing useable R-Y and B-Y chrominance outputs. These may be modulated at modulator 1906 and provided to a shared digital to analog converter 1907 for providing luminance and chrominance outputs Y and C or separately output to high definition television receiver apparatus.

A high definition television receiver is shown in block diagram form according to Figure 20. Luminance samples 4 output from the decoder of Figure 19 are input to the high definition receiver along with separate R-Y and B-Y inputs. According to the block diagram, there is provided a scan conversion apparatus 2001 comprising in combination a field store 201, a motion detector 2003 and an interpo¬ lating algorithm processor 2004. The intent is to provide line doubling in the vertical dimension as expeditiously as possible. Scan conver¬ sion apparatus may be employed as disclosed in application Serial No. 255,238, filed October 11, 1988 and incorporated herein by refer¬ ence. As a result of the interpolation process which requires a field store and associated delay, another luminance signal path comprises delay and compression circuits 2005, 2006 for 2:1 compression which is mixed with the scan converter output via compression circuit 2007 at multiplexer 2008 for display.

Chrominance information is interpolated in the vertical dimen¬ sion at vertical interpolator 2010 compressed 4:1 at compression cir¬ cuit 2011 and mixed with the chrominance input signal for display at multiplexer 2012.

Referring to Figure 21, an overall end user location is shown in pictorial form. A high definition B-MAC signal 2101 may be received and displayed alternatively to a video cassette recoder output 2102 or a land-line cable input 2103.. A high definition television receiver is shown which, responsive to a user-selected input entered via selector 2104, applies, for example, the already described adaptive scan con¬ version methods for line doubling at scan converter 2105 and so dis¬ plays a high definition 16:9 aspect ratio image on the receiver. Also, an of -the-air broadcast television signal 2108 may be received by antenna, tuned at tuner 2109, decoded if necessary at NTSC decoder 2110 and provided via the same scan conversion apparatus 2105 for display. A terrestrial high definition decoder 2111 is required if the broadcast signal 2108 is high definition. Consequently, a user equipped with a high definition television receiver is able to improve resolution of a displayed image regardless of whether the received signal is high definition, standard NTSC low resolution, or encoded in high or low resolution multiplexed analog components format.

Claims

1. Apparatus for encoding a high definition multiplexed analog component television signal for transmission characterized by luminance processing means for processing a television luminance component signal by decimating predetermined samples for transmission, chrominance processing means for processing a televi¬ sion chrominance component signal by decimating predetermined samples for transmission, multiplexing means for combining the processed lumi¬ nance and chrominance component signals for transmission, and a skew-symmetric filter portion, responsive to the mul¬ tiplexing means and coupled to a complimentary filter portion, for iltering the multiplexed signal for transmission.

2. Apparatus for encoding a high definition television sig¬ nal for transmission according to claim 1, the luminance processing means comprising a vertical filter interpolation circuit for filtering an incoming video signal, a line decimation circuit, responsive to the vertical fil¬ ter interpolation circuit, for decimating alternate lines, a diagonal filter, responsive to the line decimation cir¬ cuit, for providing a quincunx video sample structure, and a time compression circuit, responsive to the diagonal filter, for compressing video signal samples for transmission.

3. Apparatus for encoding a high definition television sig¬ nal for transmission according to claim 1, the chrominance processing means comprising a vertical filter interpolation circuit for filtering an incoming video signal, a line decimation circuit, responsive to the vertical fil¬ ter interpolation circuit, for decimating each of U and V components every other line, a horizontal filter, responsive to the line decimation circuit, for filtering at a color center frequency, and a time compression circuit, responsive to the horizontal filter for compressing high frequency color samples at a ratio of 3:1 for transmission.

4. Apparatus for encoding a high definition multiplexed analog components television signal for transmission characterized by luminance processing means for processing a television luminance component signal by decimating predetermined samples for transmission, chrominance processing means for processing a televi¬ sion chrominance component signal by decimating predetermined samples for transmission, and a skew-symmetric filter portion, responsive to the lumi¬ nance processing means and the chrominance processing means and coupled to a complimentary skew-symmetric filter portion, for filter¬ ing the high definition multiplexed analog component signal for transmission.

5. Apparatus for decoding a high definition multiplexed analog component television signal upon reception, predetermined television signal samples having been decimated prior to transmission, the apparatus characterized by: a skew-symmetric filter portion, coupled to a compli¬ mentary skew-symmetric filter portion, for filtering the received television signal, sampling means, responsive to the skew-symmetric fil¬ ter portion, for sampling the filtered signal, luminance processing means for processing the sampled signal by interpolating decimated luminance samples and chrominance processing means for processing the sam¬ pled signal by interpolating missing chrominance samples.

6. Apparatus according to claim 5, the luminance process¬ ing means comprising: a time expansion circuit and a diagonal filter and interpolator for restoring deci¬ mated luminance samples.

7. Apparatus according to claim 5, the chrominance pro¬ cessing means comprising: a time expansion circuit and a vertical ilter and interpolator for resorting decimated chrominance samples.

8. Apparatus coupled to a high definition multiplexed ana¬ log component television signal decoder for processing luminance and chrominance component signals for display, the apparatus character¬ ized by: a luminance processing path comprising a motion detec¬ tion and interpolation circuit for interpolating luminance samples determined from motion between first and second fields and a chrominance processing path comprising a vertical interpolation circuit for interpolating chrominance samples.

9. Processing apparatus according to claim 8, the motion detection and interpolation circuit comprising: a field store for sorting a first video ield and an interpolator responsive to the motion detector for interpolating samples rom the stored video field and the second video field.

10. Processing apparatus according to claim 8 further com¬ prising first and second compression circuits for compressing a delayed luminance input signal and the output of the motion detection and interpolation circuit respectively and a multiplexer for multiplexing respective outputs of the first and second compression circuits.