DD298577A5 - DEVICE FOR RECEIVING A TV CHANNEL - Google Patents

Abstract

The invention relates to a device for receiving a television signal for a large screen with Randteilbildinforamtionen and dominant main picture information, wherein the television picture has a greater picture-to-page ratio. According to the invention, means are provided which initially separate the edge subframe information and the dominant main image information; each of which generates an output signal from the separated edge and main picture information, and finally combines the output signals to form an image display signal {color television signal; Grosz screen; Picture-aspect ratio; Edge field information; dominant main picture information; Fernsehsignalempfaenger; combined image signal}

Description

Video signal processing means (1344) responsive to the first and second components for generating an image display signal.

3. Apparatus according to claim 2 for receiving the television signal, further comprising a processed during the image change third component with auxiliary image information, characterized in that the subcarrier is phase-controlled and modulated by the second and third components; the demodulating means (1326) demodulates the separate subcarrier to provide the second and third components; and

in that the video signal processing means (1344) is responsive to the first, second and third components for generating the bijdication signal.

For this 27 pages drawings

The present invention relates to a television signal reception apparatus which embodies a large screen image having side area image information and dominant non-area area image information and has a larger aspect ratio than a conventional television image.

A conventional color television receiver, such as a receiver tuned to the NTSC standard broadcast in the US and elsewhere, has an aspect ratio (the ratio of the width to the height of the displayed picture) of 4: 3. Recently, however, there has been an interest in higher aspect ratios for television receiver systems, such as the 2: 1.16: S or 5: 3 ratios, as such higher aspect ratios are closer to or equal to the visual field of the human eye than the 4: 3 aspect ratio Aspect ratio of a conventional television receiver. The video information signals with an aspect ratio of 5: 3 have received special attention because this ratio approaches that of a motion picture film, and such signals can then be transmitted and received without clipping the picture information. However, widescreen television systems that simply send signals with a higher aspect ratio compared to conventional systems are incompatible with receivers that have the conventional aspect ratio. This makes a widespread introduction of widescreen systems difficult. It is therefore desirable to have a widescreen system that is compatible with conventional television receivers. Such a system is disclosed in U.S. Patent 4,782,383 to MA Isnardi entitled "Apparatus for processing high frequency edge information in a widescreen television system" published November 1, 1988. There is an even greater desire After that, in such a compatible widescreen system, measures are taken to improve or enhance the resolution of the reproduced image to provide additional image detail. For example, an extended definition widescreen EDTV system may include a progressive-wave device A system of this kind is disclosed in a paper by MA Isnardi et al., entitled "Encodeing for compatibility and recovery in the ACTV syst 'η", and in IEEE Transactions on Broadcasting, Vol , No. 4, December 1987, pages 116 to 123. The system uses signal time compression and intraframe processing of the image information of the center and edge portions. The invention is therefore based on the object of developing a verbossm ies device for receiving a color television signal for a wide-screen image with edge sub-picture information and body image information.

According to the invention the object is achieved in that

Means for separating the time-averaged side-view image information and the dominant non-side-area image information; first signal processing means for generating a second output signal in response to the separated non-side area image information; and

Means are provided for combining the first and second output signals to produce an image display signal.

Further advantageous features of the invention will become apparent from the dependent claims.

The inventive solution allows for increased uniformity of spatial resolution between the reproduced center and margin information by substantially reducing or eliminating unwanted diagonal image artifacts in the rendered margin information.

The invention is disclosed in the context of a compatible widescreen EDTV television system employing time compression and intraframe signal processing, for example the averaging technique. The wide-screen EDTV signal includes a plurality of components, including a first main component that includes the middle portion and the time-compressed edge portion information, and an additional second component that contains the edge portion information. In the main component, only the sub-part information is subjected to intra-frame processing. The portion of the time compressed edge portion information of the main component is not subjected to intraframe processing. In a disclosed preferred embodiment of a compatible widescreen EDTV television system in accordance with the principles of the present invention, an original high resolution, progressively scanned widescreen signal is encoded to include four components. The four components are processed separately before being reunited in a single signal transmission channel.

A first component is a 2: 1 interlaced main signal with a standard aspect ratio of 4: 3. This component consists of a middle portion of the wide-screen signal, which is time stretched to occupy nearly the entire active line duration of the 4: 3 aspect ratio, and of horizontal low-frequency edge portion information, which is temporally compressed in the left and right horizontal frame oversampling areas, such information being hidden from the block in a standard television receiver display. Only the middle part of this component is subjected to an intraframe message above a given frequency.

A second component is an additional signal in the 2: 1 interlaced format, which comprises the high-frequency information of the left and right edge portions, which are each temporally stretched up to half of the active line duration. In this way, the stretched edge part information occupies substantially the entire active line duration. This component is "mapped" to occupy the same time period as the middle portion of the first component, and is subjected to an intraframe message.

A third component is a 2: 1 interlaced additional signal derived from the wide-screen signal source, comprising the high-frequency horizontal luminance detail information between approximately 5.0MHz and 6.0MHz. This component is also mapped to occupy the same time period as the middle part of the first component, and it is subjected to an intraframe message. The intraframe-averaged second and third components undergo a quadrature modulation phase-controlled alternating subcarrier which is combined with the intraframe-averaged first component.

An arbitrary fourth component is a 2: 1 interlaced additional helper signal that includes a temporal field difference of luminance detail information to help recover the missing image information in a widescreen EDTV receiver.

In a widescreen EDTV receiver, a composite signal consisting of the described four components is decoded into the four subcomponents. The decoded components are separately processed and used to develop a widescreen signal representing an image with increased resolution.

embodiments

Figure 1 provides a general overview of a compatible widescreen EDTV coding system including a device according to the present invention;

Figure 1 a shows a detailed block diagram of the coder for the disclosed system; Figures 1 b through 1 e contain useful diagrams for understanding the operation of the disclosed system; Figures 2 to 5 show signal waveforms and diagrams useful in understanding the operation of the disclosed system;

Figure 13 is a block diagram of a portion of the decoding apparatus of the widescreen EDTV receiver, and Figures 6 through 12 and 14 through 24 illustrate aspects of the disclosed system in detail.

A system designed for that. Images with a large aspect ratio of, for example, 5: 3 over a standard, e.g. B. NTSC broadcast channel, on the one hand should achieve high quality image reproduction on a widescreen receiver and on the other hand greatly reduce visible noise or eliminate completely when the image is played back with a standard aspect ratio of 4: 3. The application of the signal compression technique to the fringes of an image has the advantage of the horizontal oversampling range of a standard NTSC furnace receiver; however, this can be at the expense of image resolution in the edge portions of a reconstructed wide-screen image. Since compression in time results in a frequency-domain stretch, when processed in a standard television channel, only low-frequency components will be left which have a smaller bandwidth compared to that required for a wide-screen signal. Therefore, when the compressed edge portions of a compatible wide-screen signal are stretched in a wide-screen receiver, there is a noticeable difference between the resolution or high-frequency portion of the center portion of a reproduced wide-screen image and the edge portions, unless steps are taken to avoid this effect. This significant difference stems from the fact that low-frequency edge information is recovered while high-frequency information is lost due to the band-limiting effects of the video channel.

In the arrangement of Figure 1, the elements which correspond to those of the detailed arrangement of Figure 1 a, indicated by the same reference numerals. As shown in Fig. 1, an original wide-screen progressive scan signal having left, right and center part information is processed to develop four separate encoding components. These four components have been described above and are shown pictorially in FIG. The processing of the first component (including the time-expanded center portion information and the time-compressed low-frequency edge portion information) is performed such that in this example the resulting luminance bandwidth does not exceed the NTSC luminance bandwidth of 4.2 MHz. This signal is color coded in standard NTSC format, and the luminance and chrominance components of this signal are appropriately prefiltered (eg, using field comb filters) to provide improved luminance / chrominance separation in both the standard NTSC format and NTSC format also in the widescreen receivers to ensure. The temporal extension of the second component (high-frequency information of the edge portions) reduces its horizontal bandwidth to about 1.16 MHz. This component is spatially uncorrelated with the main signal (the first component), and thus special security measures are taken to prevent its visibility on the standard NTSC receivers, as will be discussed.

The extended high-frequency portion of the luminance information of the third component ranging from 5.0 to 6.0 MHz is first shifted downwardly in a frequency range of 0 to 1.0 MHz, before further processing takes place. The fourth component (temporal field-delay helper signal) is mapped into 4 ^ standard format to correlate with the main signal component, thereby obscuring its visibility on standard NTSCC receivers, and is horizontally limited in bandwidth to 750kHz ,

As will be explained in more detail below, the first, second, and third components are processed by respective infra-frame averaging devices 38, 64 and 76 (a type of vertical-temporal (Vt) filter) to provide vertical-temporal crosstalk between the main and auxiliary signal components on a widescreen receiver. The sub-part information of the first component is subjected to an intra-frame message above approximately 1.5 MHz. The second and third intraframe-averaged components, labeled X and Z, undergo non-linear amplitude compression in a block 80 prior to quadrature modulation of an alternate 3.108 MHz subcarrier ASC with a periodically alternating (inverting) field phase. In an adder stage 40, a modulated signal (M) from block 80 is added to the intraframe-averaged first component (N). An output resulting therefrom is a 4.2 MHz bandwidth baseband signal (NTSCF) which, together with a 750kHz fourth pass component (YTN) from a filter 79, quadrature modulates an RF picture carrier in block 57 to form an NTSC to produce a compatible RF signal that can be transmitted to a standard NTSC or progressive scan widescreen receiver over a single standard bandwidth broadcast channel.

The application of temporal compression to the first component allows the low-frequency edge information to be completely injected into the horizontal oversampling area of a standard NTSC signal. The high-frequency edge part information of the second component and the high-frequency luminance dot information of the third component are spectrally divided with the standard NTSC signal through the video transmission channel in a manner transparent to a standard receiver, by applying a quadrature modulation method to an alternating one Subcarrier in block 80, as will be explained. When received with a standard NTSC receiver, only the middle portion of the main signal (the first component) is visible. The second and third components may cause a low amplitude interference pattern, but this is not noticeable at normal viewing distances and in normal settings of the image controls.

The fourth component is completely suppressed in receivers with synchronous video detectors. In receivers with envelope rectifiers, the fourth component is processed but not perceived because it is correlated to the main signal.

The main signal (component 1) has an active standard NTSC horizontal line interval of approximately 52 microseconds (\ is) . Only the high frequency information of this component is subjected to intraframe averaging above about 1.5 MHz. The temporally compressed low-frequency edge part information of this component is not subjected to the process of intra-frame averaging. It has been found that such selective intraframe processing of the main component improves the resolution of the diagonal edge portion image information by eliminating unwanted jagged diagonal artifacts which are sometimes attributed to "bursts" which would otherwise be produced in a reconstructed image if the compressed edge portion information of the main signal has undergone intraframe averaging.

In this regard, it should be noted that the low frequency edge portion information of the main signal component has been temporally compressed with an edge compression factor (SCF) of approximately six. If such time-compressed information is subjected to intra-frame averaging before being time-stretched in the image-retrieving receiver, the retrieved edge-portion image information would have jagged diagonals because that at horizontal frequency at which intra-frame averaging was started. then approximately SCF times lower than for the middle part. The diagonal image information is increasingly "jagged" as the frequencies at which intraframe averaging is performed, for example, when the main signal is subjected to intraframe averaging at frequencies above 1.5 MHz and the low-frequency edge component information from component 1 With a SCF factor of six, the intraframe averaging of edge information effectively begins at a much lower frequency of 250KHz (1.5MHz / SCF), resulting in jagged diagonals, so jagged diagonals in the recovered edge portions would become more noticeable Since the component 1 does not undergo intraframe averaging in the temporally compressed edge portions, the entire range of the original frequency in these ranges (0 to 700KHz) retains the full vertical resolution without jamming due to jagged diagonal artifacts left and right high fre is shown to occupy the same time period as the center portion of component 1. In this way, the left and right edge sub-levels are time-stretched to execute the entire center sub-area, whereby the component 2 has an active horizontal scanning interval of approximately 50ps, which corresponds to the horizontal scanning interval of the center portion of the component 1. For this purpose, the edge strain factor (SEF) is about 4.32, compared to an edge strain factor of about

4.49, which would be required to stretch the left and right edge part information of component 2 to the full active line duration of 52 \ is .

Both components 2 and 3 are mapped into the middle portion because of the intraframe processing performed on component 1 and on components 2 and 3. As will be explained below, the intraframe averaging is a process which facilitates the separation of the two previously combined signal components, in this case for example the main signal N and the modulated additional signal M. Since the range of intraframe processing in component 1 has been reduced, to include only the mid-portion of 50με, the mapping of the modulation components 2 and 3 is similarly modified to include only the mid-portion. As mentioned above, the component 3 is mapped to coincide with the mid-portion interval by linear temporal compression of the extended horizontal luminance information to 50 \ is . The time compression of component 3 from 52 ps to 50 I is dispensed with part of the spatial correlation with the main component 1, but it can be determined with certainty that the middle and edge portions of the reconstructed image have a similar horizontal resolution. Although the spatial correlation between components 1 and 3 is desirable to mask the effects of cross talk between the alternate subcarrier and the main signal, the importance of maintaining perfect spatial correlation of component 3 is reduced because the alternating subcarrier already contains uncorrelated information in Form of component 2 contains. The measure of the spatial correlation given in component 3 is insignificant and is enhanced by the resulting similar horizontal resolution of the center and edge portions. Component 4 is not intraframe averaged and remains unchanged, having the full active line duration of 52 ms, consistent with the main signal. In the decoder described in conjunction with Fig. 13, the intraframe processing is performed only with respect to the center portion for separating the signals M and N. After demodulating the component M into the subcomponents 2 and 3, the components become 2 and 3 are mapped into their original time slots, ie, to assume a complete active line interval of 52μβ.

Figure 1 b illustrates the RF spectrum of the disclosed EDTV wide-screen system including the additional information compared to the RF spectrum of a standard NTSC system. In the spectrum of the disclosed system, the edge sub-levels and the additional high-frequency horizontal luminance detail information extend approximately 1.16 MHz on each side of the alternate subcarrier frequency (ASC) of 3.108 MHz. The V-T helper signal information (component 4) extends 75 okHz on each side of the picture carrier frequency of the main signal.

A progressive scan widescreen receiver includes means for restoring the original progressive widescreen strobe signal. Compared to a standard NTSC signal, the reconstituted widescreen signal has left and right edge portions with a standard NTSC resolution and a center portion with a 4: 3 frame ratio with higher horizontal and vertical luminance detail, especially in fixed areas of an image. Two fundamental considerations dominate the signal processing technique associated with the development and processing of the first, second, third and fourth signal components. Both of these considerations are compatibility with existing receivers and recoverability in the receiver.

Full compatibility also means compatibility of the receiver and transmitter so that existing standard receivers can receive widescreen EDTV signals and produce standard reproduction without special adaptation measures. For example, compatibility in this sense requires that the image sampling format in the transmitter be substantially the same as the image sampling format in the receiver, or at least within the tolerance of the receiver-side image sampling format. Compatibility also means that special components that are standard components are hidden physically or in terms of perceptibility in the main signal when played back using a standard receiver. In order to achieve compatibility in the latter sense, the disclosed system uses the following methods to hide the additional components.

As already described above, the edge part depths are physically hidden in the normal horizontal oversampling area of a standard receiver. The component 2, which contains little energy compared to the component of the edge part depths, and the component 3, which is normally a high-energy low-energy detail signal, are subjected to amplitude compression and quadrature modulation to an alternating subcarrier at 3.108 MHz, which is an interlaced scanning frequency (odd-numbered) Multiple of half the horizontal line frequency). The frequency, phase and amplitude of the alternating hill carrier are chosen so that the visibility of the modulated alternating subcarrier signal is reduced as much as possible, for example by controlling the phase of the alternating subcarrier from field to field so that every 180 ° from one field to the other the next changes, unlike the phase of the chroma support from one field to the next. Although the components of the modulated alternating subcarrier lie entirely within the chrominance passband (2.0 to 4.2 MHz), the components of the modulated alternating subcarrier are noticeably hidden because they are reproduced as complementary field-flicker color fuzz, which is from the human eye is not perceived at normal levels of chroma saturation. In addition, nonlinear amplitude compression of the modulation components prior to amplitude modulation advantageously reduces instantaneous amplitude overshoots to an acceptably lower level. The component 3 is spatially correlated with respect to the information of the center region of the component 1, and it is spatially correlated with respect to the left and right information components of the component 1 spatially. This is done with the help of a picture format encoder, as will be explained.

Component 4, the "helper" signal, is also hidden due to the temporal extension of the mid-level information to ensure conformance to the 4: 3 standard format, thereby spatially correlating the fourth component with the main signal Component 4 becomes standard receivers is suppressed with synchronous detectors while hidden in standard receivers with envelope detectors due to their spatial correlation with the main perception signal.

The recovery of components 1,2 and 3 in a progressive scan widescreen receiver is enabled by the application of a process of intraframe processing in the transmitter and receiver. This process is in the transmitter system of Figures 1 and 1 a by the associated elements 38,64 and 76 and assigned in the receiver by

Elements carried out, as will be explained. Intraframe averaging is a signal handling technique in which two visually correlated signals for mutual combining are pretreated so that they are later recognized quickly and accurately, for example, by means of a field memory device, such that they are free of vertical-sound crosstalk, as well then, when it comes to image signals and movement is present in the images. The type of signal handling used for this purpose essentially involves the action of making two signals identical on a field basis, i. generate two samples with identical values one frame apart. Intraframe averaging is a suitable technique for achieving this goal, but other techniques may be used. Intraframe averaging is basically a process of linear time-varying digital prefiltering and postfiltering to ensure the accurate recovery of the two visually correlated combined signals. Horizontal crosstalk is eliminated by guard bands between horizontal prefilters in the transmitter-side encoder and between horizontal post-filters in the receiver-side encoder.

The intraframe message is a form of pairwise pixel processing. The process of intraframe averaging in the time plane is illustrated generally in Figure 1c, where pairs of fields are made identical by dividing the mean of the picture elements (A, B and C, D) which are 262 horizontal periods (262 H) apart. is formed. The mean scores the original values in each pairwise group. Figure 1d illustrates the process of intraframe averaging associated with the system of Figure 1. Beginning with components 2 and 3, pixel pairs (pixels) which are within 262H of each other within a frame are averaged, and the mean (e.g. X1, X3 and Z1, Z3) replaces the original pixel values. This vertical-temporal mixing occurs within a frame and does not exceed the frame boundaries. In the case of the component 1, the intraframe averaging is performed on the information of the center part above approximately 1.5 MHz so as not to affect the vertical detail information of low frequency. In the case of components 1 and 2, intraframe averaging is performed on a composite signal including the luminance (y) and chrominance (c) components throughout the chrominance band. The chrominance component of the composite signal survives intraframe averaging because the pixels spaced apart by 262H are "in-phase" with the color subcarrier. The phase of the new alternating subcarrier is controlled to be accurate for 262H discrete pixels Therefore, if components 2 and 3 are added to component 1 after quadrature modulation in unit 40, the pixels separated by 262 H have the form (M -I-A) and ( M - A), where M is a sampling pattern of the composite main signal above 1.5 MHz and A is a sampling pattern of the modulated additional signal.

Intraframe averaging substantially eliminates vertical-temporal crosstalk, even with motion in the picture. In this regard, the process of intraframe averaging produces identical samples spaced 262H apart. In the receiver, it is a simple matter to accurately recover the information content of these samples by processing the 262H apart pixel samples within a frame, as it still does is explained; thereby the information of the main and auxiliary signal is recovered. In a decoder in the receiver, the original information which has undergone intraframe averaging can be recovered virtually intact via the process of intraframe processing, since the well correlated original information has been made substantially identical from field to field.

Also in the receiver, the RF channel is quadrature demodulated using an RF synchronous detector. The component 4 is thereby separated from the other three components. The intraframe processing is used to separate the component 1 from the modulated components 2 and 3, and using quadrature demodulation, the components 2 and 3 are separated from each other, as will be discussed with reference to FIG. After the four components have been recovered, the composite signals are decoded according to the NTSC standard and separated into the luminance and chrominance components. An inverse mapping of all components is performed to restore the wide aspect ratio, and the high frequency portions of the edge portion information are combined with the low frequency portions to recover the complete resolution in the edge portions. The enhanced high-frequency luminance detail information is shifted to its original frequency range and added to the luminance signal, which is then converted to the progressive-scan format using temporal interpolation and the helper signal. The chrominance signal is converted to the progressive scan format using temporal interpolation without additional assistance. Finally, the progressive scan format luminance and chrominance signals are converted to the analog form and matrices to produce the R, G, and B color image signals for display on a progressive scan widescreen display device.

Before the compatible widescreen coding system according to FIG. 1a is described in more detail, the signal forms A and B of FIG. 2 are considered. The signal A is a wide-screen signal corresponding to the aspect ratio 5: 3 which is to be converted to an NTSC standard compatible signal for a 4: 3 aspect ratio signaled by signal B. The wide-screen signal A includes a central portion designated primary image information occupying an interval TC and left and right marginal portions having secondary image information occupying the intervals TS. In this example, the left and right edge portions have substantially equal aspect ratios that are smaller than those of the dominant center portion that adjusts therebetween.

The wide picture signal A is converted into the NTSC signal B by completely) compressing a certain edge part information into the horizontal oversampling areas with the time intervals TO. The standard NTSC signal has an active line interval TA (with a duration of 52.6 microseconds), which includes oversampling intervals TO, a playback time interval TD containing the video information to be displayed, and a total horizontal line duration interval TH having a duration of 63.556 Includes microseconds. The intervals TA and TH are the same intervals for both the widescreen and the standard NTSC signals. It has been found that almost all consumer television receivers have an oversampling interval occupying at least 4% of the total active line time TA, i.e., at least 4% of the total active line time TA. H. 2% oversampling on the left and right margins. At an interlaced scan rate of 4 χ fsc (where fsc is the frequency of the color subcarrier), each horizontal line interval contains 910 pixels (pixels), of which 754 form the active horizontal line image information to be rendered.

The bridge image EDTV system is shown in more detail in FIG. 1a. In the following, reference is made to FIG. 1 a. A progressive scan 525-line and 60-field-per-second widescreen camera 10 produces a widescreen color signal with components R, G and B, and in this example a 5: 3 widescreen aspect ratio. An interlaced scanning signal source could also be used, but a progressive scan signal source provides better results. A widescreen camera has a larger aspect ratio and a larger video bandwidth compared to a standard NTSC camera, where the video bandwidth of a widescreen camera is proportional to the product of its aspect ratio and the total number of lines per frame, among other factors. At the same scanning speed of the wide-screen camera, increasing the aspect ratio results in a corresponding increase in its video bandwidth, as well as greater horizontal compression of the image information when the signal is reproduced on a standard 4: 3 aspect ratio television receiver. For these reasons, it is necessary to modify the wide screen signal if it is to be fully NTSC compatible. The color video signal processed by the encoding system of Figure 1 contains both luminance and chrominance signal component tone. The luminance and chrominance signals contain both low and high frequency information, which in the following description will be referred to as "lows" and "highs". The progressive scan widescreen color video signals having a wide bandwidth and coming from the camera 10 are matrixed in a unit 12 to derive the luminance component Y and the color difference signal components I and Q from the color signals R, G and B. The progressive scan broadband signals are sampled at eight times the frequency (8 * fsc) of the chrominance subcarrier and individually converted from analog to digital (binary) form by separate analog-to-digital converters (A / D) in an AD unit 14 before they are filtered individually by means of separate vertical-temporal low-pass filters in a filter unit 16 to produce filtered signals YF, IF and QF. These signals each have the shape as indicated by the waveform A in FIG. The separate filters are linear time-invariant 3x3 filters of the type shown in Figure 10 and discussed below. These filters easily reduce the vertical-temporal resolution, particularly the diagonal vertical-temporal resolution, by unwanted interlaced scanning artifacts (e.g., flickers, jagged edges, and other pincushion-related phenomena) in the main signal (component 1 in Figure 1) after conversion of the progressive ones To prevent sampling in the interlace scan. The filters preserve almost the full vertical resolution in still parts of the image.

The center stretch factor (CEF) is a function of the difference between the width of an image rendered by a wide-screen receiver and the width of an image rendered by a standard receiver. The image width of a widescreen display with an aspect ratio of 5: 3 is 1.25 times larger than the image width of a standard display with an aspect ratio of 4: 3. This factor of 1.25 is a preliminary factor for the midsection stretch that must be adjusted to accommodate the oversampling range of a standard receiver and to account for intentional slight overlap of the area boundaries between the center and edge portions, as will be discussed , These considerations dictate a CEF of 1.19.

The progressive scan signals from the filtering network 16 have a bandwidth of 0 to 14.32 MHz and are respectively converted to 2: 1 interlaced format signals by means of progressive-interlaced (PZ) converters 17 a, 17 b and 17c, the details of which will be discussed in connection with FIGS. 22 and 23. The bandwidth of the output signals IF ', QF' and YF 'from the PZ converters 17a to 17c is in a range of 0 to 7.16 MHz because the horizontal scanning frequency for interlaced scanning signals is one-half of the signals of the progressive scan , In the conversion process, the progressive scan signal is subsampled by taking one-half of the available pixel scan patterns to generate the main signal in 2: 1 interlaced format.

Specifically, each progressive scan signal is converted to the 2: 1 interlaced format by maintaining either the odd or even lines in each field, and reading out the retained pixels at a frequency of 4 x fsc (14.32 MHz). All subsequent digital processing of the interlaced scan signals is at a frequency of 4 χ fsc.

The network 17c also contains an error prediction network. An output of network 17c, YF ', is the interlaced luminance version of the pre-filtered progressive scan component. Another output (luminance) signal of the network 17c, YT, contains the temporal information derived from the field difference information representing a temporal prediction or interpolation error between actual and predicted values of the luminance sampling patterns that are present in the receiver. " The prediction is based on a time averaging of the amplitudes of the "before" and "after" picture elements available in the receiver.The signal YT, a luminance "helper" signal, carries In principle, it helps to recover the progressive scan signal in the receiver and, in principle, accounts for an error which the receiver can expect to make with respect to non-stationary image signals, and facilitates the cancellation of such error in the receiver. In the still parts of an image the error is zero, and in the receiver the full recovery is done. It has been found that a chrominance helper signal is not practically necessary and that a luminance helper signal is sufficient to achieve good results since the human eye is less sensitive to missing vertical or temporal chrominance details. Figure 2 a illustrates the algorithm used ι ~ development of a yeast signal YT. In the following, reference is made to FIG. 2a. The picture elements A, X and B in the progressive scan signal occupy the same spatial position in an image. The black pixels, such as A and B, are transmitted as the main signal and are available in the receiver. A white picture element, for example X, is not transmitted and predicted by a frame averaging time (A + B) / 2. That is, in the encoder, a prediction is made for "missing" picture elements X by averaging the amplitudes of the "before" and "after" picture elements A and B. The prediction value (A + B) / 2 becomes the actual value X is subtracted to produce a prediction error signal corresponding to the helper signal having an amplitude according to the expression X - (A + B) / 2 This term identifies the temporal field difference information other than the time all-time averaging information The helper signal is output by means of a 750 The band-limiting of the helper signal to 750 kHz is necessary to keep this signal from interfering with the next lower RF channel after this signal modulates onto the RF picture carrier In the receiver, using an average of the

Sampling patterns A and B provide a similar prediction of the missing picture element X, and the prediction error is added to the prediction. That is, X is retrieved by adding the prediction error X - (A + B) / 2 to the time average (A + B) / 2. In this way, the helper signal facilitates the interlaced conversion to the progressive scan format.

The helper signal, which is advantageously generated by the disclosed temporal prediction algorithm, is a low energy content signal compared to a predictive signal generated by other algorithms, such as the algorithm for generating a line differential signal generated by M.Tsinberg in one Article "ENTSC Two-Chani.el Compatible HDTV System", published in IEEE Transtetions on Consumer Electronics, Vol. CE-33, No.3, August 1987, pages 146 to 153. In still areas of an image, the error energy is zero because the prediction is perfect, a low-energy state is manifested by static and virtually static images (eg a news bulletin with a reporter against a static background).

It has been found that the disclosed algorithm causes the least disturbing artifacts after image reconstruction in the receiver, and that the helper signal generated by this disclosed algorithm retains its usefulness after being band limited (filtered) to about 75 oKHz. The helper signal generated by the disclosed algorithm advantageously has zero energy at still image information, and as a result, a helper signal associated with a still image is not affected by filtering.

Even if the helper signal is not transmitted, the reconstruction of the widescreen is much better. In such a case, still parts of the image appear much sharper than in an NTSC standard image, however, moving parts become somewhat "softer" and may contain a "beating" artifact. Thus, a broadcasting station does not need to transmit the Holfersignal from the beginning, but may choose to improve the RF transmission at a later time. The disclosed temporal prediction system is useful for both progressive scanning and interlaced scanning at higher than the standard line frequencies, but it works best with a progressive scan source in which pixels A, X and B occupy the same spatial position in one picture, resulting in a perfect prediction for still pictures. However, the temporal prediction will be imperfect even in still areas of an image when the original widescreen comes from an interlaced scanning signal source. In such a case, the helper signal will have more energy and will introduce small artifacts in still parts of a reconstructed image. Although experiments have shown that the use of an interlaced signal source gives acceptable results with artifacts that are only noticeable upon closer inspection, a source of progressive scan signals will bring less artifacts and better results.

In the arrangement of Figure 1 a, the interlaced image signals IF ', QF' Uiid YF 'present in the interlace format are filtered by the PZ converters 17a to 17c respectively in an associated horizontal Tiefpaßfiiter 19a or 19b and 19c to a signal IF to produce a 0 to 60 oKHz bandwidth, a 0 to 60 oKHz bandwidth QF signal, and a 0 to 5 megahertz bandwidth YF signal, which are then subjected to an image encoding process, each one of them This is done by means of an image format encoder which includes an edge center portion separator and processing unit 18. Briefly, the center area of each wide line is time expanded and into the displayed portion of the active line duration with an aspect ratio of 4: 3. The time stretching causes a reduction of the bandwidth, so that the interlaced frequencies of the original Br video signal is made compatible with the bandwidth of the NTSC standard. The edge portions are split into horizontal frequency bands so that the component containing the high frequency components ("highs") of the color of the I and Q signals has a bandwidth of 83kHz to 60kHz (as shown for the signal IH in Figure 7) ) and the component containing the "heights" of the luminance signal Y has a bandwidth of 70OkHz to 5.0MHz (as shown for signal YH in Fig. 6). The low-frequency components ("lows") of the edge portions, ie, the signals YO, IO, and QO developed according to Figures 6 and 7, contain a DC component and are time-compressed into the left and right horizontal oversampling areas of each row Details of this aspect ratio encoding process are discussed below: For the consideration of the encoding details described below, it will also be helpful to consider Figure 1e which illustrates the process of encoding the components 1,2,3 and The filtered interlaced signals IF ", QF" and YF "are preprocessed by the edge portion / center portion separator and processor 18 to produce three sets of output signals: YE, IE and QE; YO. IO and QO and YH, IH and QH). The first two sets of signals (YE, IE, QE, and YO, IO, QO) are processed into a signal containing a full bandwidth midsection component and the luminance depths of the edge portions pressed into the horizontal oversampling areas. The third signal group (YH, IH, QH) is processed into a signal containing the edge-part heights. When these signals are combined, a NTSC-compatible widescreen signal with a 4: 3 aspect ratio is obtained. Details of the circuits forming the unit 18 are shown in Figures 6, 7 and 8 and will be described in connection with these Figures 6, 7 and 8.

The signals YE, IE and QE contain the complete center part information and have the same format as shown in Figure 3 for the signal YE. The signal YE is derived from the signal YF "in short as follows: The wide-field signal YF" includes picture elements 1 to 754 appearing during the active line interval of the wide-screen signal and including both edge-part and center-part information. The wideband center part information (picture elements 75 to 680) is extracted as the center part luminance signal YC by a time demultiplexing method. The YC signal is time stretched by the center portion expansion factor 1.19 (i.e., 5.0 MHz: 4.2 MHz) to produce the NTSC compatible center sub-signal YE. The signal YE has an NTSC compatible bandwidth (Obis 4.2 MHz) due to the time expansion by a factor of 1.19. The signal YE occupies the picture display interval TD (Fig. 2) between the oversampling areas TO. The signals IE and QE are derived from the signals IF "and QF" and processed in a similar manner as the signal YE.

The signals Y0, IO and Q0 contain the low-frequency parts ("depths") of the edge-part information, which are inserted in the left and right horizontal oversampling areas. The signals Y0, IO and Q0 have the same format as that for the signal Y0 is shown in Fig. 3. The signal Y0 is derived from the signal YF "in short as follows. The

Widescreen signal YF "contains left edge portion information in picture elements 1 to 84 and right edge portion information in picture elements 671 to 754. As will be explained, signal YF" is low-pass filtered to produce a luminance depth signal having a bandwidth of 0 to 70OkHz. from which is then extracted by means of a time-division multiplexing a left j and a right edge portion depth signal (signal YL 'in Fig. 3). The luminance depth signal YL 'is time-minimized to produce the edge portion depth signal YO with the compressed low-frequency information in the oversampling areas, that is, in the picture elements 1 to 14 and 741 to 754. The compressed edge portion depth signal is proportional to the measure of Time compression increased bandwidth. The signals IO and QO are developed from the signals IF "and QF" and processed in the same way as the signal YO.

The signals YE, IE, QE and YO, IO and QO are generated by a border part / center part combining circuit 28, e.g. a time division multiplexer, combined to generate signals YN, IN and QN with an NTSC compatible bandwidth and a 4: 3 aspect ratio. These signals take the form of the signal YN shown in FIG. Combination circuit 28 also includes appropriate signal delays to equalize the durations of the signals to be combined. Such compensating signal delays are also found elsewhere in the system where signal transit times are required.

A modulator 30, a bandpass filter 32, a horizontal / vertical / temporal notch filter 34, and the combining circuit 36 constitute an improved NTSC signal coder 31. The chrominance signal components IN and QN are modulated by the modulator 30 onto a subcarrier SC of the NTSC color subcarrier frequency 3.58 MHz modulated to produce a modulated signal CN. The modulator 30 is of conventional design and will be described in conjunction with FIG. The modulated signal CN undergoes vertical (V) and temporal (T) dimension bandpass filtering in the two-dimensional (VT) filter 32 which removes crosstalk artifacts in the interlaced chrominance signal before applying that signal as a signal CP to a chrominance signal input of the combining circuit 36 becomes. The luminance signal YN undergoes band-stop filtering in horizontal (H), vertical (V), and temporal (T) dimensions by means of the three-dimensional H-V-T band notch filter 34 before being applied as a signal YP to a luminance input of the combining circuit 36. The filtering of the luminance signal YN and the chrominance color difference signals IN and QN is intended to ensure that the crosstalk between luminance and chrominance is substantially reduced after the subsequent NTSC coding. Multi-dimensional spatio-temporal filters such as the H-V-T filter 34 and the V-T filter 32 in Figure 1 a have a structure as shown in Figure 10 and described below.

The H-V-T band-stop filter 34 in Fig. 1a has the structure shown in Fig. 10b and removes the frequency components of upward-moving diagonals from the luminance signal YN. These frequency components are similar in appearance to components of the color subcarrier and are removed to create a gap in the frequency spectrum where modulated chrominance information is inserted. The removal of the frequency components of upwardly moving diagonals from the luminance signal YN does not visibly deteriorate a reproduced image because it has been found that the human eye is practically insensitive to such frequency components. The filter 34 has a cut-off frequency of about 1.5 MHz in order not to affect the vertical detail information in the luminance.

The V-T bandpass filter 32 reduces the chrominance bandwidth so that the modulated chrominance information of the edge portions can be inserted into the gap created by the filter 34 in the luminance spectrum, as described above. The filter 32 reduces the vertical and temporal resolution of the chrominance information, so that steep and moving edges are easily blurred, but with little or no consequences because the human eye is insensitive to such effects. The combiner circuit 36 provides at its output a mid-section / edge-subcircuit signal C / SL, which contains NTSC-compatible information to be reproduced, derived from the center portion of the wide-screen signal, and compressed edge sub-lows (both luminance and chrominance) derived from the Are now derived in the left and right horizontal oversampling areas, not being seen by a viewer on an NTSC receiver. The compressed edge portion depths in the oversampling area form part of the edge portion information for widescreen reproduction. The other component, the margin heights, is developed by the processing unit 18, as described below. The edge part height signals YH (luminance levels), IH (I-heights) and QH (Q-heights) are shown in FIG. Figures 6, 7 and 8 show an arrangement for developing these signals, as will be described. The signals YH, IH and QH shown in FIG. 4 contain high-frequency information of the left edge part in the picture elements 1 to 84 and high-frequency information of the right edge part in the picture elements 671 to 754.

The mid-portion of the signal C / SL is processed in the intra-frame averaging circuit 38 to produce a signal which is fed to the input of the adder 40. The intraframe-averaged signal N is substantially identical to the signal C / SL because of the highly visible correlation of the intraframe image information of the signal C / SL. The averaging circuit 38 averages the signal C / SL above about 1.5 MHz and helps to reduce or eliminate vertical time crosstalk between the main signal and the supplemental signals. The high pass frequency range of 1.5MHz and above, in which the intraframe averaging circuit 38 operates, has been chosen to provide full intraframe averaging for information at 2 MH and above to prevent luminance vertical detail information from being passed through the process the intraframe message is affected. Horizontal crosstalk is eliminated by a 200 kHz guard band between a filter belonging to the intra-name averaging circuit 38 in the coder 31 and a filter belonging to an intra-frame processing unit in the decoder of Figure 13. Figure 11b shows details of the intraframe averaging circuit 38 provided for the heights. Figures 11b and 13 are explained below. The signals IH, QH and XH are brought into the NTSC format by means of an NTSC encoder 60, which is similar to the encoder 31. Specifically, the encoder 60 includes means of the type shown in Figure 9, and means for quadrature modulating the heights of the edge portion chrominance information to the heights of the edge portion luminance information at 3.58 MHz to produce a sign 'NTSCH. which represents the heights of the margin information in NTSC format. This signal is shown in FIG.

The use of multi-dimensional bandpass filtering in NTSC encoders 31 and 60 advantageously allows the luminance and chrominance components to be separated from each other substantially free of crosstalk in the receiver if the receiver uses complementary multidimensional filtering to separate luminance and chrominance filters

Chrominance information used. The use of complementary filters for Jie coding and decoding of luminance and chrominance are referred to as cooperative processing; it is described in detail by CH Strolle in an article entitled "Cooperative Processing for Improved Chrominance / Luminance Separation," published in the SMPTE Journal, Vol. 95, No. 8, August 1986, pp. 782-789. Even standard receivers which Using conventional notch filters and row comb filters benefit from the use of such multi-dimensional filtering in the coder by showing less crosstalk between chrominance and luminance.

The signal NTSCH is time stretched in a unit 62 to produce a stretched Randieil height signal ESH having an active horizontal row interval of 50 μs, ie less than the standard active NTSC line interval of approximately 52 μs. More specifically, as shown in Fig. 5, the stretching is performed by an "imaging" process in which the left edge subpixels 1 to 84 of the signal NTSCH are mapped to the pixel positions 15 to 377 of the signal ESH, ie, the heights of the left edge portion The signal NTSCH is stretched to occupy approximately half of the line duration of the signal ESH, and the part (pixels 671 to 754) of the signal NTSCH associated with the right edge portion is processed in a similar manner The time expansion process reduces the horizontal bandwidth of the signal information ESH (compared to that of signal NTSCH) by a factor of 363/84.

The imaging process by which the temporal stretching takes place can be realized by means of the type shown in FIGS. 12 to 12d and described below in connection with these figures. The signal ESH undergoes intraframe averaging in a network 64 of the type shown in Figure 11a to produce a signal X, as shown in Figure 5. This average signal X is substantially identical to the signal ESH, because of the highly visible Konelation dor image information of the signal ESH within a frame. The signal X is applied to a signal input of a quadrature modulator 80.

The signal YF 'is also filtered by a horizontal bandpass filter 70 having a bandwidth of 5 to 6.0 MHz. The output of the filter 70, containing horizontal luminance levels, is applied to an amplitude modulator 72 in which it modulates the amplitude of a 5 MHz carrier signal f _c . The modulator 72 has a low-pass filter on the output side with a cutoff frequency of approximately 1.0 MHz in order to obtain a signal in the band range 0 to 1.0 MHz at the output of the modulator 72. The upper ("folded over") sideband (5.0-6.0 MHz) of the modulation process is removed by the 1.0 MHz low-pass filter, effectively, the horizontal luminance-treble frequencies are in the range of 5.0-6, The carrier amplitude should be high enough to maintain the original signal amplitudes after filtering by the 1.0 MHz low-pass filter, that is to say, a 0 MHz is shifted in the frequency range 0 to 1.0 MHz by the amplitude modulation process and the subsequent low-pass filtering The frequency shifted signal of the horizontal luminance levels from the unit 72 is operated (time compressed) by means of a picture format encoder 74. That is, the encoder 74 encodes the frequency shifted horizontal luminance levels such that this signal has an active line interval of 50με and thus less than the default active NTSC interval of 52.6ps, which is done by application tion of methods, which are explained in connection with Figures 6 to 8. When the input signal at the encoder 74 is time compressed therein, its bandwidth increases from approximately 1, OMIIz to 1.1 MHz at the output of the encoder 74. The signal from the encoder 74 is amplified by means of a device 76 similar to that shown in FIG 11a, undergoes intraframe averaging before being sent to the unit 60 as a signal. The intra-frame averaged signal Z is substantially identical to the signal from the encoder 74 because of the highly visible correlation of the image information within one frame of the signal coming from the encoder 74. The modulation signal X is a composite signal containing luminance and chrominance information, and the modulation signal has substantially the same bandwidth of approximately 0 to 1.1 MHz. As will be explained in connection with FIG. 24, in the case of large amplitude fluctuations of the two additional signals X and Z, the unit 80 performs non-linear amplitude compression according to a gamma function, before these signals quadrature modulate an alternating subcarrier signal ASC. A gamma value of 0.7 is used whereby the absolute value of each sample is raised to the 0.7th power and multiplied by the sign of the original sample value. Gamma Compression reduces the visibility of potentially large amplitude swings of the modulated signals in normal receivers and allows predicted retrieval in the wide-screen receiver because the inverse function of the gamma function used in the encoder is predictable and can be easily realized in the decoder of the receiver.

The amplitude-compressed signals are then quadrature-modulated onto the 3.1075 MHz alternated phase alternating subcarrier ASC, the frequency of which is an odd multiple of half the horizontal line frequency (395 χ H / 2). The phase of the alternating subcarrier is changed from one frame to the next field by 180 °, unlike the phase of the color subcarrier, which does not change from field to field The field alternating phase of the alternating subcarrier allows the modulating information of the auxiliary signals X. This results in complementary phase information components A1, -A1 and A3, -A3 of the modulated additional signal, which makes it possible to separate the additional information in the receiver without using a relatively uncomplicated field memory The quadrature-modulated signal M becomes in the addition stage 40 is added to the signal N. The resulting signal NT3CF is a 4.2 MHz NTSC-compatible signal The non-linear gamma function described is used to compress large amplitudes and forms part of a non-linear compressing and stretching companding system (compression / expansion), which as a further component for the expansion of the amplitude contains a complementary gamma function in the decoder of a widescreen receiver, as will be described below. It has been found that the disclosed non-linear companding system significantly reduces the impact of non-standard additional information on standard image information without visibly degrading an image by noise effects. The companding system uses a non-linear gamma function to compress large amplitude excursions of the additional non-standard high-frequency wide-screen information in the combiner and a complementary non-linear gamma function to correspondingly stretch such high-frequency information in the dukoder. The result is that large amplitudes of the high frequency additional information less disturb the existing standard video information in the described compatible widescreen system, splitting the non-subordinate extra wide-screen information into low-frequency parts and high-frequency parts subject to companding

become. In the decoder, nonlinear amplitude expansion of the compressed high frequency information does not result in excessively noticeable noise. That is, large amplitudes of high frequency information are typically associated with high contrast image edges, and the human eye is insensitive to the noise at such edges. The companding process described also advantageously reduces cross-modulus products between the alternating subcarrier and the chrominance subcarrier, thereby also reducing visible beat phenomena associated therewith.

The luminance detail signal YT in Figure 1 a has a bandwidth of 7.18 MHz and is encoded by means of a picture format encoder 78 on the 4: 3 picture format (for example, in the same manner as shown in Figure 6) and then in a filter 79 horizontally low pass filtering 750KHz to produce a signal YTN. The edge portions are low-pass filtered prior to their time compression to 12SkHz, which occurs in an input low-pass filter of the picture format encoder 78, which corresponds to the input filter 610 of the device shown in Figure 6, but has a cut-off frequency of 125kHz. The heights of the round parts are suppressed. Thus, the signal YTN is spatially related to the main signal C / Sl. correlated.

The YTN and NTSCF signals are converted from the (binary) digital form to the analog form with HiKo by digital-to-analog converters (D / A) 63 and 54 before being passed to an RF quadrature modulator 57 for transmission to a television RF Aufzumodulieren carrier signal. The RF-modulated signal is then applied to a transmitter 55 to radiate it via an antenna 56.

The alternating subcarrier ASC used in the modulator 80 is horizontally synchronized, and the frequency is chosen to provide sufficient separation. Edge part and center part information (e.g., 20 to 3OdB), and that there is no significant effect on a picture rendered with a standard NTSC receiver. The ASC frequency should preferably be an interlaced frequency with an odd multiple of half the horizontal line frequency, so that no disturbances are caused which could affect the quality of a reproduced image.

The quadrature modulation thus performed in the unit 80 advantageously permits the simultaneous transmission of two narrowband signals. The temporal stretching and modulation of high frequency signals results in a reduction in bandwidth, in accordance with the narrowband requirements of quadrature modulation. The more the bandwidth decreases, the less likely it is that interference between carrier and modulation signals will occur. In addition, the DC component of the edge portion information, which typically occurs with high energy, is forced into the oversampling area and not used as a modulation signal. Thus, the energy of the modulation signal and therefore the associated interference is very reduced.

The encoded and transmitted via the antenna 56 NTSC-compatible widescreen signal is to be received by both NTSC receivers and widescreen receivers, as shown in FIG. Referring now to FIG. 13, a broadcast widescreen enhanced wide-screen television signal is intercepted with an antenna 1310 and fed to the antenna input of an NTSC receiver 1312. The receiver 1312 normally processes the compatible widescreen design?) To render a 4: 3 aspect ratio image, with the wide part edge information (ie, the "depths") being pressed into the horizontal oversampling areas out of sight of the observer and in part (ie, the "heights") are included in the modulated alternating subcarrier signal, which does not interfere with the operation of the standard receiver. The extended-width widescreen (EDTV) widescreen signal intercepted by the antenna 1310 is also applied to a progressive-scan wide-screen receiver 1320 which captures an image having a high aspect ratio of e.g. B. 5: 3 can play. The received widescreen signal undergoes first processing in an input unit 1322 including an RF tuner and amplifier circuits, a baseband video signal generating synchronous video modulator (quadrature demodulator), and analog-to-digital converter circuits (AD wall); to generate a baseband video signal (NTSCF) in binary form. The A / D converter circuits operate at a sample sampling frequency equal to four times the color subcarrier frequency (4x fsc).

The signal NTSCF is supplied to an intraframe processing unit 1324 in which image lines are processed in a frequency range above 1.7 MHz, each being spaced apart by 262 horizontal line periods (262 H) within a respective frame, so that the main signal N and the quadrature modulated additional signal M virtually free of vertical and temporal crosstalk (VT crosstalk). Between the lying at 1.7 MHz lower limit of the operating frequency of the unit 1324 and lying at 1.5 MHz lower limit of the operating frequency of the unit 38 in the encoder of Figure 1a is a 20OkHz wide security band against horizontal crosstalk. The recovered signal N contains information which is substantially visually identical with the image information of the main signal C / SL, because of the good correlation of the visual information within a frame of the original main signal C / SL, which detects the intraframe averaging in the encoder Figure 1 has experienced a.

The signal M is applied to a quadrature demodulator and amplitude dilation unit 1326 to demodulate the additional signals X and Z by means of an alternating subcarrier ASC whose phase changes from field to field, similar to the subcarrier signal ASC described in connection with Figure 1a. The domodulated signals X and Z contain information which is visually substantially identical to the image information of the signal ESH or the output signal of the unit 74 in Figure 1a, due to the correlation of the highly visible information within each one frame in these signals, which have undergone an intraframe averaging in the encoder of Figure 1 a. The unit 1326 also includes a 1.5 MHz low pass filter to remove unwanted high frequency dual frequency demodulating products of the alternating subcarrier, and an amplitude stretch circuit to stretch the (previously compressed) demodulated signals using an inverse gamma function, which is the inverse is the non-linear compression function used by the unit 80 in Figure 1 (ie has a gamma value of 1 / 0.7 = 1.429). A unit 1328 effects time compression of the color coded high frequency components of the edge portion information so that they again occupy their original time slots, thereby recovering the signal NTSCH. The unit 1328 compresses the signal NTSCH temporally by the same amount by which the unit 62 in FIG. 1a has temporally expanded the signal NTSCH. An encoder 1330 decodes the signal Z containing the horizontal-high frequency components of the luminance (Y-height) into the

Widescreen format, by temporally extending this signal by the same amount as the time compression of the corresponding component in the encoder of Figure 1a was and is indicated in Figure 17 using the imaging method described therein.

A modulator 1332 outputs the signal from the decoder 1330 by means of amplitude modulation to a 5.0 MHz carrier f _c . The amplitude modulated signal is then high pass filtered in a filter 1334 with a cut-off frequency of 5.0 MHz to suppress the lower sideband. In the output signal of the filter 1334, the frequencies of the center part information of 5.0 to 6.0 MHz and the frequencies of the edge part information of 5.0 to 6.0 MHz are recovered. The signal provided by the filter 1334 is given to an adder 1336.

The Dac signal NTSCH from the compressor 1328 is applied to a unit 1340 to separate the luminance levels from the chrominance levels and generate signals YH, IH, and QH iu. This can be done by means of the arrangement shown in FIG.

The signal N from the unit 1324 is separated into its luminance and chrominance components YN, IN and QN in a luminance / chrominance separator 1342. The separator 1342 may be similar to the separator 1340 and use an arrangement of the type shown in FIG.

The signals YH, IH, QH and YN, IN, QN are applied to the inputs of a Y-I-Q picture format decoder 1344, which decodes the luminance and chrominance components into the wide format. The edge part depths are time stretched, the center part information is time compressed, the edge part heights are added to the edge part depths, and the edge parts are joined to the middle part to form the ten pixel wide overlap area, using the method shown in FIG illustrated principles. Details of the decoder 1344 are shown in FIG.

The signal YF 'is fed to the adder 1336, where it is summed with the signal coming from the filter 1334.

Through this process, retrieved information with extended high frequency horizontal luminance details is added to the decoded luminance signal YF '.

The YF ', IF' and QF 'signals are converted to progressive scan format using framing converters 1350, 1352 and 1354, respectively, in the interlaced scanning format. The luminance image-to-noise converter 1350 also responds to the "helper" luminance signal YT from the image format decoder 1360, which decodes the decoded "helper signal" YTN. The coder 1360 decodes the signal YTN into the widescreen format and has a structure similar to the structure shown in FIG.

The I and Q component image-to-noise converters 1352 and 1354, respectively, place the interlaced signals in the progressive scan format by forming the time average of lines that are one frame apart so as to provide the information for the respective missing line to get progressive sampling; this can be done by means of a device of the type shown in FIG.

The luma image-to-noise converter 1350 is similar to the arrangement shown in FIG. 20, except that the signal YT is added as shown in the arrangement of FIG. In this unit, a sample of the "helper" signal YT is added to a time average to help reconstruct a missing progressive scan pixel. The complete time details are recovered within the band of horizontal frequencies recorded in the encoded row differential signal (7GOKHz, after encoding). Above this band of horizontal frequencies, the signal YT is equal to zero, so that the missing picture element is reconstructed by time averaging.

The widescreen signals YF, IF and QF, now in the progressive scan format, are converted to the analog form by a digital-to-analog converter 1362 before being fed to a video signal processor and matrix amplifier unit 1364. The video signal processor in unit 1364 includes signal amplification circuitry, DC level shifting, peaking, brightness control, contrast control, and other conventional video signal processing. The matrix amplifier 1364 combines the luminance signal YF with the color difference signals IF and QF to produce the video color signals R, G and B. These color signals are amplified in the unit 1364 by drivers to a level more suitable for directly driving a wide-format color-capable display device, such as a video recorder. a widescreen picture tube, is suitable.

FIG. 6 shows an arrangement contained in the processing unit 18 of FIG. 1 a for developing the signals YE, YO and YH from the wide-band widescreen signal YF. The signal YF "undergoes high-pass low-pass filtering in an input filter 610 having a gene frequency of 7COKHz to produce the low-frequency luminance component signal YL, which is fed to one input of a subtractive combining circuit 612. The signal YF" is applied to the other input of the Combination circuit 612 and a time-demultiplexing circuit 616, after having been delayed in a unit 614 to the by processing in Fiue. 610 caused signal delay to compensate. The combination of the delayed signal YF "and the filtered signal YL would be" "the signal containing the high-frequency luminance components! YH at the output of combination circuit 612.

The delayed signal YF "and the signals YH and YL are supplied to separate inputs of the demultiplexer 616, which includes individual demultiplexer units (abbreviated to DEMUX) 618, 620 and 621 to process the signals YF", YH and YL. The details of the demultiplexing device 616 will be explained in connection with FIG. The demultiplexer units 618, 620 and 621, respectively, provide the full-bandwidth center sub-signal YC, the margin sub-signal YH and the margin sub-signal YL 'as shown in Figures 3 and 4.

The signal YC is time stretched in a time expander 622 to produce the signal YE. This stretching is done with a center stretch factor that is sized to leave room for the left and right horizontal oversampling areas. This center-portion expansion factor (1.19) is the ratio of the desired width of the signal YE (image elements 15 to 740 to the width of the signal YC (pixels 75 to C80) as shown in FIG.

The signal YL 'is time compressed in a time compressor 628 by an edge compression factor to produce the signal YO. The edge compression factor (6.0) is the ratio of the width of the corresponding part of the signal YL '(eg, the left picture elements 1 to 84) to the desired width of the signal Y0 (eg, the left picture elements 1 to 14) ), as shown in FIG. The time stretchers 622, 624 and 626 and the time compressors 628 may be of a type as shown in FIG. 12 and described below.

The signals IE, IH, IO and QE, QH, QO are developed from the signals IF "and QF" in a similar manner as the

Signals YE, YH and YO with the aid of the device according to Fig. 6. Considering Fig. 7, which shows a device for generating the signals IE, IH and IO from the signal IF ", the development of the signals QE, QH and QO the signal QF "occurs in a similar way.

Referring to Fig. 7, wideband wideband signal IF "is applied to demultiplexing circuit 716 after delay in unit 714 and also subtractively combined in subtractive combining circuit 712 with low frequency signal IL from low pass filter 710 to produce high frequency signal IK IF "and the signals IH ui. IIL are respectively decoded by demultiplexers 718, 720 and 721 in the demultiplex circuit 716 so as to obtain the signals IC, IH and IL '. The signal IC is time-expanded in a time-stretching means 722 to generate the signal IE, and the signal IL 'is time-compressed in a time compressor 728 to produce the signal IO. The stretching of the signal IC takes place with a similar mean stretch factor as the described stretch of the signal YC, and the compression of the signal IL 'is done with a similar edge part compression factor as the described compression of the signal YL'.

FIG. 8 shows a demultiplexing circuit 816 which can be used for the circuit 616 in FIG. 6 and the circuit 716 in FIG. Figure 8 shows the circuit in its application as the demultiplexer circuit 616 of Figure 6. The input signal YF "contains 754 picture elements which define the picture information The picture elements 1 to 84 define the left edge part, the picture elements 671 to 754 define the right edge part, and the pixels 75 to 680 define the central portion which slightly overlaps the left and right edge portions, and the signals IF "and QF" show a similar overlap As will be described, such an area overlap will allow the center portion and edge portions in the receiver The demultiplexing circuit 816 includes first, second and third demultiplexing units (DEMUX) 810, 812 and 814, respectively, for the information of the left edge portion and the center portion, respectively Each demultiplexer unit has an input "A" to which the signal YH or YF "and YL, respectively, and an input" B "for applying a blanking signal (BLK). The blanking signal can be, for example, a logic value 0 or ground potential. The unit 810 extracts the output signal YH including the heights of the left and right edge portions from the input signal YH, as long as the select input SEL from a count comparator 817 receives a first control signal indicating the presence of the left edge portion pixels 1 to 84 and the output Display pixels 671 to 754 of the right edge portion. At other times, a second control signal from count comparator 817 causes the blanking signal BLK from input B to be coupled to the output of unit 810 instead of signal YH from input A. The unit 814 and a count comparator 820 similarly operate to derive the edge portion depth signal YL 'from the signal YL. The unit 812 couples the signal YF "from its input A to its output to produce the midsection signal YC only when a control signal from a count comparator 818 indicates the presence of the mid-section pixels 75-680. The count comparators 817, 818 and 820 are shown with the video '. _d , al YF "is synchronized by a pulse signal from the output of a counter 822 responsive to a clock signal of four times the color subcarrier frequency (4x fsc) and a horizontal line sync signal H derived from the video signal YF" The counter 822 has an initial offset by starting with the count-100 corresponding to the 100 pixels from the beginning of the negative-going horizontal sync pulse at time T _H s to the end of the horizontal blanking interval at which the picture element 1 is at the beginning of the horizontal line playback interval Thus, the counter 822 is at the count of "1" at the beginning of the line rendering interval. Other meter arrangements can also be tapped. The principles employed by the demultiplexing circuit 816 may also be used in multiplexing circuits for performing a reverse signal combining operation, as described, for example, in US Pat. B. runs in the edge part-central part combination circuit 28 of Figure 1a.

FIG. 9 shows details of the modulator 30 in the encoders 31 and 60 of FIG. 1a. According to FIG. 9, signals IN and ON, which appear at four times the chrominance carrier frequency (4x fsc), are respectively applied to the signal input of an assigned latch (latch circuit) 910 and 912, respectively. The latches 910 and 912 also receive clock signals of frequency 4x fsc to supply the signals 2xfsc supplied to an inverting switching signal input of the latch circuit 910 and a non-inverting switching signal input of the latch circuit 912. *** " The signal outputs of the latch circuits 910 and 912 are combined into a single output line on which the signals I and Q appear alternately and go to the signal inputs of a non-inverting latch 914 and an inverting latch 916. These latches are clocked at frequency 4χfsc and each of them receives a chrominance carrier frequency switching signal fsc, the first at an inverting input and the second at a non-inverting input. The non-inverting latch 914 provides at its output in alternating sequence signals I and Q of positive polarity, and the inverting latch 916 provides at its output in an alternating sequence signals I and Q of negative polarity, that is signals -1, -Q. The outputs of the latch circuits 914 and 916 are combined into a single output line on which the signals I and Q appear alternately and in pairs alternating between opposite polarities, that is, in sequence I, Q, -I, -Q ,. .. etc., so as to form the signal CN. This signal is filtered in the filter 32 before being combined in unit 36 with a filtered version of the luminance signal YN to produce the encoded NTSC signal C / SL in the form Y + I, Y + Q, Y - I, Y - Q ₁ Y + I Y + Q _1,.., etc. the figure 10 shows a filter for vertical-temporal filtering (vertical-temporal filter or abbreviated VT) filter which by setting weight coefficients a 1 to a9 to a VT bandpass filter , a VT notch filter or a VT low pass filter. The table in Figure 10a shows the weight coefficients for VT boundary pass filtering and VT bandstop filtering as used in the system disclosed herein. An HVT notch filter, such as a filter 34 of Figure 19, and HVT bandpass filters included in the decoding system of Figure 13, for example, each comprise the combination of a horizontal lowpass filter 1020 and a VT bandstop filter 1021 as shown in Figure 10b and the combination of a horizontal bandpass filter 1030 and a VT bandpass filter 1031 as shown in Figure 10c. In the HVT bandstop filter of Figure 10b, the horizontal low pass filter 1020 has a given cutoff frequency and provides a filtered low frequency signal component. This signal is subtractively combined in a combining circuit 1023 with a delayed version of the input signal from a delay unit 1022 to produce a high frequency signal component. The low-frequency component is by means of a network 1024 by a Voüoildpcriode

delayed before being applied to an additive combining circuit 1025 to provide an output signal which has undergone H-V-T band-stop filtering. The V-T filter 1021 has the weight coefficient indicated in Figure 10a for the V-T band-stop filter. An H-V-T bandpass filter, as it is z. 13, there is shown in FIG. 10c a horizontal bandpass filter 1030 having a given cut-off frequency in cascade with a V-T bandpass filter 1031 having the weight coefficients indicated in the V-T bandpass filter table of FIG.

The filter of Fig. 10 includes a plurality of cascaded memory units (M) 1010a to 1010h for progressively receiving signal delays at individual taps 11 to 19 to provide a total filter delay. The signals taken at the taps are respectively supplied to an input of associated multiplier circuits 1012a to 1012i. Another input of each multiplying circuit receives a prescribed weighting signal a 1 to a9, depending on the nature of the filtering operation to be performed. The nature of the filtering process also dictates the delays introduced by the storage units 1010a-1010h.

Horizontal dimension filters contain pixel memory elements such that the total filter delay is shorter than one horizontal line period (1H). Vertical dimension filters contain only line storage elements, and temporal dimension filters contain only full-screen storage elements. Thus, a three-dimensional HVT filter consists of a combination of pixel memory elements (<1H), row memory elements (1H), and frame memory elements (> 1H), while a VT filter contains only the last two types of memory elements , The weighted tapped (mutually delayed) signals from the elements 1012 a to 1012 i are combined in an adder stage 1015 to produce a filtered output signal. to deliver.

Such filters are non-recursive filters with finite impulse response (so-called Fl-R filters). The nature of the delay introduced by the memory elements depends on the type of signal to be filtered, and in the present example also on what degree of crosstalk between the luminance, the chrominance and the i.threshold edge signals can be tolerated. The sharpness of the filter boundary characteristic and thus the sharpness of the filtering is improved by increasing the number of cascaded memory elements.

FIG. 10d shows one of the separate filters in the network 16 of FIG. 1a. This filter includes cascaded memory units (delay units) 1040a-1040d and associated multiplying circuits 1042a-1042e with designated weighting factors a 1 through a 5, which receive the signals from taps 11 through t5. Further, a combination circuit 1045 is provided which sums the weighted outputs of the multiplier circuits a 1 to a 5 to form an output therefrom.

FIG. 11a shows an intraframe averaging device which is suitable for the intraframe averaging devices 64 and 78 of FIG. 1a. A composite video signal is applied to the delay network having the 262 H delay elements 1110 and 1112 and also to one input of a multiplexer (MUX) 1115, which is switched to a 30 Hz switching signal at a field rate of one. The 30 Hz MUX switch signal is vertically synchronized in response to vertical interval sync pulses associated with the composite input video signal. Another input of the multiplexer (MUX) 1115 receives a signal from an output of the delay element 1112. In addition, a combination circuit 1118 combines outputs from the multiplexer 1115 and a point of center tap between the delay elements 1110 and 1112, after these signals have an averaging factor 1/2 have been evaluated. This weighting factor may be provided by appropriate matrix networks within the combinational circuit 1118 or by means of signal multiplying circuits respectively arranged in the input signal paths of the combinational circuit 1118.

The signals "Y1 + CV and" Y2 + C2 "are composite color video signals in spaced apart first 262 H in successive first and second fields, and the signal" M1 "is an intraframe-averaged output signal as shown, for example, in FIG is shown. During a first field, the multiplexer 111 u is in an input position "1" and transmits the signal Y 2 + C2 to the combinational circuit 1118 in which it is summed with the signal tap Y 1 + C 1 to generate an output side average signal M1 In the next field, the center tap between the delay elements 1110 and 1112 contains the signal value Y2 + C2 and the multiplexer 1115 is in position "2" for selecting the signal path from the output of the delay element 1112 containing the signal value Y1 + C1, thereby obtaining the same averaging value of the Signal M1 is supplied from the output of combiner 1118. The described device generates identical picture elements located at a distance of 262 H and is not limited in the application of an averaging process. Any weight values may be used to produce a desired weighted combination of pixel pairs, and delays other than 262 H may be used (along with an associated change in the switching frequency of the multiplexer), depending on requirements of a particular system. FIG. 11b illustrates a frequency selective intraframe averaging device suitable as intraframe averaging device 38 of FIG. 1a. Figure 11b contains the arrangement of Figure 11a except that the combination circuit 1128 combines the subtractive and not the additive signal combination, and that a filter 1130, a gate 1132 and a combiner 1134 are included. In short, for the output of the combinational circuit 1128, a difference of the fields and not an average value as in the arrangement of Figure 11a is representative. This difference is essentially an erasure term added backwards to the signal Y1 + C1 in the combiner 1134 for canceling the difference between successive fields to ensure that the portions of the successive fields are averaged identically. Filter 1130 filters the cancellation term from the output of combiner 1128 to limit the averaging process to a desired frequency range. The gate 1132 is controlled to detect when the averaging process occurs in an image interval, in this case in the middle portion, except for the temporally compressed edge portions. In particular, the input signal paths to combining circuit 1128 have signal weight coefficients of '/ 2 and -' / 2 as shown, so that the output from combining circuit 1128 corresponds to the difference in information content of the input signal to combiner 1128, which in temporal time in adjacent fields separated by 262 H from each other. The complementary weighting coefficients may be provided to the combiner circuit 1128 by using the signal multiplying circuits in the respective input paths

or by the arrangement of the combination circuit 1128 as a differential amplifier. The output signal from combiner 1128 is filtered in a 1.5 MHz horizontal high pass filter 1130 before being applied to electronic transfer device 1132. Gate 1132 is responsive to a switching control signal for passing the high frequency signal from the output of filter 1130 only during the mid portion of the main signal (component 1). At this time, the gate 1132 is open (conducting). Gate 1132 is closed (nonconducting) during the time-compressed edge portions of the main signal, for example, during the illustrated positive pulse pulse *; rvalle of the control signal. The output from gate 1132 is summed in combination circuit 1134 with the composite video signal appearing at the point of center tap between delay elements 1120 and 1122. The control signal of the gate is vertically synchronized with the associated composite video signal, triggered by the vertical interval sync pulse. The control signal of the gate is also synchronized horizontally. The horizontal synchronization can be achieved by responding to the horizontal line synchronizing pulse component of the composite input video signal including a picture element counter to determine the timing of the positive pulse components of the gate control signal following each horizontal line synchronizing pulse. Between a horizontal line sync pulse and the first picture element, a predetermined time interval can easily be accommodated.

In the following, reference is again made to FIG. 1 d together with FIG. 11 b. If multiplexer 1125 is in position 1 as shown and gate 1132 is closed, then only the composite video signal Y1 + C1 will appear from the center tap between elements 1120 and 1122 at the output of combiner 1134. Therefore, the output is at this time the combining circuit 1134, the unchanged compressed edge portion information of the composite video signal Y1 + C1, which belongs to the field 1. The output of combiner 1134 is the unchanged compressed edge portion information of the composite video as Y2 + C2 for subsequent field 2 when multiplexer 1125 occupies position 2.

When the multiplexer 1125 for field 1 is in position 1 and the gate 1132 is closed during the mid-portion interval between the edge sub-intervals, the output of the combiner 1134 contains the signal components Y1 + C1 and M1. The component Y1 + C1 contains unchanged, i. not intraframe-averaged, the middle part information at and below approximately 1.5MHz. The component M1 softens the intraframe-averaged midsection information above approximately 1.5 MHz. When the multiplexer 1125 is in position 2 during the succeeding field 2 and the gate 1132 is closed during the mid-portion interval, the output from the combiner 1134 includes the intraframe-averaged component 1 as explained above and the i'.o. component Y2 + C2. The latter component contains the unmodified (non-intraframe-averaged) middle part information at and below approximately 1.5 MHz.

FIG. 12 shows a raster imaging device that can be used for the time-stretching and time-compression devices in FIGS. 6 and 7. The waveforms referred to herein in Figure 12a illustrate the imaging process. FIG. 12a shows an input signal waveform S with a middle part between the picture elements 84 and 670. It is intended to map to the picture element positions 1 to 754 of an output waveform W by means of a time stretching process. The end pixels 1 and 670 from the waveform S are imaged directly on end pixels 1 and 754 of the waveform W. The intervening pixels are not imaged directly in a positional ratio of 1: 1 due to the temporal strain, in some cases not even in an integer positional ratio. The latter case is true, for example, when the pixel position 85, 33 of the waveform S corresponds to the integer pixel position 3 of the output waveform W. The pixel position 85, 33 of the signal S contains an integer part (85), and a fraction DX (0.33), and the pixel position 3 of the waveform W contains an integer part (3) and a fraction (0).

In the arrangement of Figure 12, a pixel counter 1210 operating at the frequency 4 x fsc provides an output write address signal M representing pixel locations (1 .. .754) of an output raster. The signal M is fed to a programmable read-only memory (PROM) 1212 which contains a look-up table of programmed values which depend on the nature of the raster image to be performed, e.g. For example, whether the image is compressing or stretching. In response to the signal M, the PROM memory 1212 provides an output read address (READ ADDRESS) signal N representing an integer and an output DX representing a fraction equal to or greater than zero but less than 1. In the case of a 6-bit signal DX (2 ^e = 64), the signal DX shows the fractions 0, Vm, Va, Vu ... * ³ / *.

The PROM memory 1212 shows a stretching or compressing representation of a video input signal S as a function of the stored values of the signal. ¹ ; N. Thus, in response to integer values of the pixel position signal M, a programmed value of the read address signal N and a programmed value of the fractional signal DX are provided. For example, to perform signal expansion, the PROM memory 1212 is designed to provide the signals N at a slower rate than the signals M. Conversely, if compression is to occur, the PROM memory 1212 provides the N signals at a faster rate as the signals M.

The video input signal S is delayed by cascaded picture element delay elements 1214a, 1214b, and 1214c to produce video signals S (N + 2), S (N + 1), and S (N), which are mutually delayed versions of the video input signal , These signals are applied to video signal inputs of associated dual port memories 1216a-1216d as they are known per se. The signal M is applied to the write address input of each of the memories 1216a to 1216d, and the signal N is applied to the read address input of each of the memories 1216a to 1216d.

The signal M determines where incoming video signal information is to be written into the memories, and the signal N determines which values are to be read from the memories. The memories are designed so that information can be written into one address while simultaneously reading out another address. The output signals S (N-1), S (N), S (N + 1) and S (N + 2) from the memories 1216a to 1216d show a time-compressed or time-compressed format depending on the read / write operation the memory 1210a-1216d, which depends on how the PROM memory 1212 is pruned

 The signals S (N-1), S (N), S (N + 1) and S (N + 2) from the memories 1216a to 121fd are stored in a four-point linear

Processing interpolator, filter 1220 and 1222 for peaking (peak filter), a PROM memory 1225 and a linear two-point interpolator 1230; Details of these parts are illustrated in FIGS. 12 b and 12 c. The peak filters 1220 and 1222 receive three signals from the group of signals S (N-1), S (N), S (N + 1), and S (N + 2) in the shown selection and also receive a peak signal PX. The value of the peak signal PX changes from 0 to 1 as a function of the value of the signal DM as shown in Figure 12d and is output from the PROM memory 1225 in response to the signal DX. The PROM memory 1225 contains a look-up table and is programmed to provide a given value of PX in response to a given value of DX.

The peak filters 1220 and 1222 provide two mutually delayed video signals S '(N) and S' (N + 1) that have been peaked to the linear two-point interpolator 1230, which also receives the signal DX. The interpolator 1230 provides a (compressed or stretched) video output signal W defined by the term

W = S '(N) + DX [S' (N + 1) -S '(N) I.

The described function of four-point interpolation and peaking is advantageously approximated to a (sinX) / X interpolation function with good resolution of the high-frequency details.

Figure 12b shows details of the peak filters 1220 and 1222 and the interpolator 1230. Referring to Figure 12b, the signals S (N-1), S (N) and S (N +) are applied to the weighting circuit 1240 in the peak filter 1220 in which these Signals (in the order enumerated) with peak coefficients - ¹ A, Vj and - ¹ A. As shown in Figure 12c, the weighting circuit 1240 includes multiplying circuits 1241a to 1241c for receiving the signals S (N-1), S ( N) and S (N + 1) are respectively multiplied by the associated peak coefficients - ¹ A and Va and ^-1 A. The outputs of the multipliers 1241a to 1241c are summed in an adder stage 1242 to produce a peak signal P (N ) which is multiplied in a further multiplying circuit 1243 by the signal PX to provide a peak signal which is then summed in an adding stage 1244 with the signal S (N) to provide the peak signal S '(N). The S Pitzenfilter 1222 has a similar structure and mode of action.

In the two-point interpolator 1230, the signal S '(N) is subtracted from the signal S' (N + 1) by a subtractor 1232 to produce a difference signal which is multiplied by a signal DX in a multiplier circuit 1234. The output of multiplier circuit 1234 is summed in an adder stage 1236 with signal S '(N) to produce output signal W.

FIG. 15 shows details of the intraframe processing device 1324 of FIG. 13. The decoding device of FIG. 15 basically corresponds to the coding device of FIG. 11b. A composite input video signal for the processing means 1324 in Fig. 15 includes in a first field the signal components "Y1 + C1" and "Mi + A1". In a subsequent second field, the input signal includes the components "Y2 + C2" and "M1-AV." The components YI + C1, M1 and Y2 + C2, M1 are components supplied from the intra-frame processor 38 as described in connection with FIG The components + A1 and -A1 represent the alternate ide hit carrier signal which is modulated with the component 2 and the component 3 of the intraframe-matched information of the units 64 and 76 for the respective subsequent fields Reference is made to Figures 1.1a and in particular to Figure 1d.

The intraframe processor of Fig. 15 operates in substantially the same manner as the arrangement of Fig. 11b previously discussed. If the multiplexer 1525 is in position 1, the component of the field difference is formed at the output of the combination circuit 1528. After filtering by high pass filter 1530 and gate combination by unit 1532, a component -A1 is obtained which, when combined in a combining circuit 1534 with signal Y1 + C1, M1 + A1, clears the modulated additional subcarrier component (+ AD In order to produce a recovered main signal Y1 + C1, M1, the component Y1 + C1 of the recovered main signal is not changed below the cutoff frequency de-o high pass filter 1530 of 1.7 MHz, and the component M1 represents the intraframe-averaged center part information above The field difference cancellation term (-AD is the recovered modulated additional signal A1 after inversion in the unity gain amplifier 1535.

The recovered main signal Y1 + C1, M1 corresponds to the signal N in FIG. 13 and is further processed in a network 1342 as explained. The recovered additional signal A1 corresponds to the signal M in FIG. 13 and is demodulated in a network 1326.

Figure 16 illustrates the operation of the network 1324 as shown in Figure 15 for the next subsequent field. In this case, the signal Y2 + C2, M1 -A1 is developed between the delay elements 1520 and 1522, and the multiplexer 1528 occupies position 2 for the receive signal Y1 + C1, M1 + A2. A recovered main signal Y2 + C2, M1 is asserted on Output of the combination circuit 1534 generates, and an opposite phase-modulated additional signal -A1 is recovered.

In FIG. 18, an HVT bandpass filter 1810, which is constructed like that of FIG. 10c and has a passband of 3.58 ± 0.5 MHz, passes the signal NTSCH to a subtractive combining circuit 1814, which also receives the signal NTSCH after it has passed a delay compensating delay device 1812. The separated luminance level signal YH appears at the output of combiner 1814. The filtered NTSCH signal from filter 1810 is quadrature demodulated in demodulator 1816 in response to chrominance subcarrier signal SC to produce chrominance levels IH and QH.

In Fig. 19, the signals YN, IN and QN are divided into the compressed edge part depths Y0, IO, Q0 and the extended center part signals YE, IE, QE by means of an edge middle part signal separator (time demultiplexer) 1940. *** " The demultiplexer 1940 may operate using the principles of the demultiplexer 816 of the previously discussed figure. The signals YO, IO, and QO are time-stretched by a time-expansion means 1942 by an edge-elongation factor (corresponding to the edge-compression factor in the encoder of Figure 1 a) to restore the original ranking of the edge-portion depths in the wide-screen signal; these recovered edge part depths are labeled YL, IL and QL. Similarly, in order to make room for the edge portions, the center part signals YE, IE and QE are in one

Time compression means 1944 time-compressing by a mid-portion compression factor (corresponding to the mean-to-the-valley expansion factor in the encoder of Figure 1 a) to restore the original spatial ordering of the mid-section signal in the wide-screen signal; the recovered mid-section signals are labeled YC, IC and QC. The time compressor 1944 and the time expander 1942 may be of a type previously described in FIG.

The spatially recovered edge portion heights YH, IH, and QH are combined with the spatially recovered edge portion depths YL, IL, and QL in a combining circuit 1946 to produce reconstructed edge portion signals YS, IS, and QS. These signals are combined with the reconstructed center-part signals YC, IC and QC in a combiner 1960 to form a fully reconstructed wide-field luminance signal YF 'and fully reconstructed wide-field color difference signals IF' and QF '. The assembly of the edge part and middle part signal components is done in a manner in which a visible edge at the boundaries between the middle part and the edge parts is virtually eliminated, as will be apparent from the following description of the combiner 1960 shown in FIG. In FIG. 20, the interleaved IF '(or QF') signals are delayed by 263H by element 2010 before being applied to one input of a dual port memory 2020. This delayed signal experiences another delay of 262 H in a 2012 element before being added to the input signal in an addition stage 2014. The output of the adder 2014 is applied to a frequency bisecting network 2016 before being applied to one input of a dual port memory 2018. The reading of the data from the memories 2020 and 2018 takes place at a speed of 8 χ fsc, while the data is written at a speed of 4 χ fsc. The output signals of the memories 2018 and 2020 are supplied to a multiplexer 2022 to produce progressive scan output signals IF (QF). In Fig. 20, there are also the waveforms of the interlaced input signal (two lines of which one includes the picture element scanning patterns C and the other of which the picture element scanning patterns X) and the progressive scan format output having the same picture element scanning patterns C and X shown.

FIG. 21 shows an arrangement that can be used as the transducer 1350 for the signal YF 'in FIG. The interlaced signal YF 'is delayed by elements 2110 and 2112 before being combined with the non-delayed version in adder 2114. The delayed signal from element 2110 is applied to a dual port memory 2120. The output of the adder 2114 is applied to a frequency bisecting network 2116 whose output is added in an adder 2118 to the signal YT. The output of the adder 2118 is supplied to a dual port memory 2122. The memories 2120 and 2122 are operated at a write frequency of 4 χ fsc and a read frequency of 8 χ fsc and provide output signals to a multiplexer 2124 which provides the signal YF in the progressive scan format.

FIG. 14 shows a peripheral part-middle combiner, which is e.g. B. can be used for the element 1960 in Figure 19. The combiner shown in Fig. 14 includes a network 1410 which generates the full-bandwidth luminance signal YF 'from the edge-portion luminance component YS and the center-part luminance component YC, and an I-signal combiner 1420 and a Q-signal combiner 1430 in structure and operation are similar to the network 1410. The center area and the edge portions are intentionally overlapped by some picture elements, e.g. B. 10 picture elements. Thus, the center and edge signals have some picture elements in redundant manner throughout the encoding and transmission process prior to their merging.

In the widescreen receiver, the center part and the edge parts are recovered from the respective signals concerned, however, because of the experienced temporal stretching, temporal compression and filtering of the signals associated with each area some pixels are distorted or distorted at the boundaries between edge parts and middle part. The waveforms of the signals YS and YC shown in Fig. 14 show the overlapping areas (UL) and the corrupted picture elements (CP, somewhat exaggerated for illustration). If there were no overlapping areas of the individual parts, the falsified picture elements would collide, so that a hem would be visible. It has been found that an overlap area of 10 pixels is wide enough to compensate for three to five falsified boundary pixels. The redundant picture elements advantageously allow a mutual blending of edge and middle parts in the overlapping area. A multiplying circuit 1411 multiplies the margin signal YS by a weight function W in the overlapping areas as indicated by the associated waveform before applying the signal YS to a combining circuit 1415. Similarly, a multiplier circuit 1412 multiplies the center sub-signal YC by a complementary weight function (1-W) in the overlap areas, as shown with the associated waveform, before passing the signal YC to the combiner circuit 1415. These weight functions have a linear ramp-shaped characteristic in the overlap areas and have values between 0 and 1. After the evaluation, the edge-part and center-part picture elements in the combination circuit 1415 are summed so that each reconstructed picture element is a linear combination of edge-part and center-part picture elements. The weight functions should preferably approach the value 1 near the inner limits of the overlap area and the value 0 near the outer limits. This ensures that the corrupted picture elements have relatively little influence on the boundaries of the reconstructed areas. The illustrated linear ramped weight function fulfills this requirement. However, the weight functions need not necessarily be linear, nonlinear weight functions curvilinear or rounded in the end regions (i.e., near the weight points 1 and 0) may also be used. Such a weighting function can be easily achieved by filtering a linear ramped weight function of the type shown.

The weight functions W and (1-W) can be easily generated by a network including a look-up table responsive to an input signal indicative of pixel positions and a subtractive combination circuit. The pixel positions in the overlap area between marginal and center portions are known, and the look-up table can be programmed accordingly to provide output values from 0 to 1 corresponding to the weight function W in response to the input signal. The input signal can be developed in various ways, e.g. By a counter synchronized with each picture line sync pulse. The complementary weight function (1-W) can be obtained by subtracting the weight function W from 1.

FIG. 22 shows an arrangement that can be used as the image converter 17c for the YF signal in FIG. 1a to convert this progressive scan format signal to the interlaced format. Figure 22 also shows a diagram of a portion of the progressive scan input signal YF with the sampling patterns A, B, C and X in the plane containing the vertical (V) and temporal (T) dimensions, as also shown in Figure 2a is. The progressive scan signal YF is subjected to a delay of 525 H in elements 2210 and 2212, respectively, to produce relatively delayed scan patterns X and A from the scan pattern B. The sample patterns B and A are summed in an adder stage 2214 before being fed to a frequency bisector network 2216. An output of the network 2216 is subtractively combined in a network 2218 with the scan pattern X to produce the signal YT. This signal is applied to an input of a switch 2220 which is operated at a frequency equal to twice the line scan frequency of the interlaced scan. Another input of the switch 2220 receives the delayed signal YF from the output of the delay element 2210. The output of the switch 2220 is supplied to a dual port memory 2222, which is operated at a write speed of 8χfsc and a read speed of 4χfsc an output to deliver the signals YF 'and YT in the interlaced format.

FIG. 23 shows an arrangement which can be used to implement the image-scanning converters 17a and 17b in FIG. 1a. Referring to Figure 23, the progressive scan format signal IF (or QF) is applied to a 525 H delay element 2310 before being applied to a dual port memory 2312 having a read speed of 4 χ fsc and a write speed of 8 χ fsc operates to produce the IF output signal IF '(or QF') in interlaced format. Also shown in Fig. 23 are waveforms of the progressive scan format input signal having a first line including the scanning patterns C and a second line including the scanning patterns X and the interlace output (the first line having the scanning pattern C stretched by half scanning line frequency). The dual port memory 2312 provides at its output only the first line (scan pattern C) of the input signal in stretched form.

FIG. 24 shows details of the unit 8C of FIG. 1a . The signals X and Z are respectively fed to the address input of an associated nonlinear amplitude compression means 2410 and 2412, respectively. Compressors 2410 and 2412 are programmable read only memories (PROM memories), each of which contains a look-up table of programmed values corresponding to the desired non-linear gamma compression function. This function is shown in FIG. 24 under unit 2412 by a curve showing the output instantaneous values as a function of the input instantaneous values. The compressed signals X and Z from the data outputs of the units 2410 and 2412 are respectively applied to the signal input of an associated multiplying circuit 2414 and 2416, respectively. The reference inputs of the multiplier circuits 2414 and 2416 receive the alternating subcarrier ASC in a mutually shifted by 90 ° phase position (quadrature phase ratio), ie in sine or cosine form. The outputs of the multipliers 2414 and 2416 are combined in a combiner 2420 to produce the quadrature modulated signal M. In the decoder of Figure 13, the compressed signals X and Z are recovered by a conventional method of quadrature demodulation and undergo complementary non-linear amplitude expansion in associated look-up PROM memories programmed with values complementary to the values in the PROM memories. Save 2410 and 2412 are.

Claims (2)

An apparatus for receiving a television signal embodying a large screen image having side area image information and dominant non-area area image information and having a larger aspect ratio than a conventional television image, said signal including a first component having the side area image information timed into an area of space override and excluding the time-delayed information during comprising dominant non-area area image information processed by the image change, characterized by means (1324) for separating currently accessed side area image information and the dominant non-area area image information; first signal processing means (1340) for generating a first output signal in response to the separated side area image information; second signal processing means (1342) for generating a second output signal in response to the separated non-side area image information; and means (1344) for combining the first and second output signals to produce an image display signal. Apparatus according to claim 1 for receiving the television signal further comprising a second component with image help information processed during the image change, the second component modulating a different subcarrier as a chrominance carrier, characterized by means (1324) responsive to the television signal for separation the first component and the modulated subcarrier; Means (1326) for demodulating the separate subcarrier to provide the second component;