FI88349C - Signal process system for wide-screen television with uniform image resolution for center and side bands - Google Patents

Description

1,383.9

Widescreen TV signal processing system with smooth flat and sideband image separation

BACKGROUND OF THE INVENTION The present invention relates to an apparatus for improving the uniformity of spatial separation between the center and side bands of a widescreen image displayed. In particular, this invention relates to an apparatus in a widescreen display system that uses time compression and in-frame signal processing methods.

A conventional television receiver, such as a receiver that conforms to the NTSC broadcast standards used in the United States and elsewhere, has a 4: 3 aspect ratio (the ratio of the width of the displayed image to the height). In the Vii-15 me, there has been an interest in using higher aspect ratios in television systems, such as 2: 1, 16: 9, or 5: 3, because such higher aspect ratios more accurately approximate the aspect ratio of the human eye than the 4: 3 aspect ratio of a conventional television receiver. Video information signals with an aspect ratio of 5: 3 have received special attention because this ratio approximates:; the aspect ratio of the film, and thus such signals can be transmitted and received without losing the image information; However, widescreen television systems that simply transmit signals with increased- · ···. aspect ratio compared to conventional systems are X ·. incompatible with the recipients of the normal aspect ratio. This makes the adoption of wider widescreen systems difficult.

30 It is therefore desirable to have a widescreen display system that is compatible with conventional television receivers; with the companies. One such system is described in U.S. Patent No. 4,782,383, entitled Apparatus for • Processing High Frequency Edge Information in a Widescreen * · "· 35 Television System granted to M.A. Isnard 1.

2 38349 November 1988. It is even more desirable to have a compatible widescreen display system that can improve or expand the resolution of the displayed image, resulting in additional image resolution. For example, such a widescreen EDTV system (extended definition television) may include a device for producing a progressively scanned image. A system of this type is described in the document entitled Encoding for Compatibility and Recoverability in the ACTV System, written by 10 M.A. Isnardi et al., IEEE Transactions on Broadcasting, Vol. BC-33, no. 4, December 1987, ss. 116 -123. The system uses signal time compression and in-frame middle and sideband image information processing.

In accordance with the principles of the present invention, there is described an apparatus for improving systems of the type described to achieve greater spatial separation uniformity between displayed center and sideband information, significantly reducing or eliminating unwanted oblique image noise in the displayed sideband information.

SUMMARY OF THE INVENTION

An apparatus in accordance with the principles of the present invention has been described in connection with a compatible widescreen EDTV-25 television system using time compression and an in-frame signal processing method, e.g., averaging. The widescreen EDTV signal includes a plurality of components including a first main component comprising midband and time compressed sideband information and a second additional component including sideband information. In the main component, only the midband information is subjected to intra-frame processing. The time-compressed sideband information of the main component is not processed in-frame.

3 8 8: τ. 0

In the described preferred implementation of a compatible widescreen EDTV television system according to the invention, the original high definition progressively scanned widescreen signal is encoded to include four components. The four components are processed separately before they are reconnected to a single signal transmission channel.

The first component is a 2: 1 interlaced main signal with a standard 4: 3 aspect ratio. This component 10 includes a center portion of a widescreen signal time expanded to use almost the entire 4: 3 aspect ratio active line time, and sideband horizontal low frequency information compressed with time to the left and right horizontal image over-15 scan areas, where such on the screen of the standard TV. Only the center of this component is subjected to an in-frame averaging above a given frequency.

The second component is an additional 2: 1 interleaved signal containing high frequency information of the left and right sidebands, both of which are expanded to half the active line time. Thus, the expanded sideband information uses substantially the entire active line time. This component is "set on" 25 to use the same time period as the center of the first component, and is taken as an in-frame average.

The third component is an additional 2: 1 interleaved signal obtained from a widescreen signal source containing high frequency horizontal luminance separation information between approximately 5.0 MHz and 6.0 MHz. This component is also "set" to use the same time period as The second and third components, which are subjected to an in-frame averaging, transversely modulate a phase-controlled alternating 4 3 8 3 -: 3 subcarrier coupled to the first component of the in-frame average.

The optional fourth component is a 2: 1 interleaved "auxiliary signal" containing luminance difference information of the temporal field separation 5 to help reconstruct the widescreen display of the missing image information in the EDTV receiver.

In a widescreen EDTV receiver, a compost signal containing the four components described is decoded into four components. The decoded components are processed separately and used to generate a widescreen signal representative of the picture with improved resolution.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 shows a general description of a compatible widescreen EDTV coding system incorporating a device according to the present invention;

Figure 1a shows a detailed block diagram of the coding part of the system to be defined; Figures 1b to 1e contain diagrams to aid in understanding the operation of the system to be defined;

Figures 2-5 include signal binding formats and diagrams to aid in understanding the operation of the system 25 to be defined;

Fig. 13 is a block diagram of a portion of a widescreen EDTV receiver decoding apparatus; and

Figures 6 to 12 and 14 to 24 illustrate aspects of the system to be defined in more detail.

A system designed to transmit wide aspect ratio images, e.g., 5: 3, via a standard transmitter channel, e.g., NTSC, should achieve high quality image display in a widescreen receiver, while greatly reducing or eliminating perceptible distortion in a standard 4: 3 display. ratio display. Use of signal compression technology 5 3 8; the side bands of the image utilize the horizontal overscan area of the screen of the standard NTSC television receiver, but this may occur at the expense of image separation of the side areas of the reconstructed widescreen image. Since time 5 compression results in an expansion in the frequency range, only low frequency components would survive processing a standard television channel using less bandwidth than is required for a widescreen signal. Thus, when the compressed sidebands of a compatible widescreen display signal are expanded in a widescreen receiver, there is a noticeable difference between the center separation of the displayed widescreen image or the high frequency content and the sidebands unless measures are taken to avoid this phenomenon. This observable difference is due to the fact that the low frequency sideband information could be recovered, but the high frequency information would be lost due to the bandwidth limiting effects of the video channel.

In the system of Figure 1, the parts common to the more detailed system of Figure 1a are denoted by the same reference numeral. As shown in Figure 1, the initial widescreen progressive scan signal with left, right, and center band information is processed to form four separate coding components. These four components were described above, and are illustrated in Fig. 1. The processing of the first component (containing the time-expanded center portion information and the time-compressed side portion low frequency information) is such that the resulting luminance bandwidth does not exceed ::: NTSC luminance bandwidth of 4.2 MHz in this example.

· This signal is color-coded in the standard NTSC format, and the luminance and chrominance components of this signal are pre-filtered appropriately (eg using a field comb filter - '/ ·: 35 mia) to provide improved luminance-chrominance - 6 8 8 7, 1 resolution and standard NTSC and widescreen receivers.

Expanding the second component (sideband high frequency information) over time reduces its horizontal bandwidth to about 1.16 MHz. This component does not spatially correlate with the main signal (first component), and certain precautions have been taken to mask its visibility in standard NTSC receivers, as will be discussed.

10 The extended high frequency luminance information content of the third component 5.0 to 6.0 MHz is first shifted down from the frequencies to the frequency range 0 to 1.0 MHz before further processing.

A fourth component (temporal field difference signal) is placed over the constant 4: 3 format to correlate with the main signal component, masking its visibility in the constant NTSC receivers, and is band-limited horizontally to 750 kHz.

As discussed in more detail below, the first, second, and third components are processed by respective in-frame averages 38, 64, and 76 (vertical-temporal-temporal (V-T) filter type) to eliminate V-T crosstalk between the main and auxiliary signal components in the widescreen receiver. From the center band information of the first 25 components, an in-frame average is taken from above about 1.5 MHz. The second and third intra-frame average components, denoted X and Z, are amplitude non-linearly compressed before cross-modulating the 3.108 MHz alternating subcarrier ASC with 30 varying (inverting) field phases in block 80. The modulated signal (M) from the block 80 component (N) in adder 40. The resulting output signal is a 4.2 MHz bandwidth baseband signal (NTSCF) which, together with the 35,750 k-7 component (YTN) from the 35,750 kHz low-pass filtered filter 79, cross-modulates the RF image carrier in the block. 57, thereby obtaining an NTSC-compatible RF signal that can be transmitted to a standard NTSC receiver or a wide-screen progressive scan receiver 5 via a single standard bandwidth transmitter channel.

The use of compression with respect to the time of the first component allows the low frequency sideband information to be compressed into the horizontal 10 overscan area of the constant NTSC signal. The high frequency sideband information of the second component and the high frequency luminance discrimination information of the third component are spectrally controlled together with the constant NTSC signal in the video transmission channel in a manner invisible in the standard receiver. When received with a standard NTSC receiver, only the center band portion of the main signal (first component) is seen. The second and third components may form a low frequency interference pattern that is not; detected at normal viewing distances and normal image control settings. The fourth component is completely eliminated in receivers with Synchronous Video Sensors. In receivers with envelope detectors, the fourth component is processed but not detected because it is correlated with the main signal.

The main signal (component 1) has a constant NTSC active line period of approximately 52 microseconds (ps). Only the high frequency information of this component above about 1.5 MHz is subjected to an in-frame average. The time-compressed sideband low-frequency information of this component is not subjected to the in-frame average process. It has been found that: **. * Such selective in-frame processing of the principal component 35 improves the separation of skewed image information by eliminating undesirable step-like skewed interference that would otherwise be produced in the reconstructed image if the main signal intake.

In this regard, it is noted that the low frequency information of the sideband of the main signal component is compressed with respect to time by a side compression multiplier (SCF) of approximately six. If such time-compressed information is subjected to an in-frame averaging before being expanded at the receiver to reconstruct the image, the reconstructed sideband information would have dashed slashes because the horizontal frequency at which the in-frame averaging of the SC began would be approximately. The skewed image information becomes increasingly distorted ("knurled") as the frequencies at which the in-frame averaging is performed decrease. For example, if the within-frame average of frequencies above 1.5 MHz is taken from the main signal and the sideband low frequency information of component 1 is compressed with time to an SCF of six, the in-frame averaging of the sideband information effectively begins at a much lower 250 kHz frequency (1.5 MHz / SCF). : 25 follows the knurled slashes. Thus, the beveled slashes would be more noticeable in the reconstructed sideband sections. Since component 1 does not take the in-frame average of the time-compressed sideband regions, the entire original frequency band in these regions (0 to 700 kHz) maintains full vertical separation without distortion due to step-like oblique disturbances.

Component 2, which contains the high frequency information of the left and right side bands, is positioned 35 to use the same time period as the center band portion of component 19893-: 9. Thus, the high left and right sideband expands over time to fill the entire center band region, while the active horizontal sweep period of component 2 is approximately 50ps, which corresponds to the horizontal sweep period of the center band portion of component 1. Therefore, the side expansion coefficient (SEF) is about 4.32, compared to about 4.49 SEF, which would be needed to expand the left and right sideband information of component 2 for the full active line time of 52 ps.

Both components 2 and 3 are located in the middle band region due to in-frame processing performed on the main component 1 and the auxiliary components 2 and 3. As will be explained below, in-frame averaging is a process that makes it possible to distinguish the two previously a combined signal component such as a main signal N and a modulated auxiliary signal M in this example. Since the in-frame processing area in component 1 is reduced to include only the 50ps centerband area, the si-20 interleaving of the modulation components 2 and 3 is similarly converted to include only the centerband area.

As mentioned above, component 3 is located. : coincide with the midband period by linearly compressing the time-extended horizontal luminance information to 50 ps. Compression of component 3 with respect to time from 52 ps to 50 ps sacrifices some spatial correlation with respect to major component 1, but more importantly, ensures that there is a similar horizontal '30 separation in the middle and side band regions of the reconstructed image. Although spatial correlation between components 1 and 3 is desirable to mask crosstalk phenomena between the alternate subcarrier and the main signal, the importance of maintaining a complete spatial correlation of component 3 is reduced because the alternate subcarrier already contains uncorrelated information from component 10 "83;" 2 formats. The amount of reduced spatial correlation of component 3 is insignificant and the resulting similar middle and side band separation weighs more. Component 4 has not undergone an in-frame averaging and py-5 causes unchanged including a full 52 ps active drinking time that coincides with the main signal.

In the decoder, as will be discussed in connection with Fig. 13, in-frame processing is performed only with respect to the center band area to separate the signals M and N. Once component M has been demodulated into components 2 and 3, components 2 and 3 are placed in their original time slices, i. to use the entire 52 ps active line period.

Figure Ib illustrates the RF spectrum of the EDTV wide-15 display system under consideration, which contains additional information compared to the RF spectrum of the standard NTSC system. In the spectrum of the present system, the horizontal luminance separation information of the high and ultra-high frequency of the sideband extends approximately 1.16 MHz on both sides of the alternating subcarrier (ASC) frequency of 3.108 MHz. The V-T auxiliary signal information (component 4) extends 750 kHz on both sides of the main carrier image carrier frequency.

The widescreen progressive scan receiver includes a device for reconstructing the original widescreen progressive scan signal. Compared to the standard NTSC signal, the left and right sidebands of the reconstructed widescreen signal have a standard NTSC resolution and a 4: 3 aspect ratio center band with • 30 excellent horizontal and vertical slurry separation, especially in stationary portions of the image.

Two fundamental benefits determine the signal processing method associated with the first,; 35 for forming and processing the second, third and fourth signal components. These benefits are compatibility with existing receivers and reproducibility at the receiver.

Full compatibility requires compatibility between the receiver and the 5 transmitters, so that existing standard receivers can receive widescreen EDTV signals and produce a standard display without special adapters. Compatibility in this sense requires, for example, that the image scan format of the transmitter be substantially the same, or within tolerance, the same as the image scan format of the receiver. Compatibility also means that additional non-standard components must be physically or visibly hidden in the main signal when displayed on standard receivers. To achieve compatibility in the latter sense, the present system uses the following methods to hide additional components.

As discussed above, the sideband lows are physically hidden in the normal horizontal-20 overscan area of the standard receiver. Component 2, which is a low-energy signal compared to the low-band component of the sideband, and component 3, which is a normal low-energy high-frequency separation signal, are amplitude-compressed and cross-modulated on an alternate auxiliary carrier at 3.108 MHz, which is the interleaving frequency halves). The frequency, phase, and amplitude of the alternating subcarrier are selected so that the visibility of the modulated alternating subcarrier signal is reduced as much as possible - 30, e.g., by shifting the phase of the alternating subcarrier from field to field by alternating 180 # from field to field, unlike chrominance from field to field. Although the modulated alternating subcarrier components are completely within the 35 chrominance emission bands (2.0 to 4.2 MHz), the modulated alternating subcarrier components are not detectable because they are displayed as complementary color flicker in the field rate that is not observed in the human eye with normal chrominance saturation levels. Also, the nonlinear amplitude compression of the modulation components before the amplitude modulation preferably reduces the instantaneous amplitude overshootings to an acceptable lower level. Component 3 is spatially correlated with the information of the center portion of component 1 and spatially correlated slightly less with respect to the left and right portions of component 1. This is done with a format ticker, as will be handled.

Component 4, the "auxiliary signal", is also hidden by expanding the center band information over time to match the standard 4: 3 format, spatially correlating component 4 with the main signal. Component 4 is removed in standard receivers with Synchronous Sensors and is hidden in detection in standard receivers with envelope detectors because it is spatially correlated with the main signal.

The reproducibility of components 1, 2 and 3 in a widescreen progressive scan receiver is performed using in-frame processing at the transmitter and receiver. This process relates to elements 38, 64 and 76 in the transmitter system of Figures 1 and 1a, and corresponding elements in the receiver, as will be discussed. Intra-frame averaging is a signal processing technique that prepares two visually correlated signals for interconnection so that they can be efficiently and accurately reproduced afterwards, for example by a field memory device, without VT (vertical-temporal) crosstalk even in the presence of motion-representative signals. . Signal handlers used for this purpose; The type of operation comprises essentially two signal generators. identical on a field basis, i.e. two samples are obtained at field intervals with the same values. Intra-frame averaging is a suitable method to achieve this goal, but other methods can be used. Intra-frame averaging is essentially a linear time-varying digital pre-filtering and post-filtering process that guarantees the recovery of two visually correlated combined signals. Horizontal crosstalk is eliminated with filter-10 divisions between the horizontal pre-filters of the transmitter encoder and the post-filters of the receiver decoder.

Intra-frame averaging is a form of processing paired pixels. The process of intra-frame averaging over the time domain is generally illustrated in Figure 1e, where field pairs are made identical by averaging pixels (A, B, and C, D) that are 262H apart. The mean replaces the original values for each pair. Figure Id illustrates the in-frame averaging process for the system of Figure 1. Starting with components 2 and 3, the pixel pairs (pixels) 262H apart within the frame are averaged, and the average (e.g., X1, X3, and Z1, Z3) replaces / the original pixel values. This V-T averaging · * ·: 25 occurs within the frame and does not exceed the frame boundaries. In the case of component 1, the in-frame averaging is performed only for the midband information above approximately 1.5 MHz so as not to affect the lower frequency vertical separation information. In the case of component-30 paths 1 and 2, in-frame averaging. is performed on a combined signal including luminance (y) and chrominance components (c) over the entire chrominance band. The chrominance component of the combined signal is retained in the intra-frame averaging because the pixels at the 262H end are "in the same phase as the color subcarrier." The lie of the new alternating subcarrier is controlled so that it is in exactly the opposite lie with pixels 262H apart, and is in a different lie than the lie of the chrominans-5 porcine carrier. Thus, when components 2 and 3 (after polygon modulation) are summed to component 1 in unit 40, the shape of the pixels spaced 262H apart is (M + A) and (M - A), where M is a sample of the combined main signal 1.5 MHz above, and A is a sample of the modulated additional 10 signal.

By taking the in-frame average, the V-T crosstalk virtually disappears, even with the e-birth of the movement. In this regard, the in-frame averaging process produces identical samples 262H from the other. In the receiver, it is a simple matter to return exactly the information content of these samples, i. without crosstalk, by processing pixel samples that are 262H apart in the frame, as will be processed, recovering the information of the main and auxiliary signals 20. In the receiver decoder, the original information of the in-frame average can be restored o-substantially intact by in-frame processing because the original highly visibly correlated information is made substantially similar to the field-'25 field.

Also in the receiver, the RF channel is demodulated transversely using a synchronous RF sensor. Component 4 is thus separated from the other three components. In-frame processing is used to separate component 1 from modulated components 2 and 3, and cross-demodulation is used to separate components 2 and 3, as will be discussed with reference to Figure 13.

After the said four components have been restored, the combined signals are NTSC decoded 15 8 8 3 -r 9 and separated into luminance and chrominance components. Inverse override is performed on the lime components to restore the widescreen aspect ratio, and the high side band highs are combined with the low to restore the full side band separation. The extended high-frequency luminance discrimination information is transferred to its original frequency range and summed into a luminance signal that is converted to a progressive scan format using temporal interpolation and an auxiliary signal. The chrominance-10 signal is converted to a progressive scan format using unassisted temporal interpolation. Finally, the progressive scan luminance and chromium signals are converted to an analog form and matrixed to produce RGB color image signals that are sampled on a widescreen progressive scan display device.

Before discussing the compatible widescreen coding system of Fig. 1a, the signal waveforms A and B of Fig. 2 are considered. Signal A is a 5: 3 aspect ratio wide-20 display signal to be converted to a standard NTSC signal with a 4: 3 aspect ratio, as indicated by signal B. The wide display signal A includes a center band portion associated with the primary image information using the period TC, and left and right side band portions · ;;; 25 related to secondary image information using '···' periods TS. In this example, the left and right sidebands have substantially the same aspect ratios that are smaller than the dominant center band centered between them.

: 30 The wide display signal A is converted to an NTSC signal B by compressing the determined sideband information into completely horizontal sweep areas related to time periods TO. The standard NTSC signal has an active line period TA (52.6 microseconds long) containing overwrite periods: TO, display time period TD containing samples 16 16 8 3- :? video information, and a total horizontal line time period TH of 63,556 microseconds. The periods TA and TH are the same for both widescreen and standard NTSC signals. It has been found that almost all consumer televisions for consumption-5 use have an over-sweep period that uses at least 4% of the total active line time TA, i. 2% left sweep on the left and right. At an interleaving sampling rate of 4 x f se (where fse is the frequency of the color subcarrier), each horizontal-10 line period contains 910 pixels (pixels), of which 754 form the image information of the active horizontal line to be displayed.

The widescreen EDTV system is shown in more detail in Figure 1a. Referring to Fig. 1a, 525, the camera 10 progressively scanning 15 lines of line 60 per second produces a widescreen color signal with its R, G, B components and a wide 5: 3 aspect ratio in this example. An interleaved signal source could also be used, but a progressive scan signal source produces better results. The widescreen camera has a higher aspect ratio and higher video bandwidth compared to a standard NTSC camera, with the widescreen camera's video bandwidth being proportional to the product of its aspect ratio and the total number of lines in the frame, among other factors. Assuming a constant speed scan of 25 with a widescreen camera, an increase in its aspect ratio causes a corresponding increase in its video bandwidth, as well as horizontal compression of image information when the signal is displayed on a standard 4: 3 aspect ratio television. For these reasons, it is necessary to convert the widescreen signal to achieve full NTSC compatibility.

The Color Video signal processed by the coding system of Figure 1 includes both luminance and chrominance signal components. The luminance and chrominance signals contain both low and high frequency information, which are referred to in the following description as "low" and "high", respectively.

The wide-bandwidth wide-screen progressive scan color video signals from the camera 10 are matrixed 5 in the unit 12 to obtain the luminance component Y and the color difference signal components I and Q from the color signals R, G, B. , and are converted from analog to digital (binary) format individually by separate analog-to-digital converters (ADCs) in the ADC unit 14 before being individually filtered by separate vertical-temporal (VT) low-pass filters in the filter unit 16, resulting in filtered sig-15 signals. YF, IF and QF. These signals are each of the shape indicated by waveform A in Figure 2. The separate filters are 3X3 linear time invariant filters of the type shown in Figure 10d, as will be discussed. These filters slightly reduce vertical-20 li-temporal separation, especially oblique VT separation, to prevent undesired interleaving interference (such as flicker, jagged edges, and other spurious reproduction phenomena) in the main signal (component 1 in Figure 1) after it is converted. ;;; 25 from progressive sweep to interlaced. Filters maintain an almost full vertical separation image in stationary parts.

The center band expansion factor (CEF) is a function of the difference between the width of the image displayed on the widescreen receiver and the width of the image displayed on the standard 30 receiver. The picture width in a 5: 3 aspect ratio wide screen is 1.25 times the width of a 4: 3 aspect ratio standard screen picture. This factor of 1.25 is the preliminary midband · / *; the expansion factor to be adjusted to take into account: ***: the overscan area of 35 standard receivers and to take into account the 18 3 8 3 n 9 intentional small overlap of the edge areas between the center and side bands, as will be explained. These factors determine the CEF value of 1.19.

The progressive scan signals from the filter circuit 5 16 use a bandwidth of 0 to 14.32 MHz and are converted to 2: 1 interleaved signals from progressive scan (P) to interleaved scan (I) converters 17a, 17b and 17c, respectively, details 23 of which are discussed below. The bandwidth of the output signals IF ', QF' and YF 'from the converters 17a-10 17c is 0 to 7.16 MHz, because the horizontal scan rate for interleaved signals is half the scan rate of the progressive scan signals. In the conversion process, the progressive scan signal is subsampled by taking half of the available pixel samples to produce a 2: 1 interlaced main signal. More specifically, each progressive scan signal is converted to a 2: 1 interlaced format by retaining either odd or even lines in each field and reading the 20 retained pixels at 4 x f se (14.32 MHz).

All subsequent digital processing of the interleaved signals takes place at a rate of 4 x fse.

Circuit 17c also includes an error prediction circuit. One output signal of circuit 17c, YF ', is an interleaved lowest-25 sampled luminance version of the prefiltered progressive scan component. The second output signal (luminance) YT of the circuit 17c contains temporal information obtained from the field separation information of the image and represents a temporal prediction error, or a temporal interpolation error between the actual and predicted values of luminance samples "missing" at the receiver, as will be described. . Prediction is based on the temporal average of the amplitudes of the "previous" and "next" pixels available at the 3535 receiver. The signal YT, a luminance "auxiliary signal", 19 3 8 3 4 9 which directs the reconstruction of the progressive scan signal at the receiver, substantially determines the error that the receiver is expected to make with respect to the moving image signals and allows such an error to be eliminated at the receiver. In the stationary parts of the image, the error is zero and the receiver performs a complete reconstruction. It has been found that a chrominance auxiliary signal is not required in practice, and that a luminance auxiliary signal is sufficient to produce good results because the human eye 10 is less sensitive to the lack of vertical or temporal separation of chrominance. Figure 2a illustrates the algorithm used to generate the auxiliary signal YT.

Referring to Figure 2a, pixels A, X and B in the progressive scan signal use the same spatial point in the image. Black pixels, such as A and B, are transmitted as the main signal and are available at the receiver. A white pixel, such as X, is not transmitted and is predicted by the temporal mean of ke-20 (A + B) / 2. This means that in the encoder, prediction is made for the "missing" pixel X by averaging the amplitudes of the "before" and "after" pixels. elements A and B. The predicted value (A + B) / 2 is subtracted from the actual value X to produce a prediction error signal 25 si corresponding to an auxiliary signal having an amplitude according to the representation X - (A + IS B) / 2. This representation determines the temporal field separation information in addition to the temporal frame average formation. The auxiliary signal is low pass filtered horizontally with a 750 kHz low pass filter and carried as auxiliary signal YT. The band-limiting of the auxiliary signal to 750 kHz is necessary to prevent this signal from interfering with the ____ next lower RF channel after this signal is modulated into an RF image carrier. A similar prediction of the missing pixel X is made at the receiver using the average of samples A and B, and a prediction error of 8888-, 9 is added to the prediction. This means that X is restored by summing the prediction error X - (A + B) / 2 to the temporal mean (A + B) / 2. Thus, the auxiliary signal allows the conversion from interlaced to progressive-5 to its sweep format.

The auxiliary signal produced by the presented temporal prediction algorithm is preferably a low energy signal compared to the prediction signal produced by some other algorithms, such as that used to produce the line difference signal as described by M. Tsinberg in "ENTSC Two-Channel Compatible HDTV System", IEEE Transactions on Consumer Electronics, Vol. . CE-33, No. 3, August 1987, ss. 146 - 153. In the stationary areas of the image, the error energy is zero because the prediction is complete. The energy content of the auxiliary signal indicates whether the video signal contains stationary or moving image information. A situation in which the auxiliary signal is low in energy occurs in stationary or substantially stationary images (such as a news source where the reporter is against a stationary background). The presented algorithm has been found to produce the least undesirable interference after image reconstruction at the receiver, and the auxiliary signal produced by the presented algorithm retains its usefulness after becoming band-limited (filtered) to about 750 kHz. The auxiliary signal produced by the presented algorithm preferably contains zero energy in the presence of stationary image information, and thus the filtering does not affect the auxiliary signal associated with the stationary image. The greatly improved recon- wired widescreen image is obtained even with an auxiliary signal. will not be sent. In such a case, the stationary parts of the image are much sharper than in a standard NTSC image, but the moving parts are slightly "softer" and may have a "pulsating" noise. Thus, the transmitting station does not need to initially transmit the auxiliary signal, but may choose to switch to RF transmission at a later time.

The described temporal prediction method is useful in both progressive scan and interleaved 5 scan systems with larger-than-constant line counts, but works best with a progressive scan source where pixels A, X, and B use the same spatial point in the image, resulting in complete prediction. case-10 sa. Temporal prediction is incomplete even in stationary parts of the image if the original widescreen image comes from an interlaced signal source. In such a case, the auxiliary signal has more energy and causes small disturbances to the stationary parts 15 of the reconstructed image. Experiments have shown that the use of an interlaced signal source produces acceptable results with interference noticeable only upon closer examination, but that the source of the progressive scan signal causes less interference and produces better results. Returning to Fig. 1a, the interleaved widescreen signals IF ', QF' and YF 'from converters 17a to 17c are filtered by horizontal low-pass filters 19a, 19b and 19c, respectively, to produce a signal IF "having a bandwidth of 0 to 600 kHz, a signal QF" having has a bandwidth ·· .: 25 0 to 600 kHz, and a signal YF "having a bandwidth of 0-5 MHz. These signals are then fed to a format encoding process which encodes each of these signals in 4: 3 format with a format encoder associated with a page-center signal. to the separator and processing unit 18. 30 Briefly, the center portion of each widescreen line is expanded with time and placed in a displayed portion of the active line time with an aspect ratio of 4: 3. Over time, the expansion causes the bandwidth to decrease so that the original widescreen interlaced frequencies come. 35 compatible with standard NTSC bandwidth.Side- 22 38 3-, ·? Seats are divided into horizontal frequency bands such that the color component of the I and Q heights has a bandwidth of 83 kHz to 600 kHz (as shown for signal IH in Figure 7) and the luminance component of the Y heights has a bandwidth of 700 kHz to 5.0 MHz (as shown for the signal YH in Figure 7). 6). Sideband low, i.e. the signals YO, 10 and QO formed as shown in Figs. 6 and 7 contain a DC component, and are compressed with time and placed in the left and right 10 horizontal sweep areas in each line. Sideband highs are processed separately. The details of this for-country coding process are immediately followed below.

In considering the following coding details 15, it is also useful to refer to Fig. 1e, which shows the coding process of components 1, 2, 3 and 4 in relation to the displayed middle and sideband information. The filtered interleaved signals IF ", QF" and YF "are processed by a side-center signal separator and processor 18, 20 to give three groups of output signals YE, IE and QE; YO, 10 and Q0; and YH, IH and QH. the array (YE, IE, QE, and YO, 10, QO) is processed to generate a signal that includes a full-bandwidth midband component and sideband luminance low · 25 compressed into horizontal sweep areas.

The third group of signals (YH, IH, QH) is processed to form a signal containing high sideband bands. When these signals are combined, an NTSC-compatible widescreen signal with a 4: 3 aspect ratio is generated. Details of the silicon: 30 forming the unit 18 are shown and discussed with reference to Figures 6, 7 and 8.

The signals YE, IE and QE contain complete center band information and have the same format as the signal YE in Fig. 3 shows. Briefly, the signal YE 35 is obtained from the signal YF "as follows. The wide display signal 23 383-;? YF" contains pixels 1 to 754 that occur during the active line period of the wide display signal, including side and center band information. The wideband bandwidth information (pixels 75 to 680) is separated as a 5 bandband luminance signal in the YC time demultiplexing process. The signal YC is expanded with respect to time by a midband expansion factor of 1.19 (i.e., 5.0 MHz + 4.2 MHz) to produce an NTSC-compatible midband signal YE. The signal YE has an NTSC compatible bandwidth of 10 (0-4.2 MHz) due to time expansion by a factor of 1.19.

The signal YE uses the image display period TD (Fig. 2) between the over-sweep areas TO. The signals IE and QE are formed from the signals IF "and QF", respectively, and are processed in the same manner as the signal YE.

15 Signals YO, 10, and QO provide low frequency sideband information ("low") located in the left and right horizontal sweep areas. The signals YO, 10 and QO have the same format as indicated by the signal YO in Fig. 3. Briefly, the signal YO 20 is obtained from the signal YF "as follows. The wide display signal YF contains left band information related to pixels 1 to 84 and right band information related to pixels 671-754. As will be discussed, the signal YF "is low pass filtered to produce a luminance low -25 signal with a bandwidth of 0 to 700 kHz, from which signal the left and right sideband low signals are separated (signal YL 'in Fig. 3) by a demultiplexing process. The luminance low signal YL 'is compressed with time to obtain a page ---- 30 band low signal YO containing compressed Mata charge frequency information in the overscan areas associated with pixels 1 to 14 and 741 to 754. The compressed low signal has an expanded bandwidth that is proportional to the amount of time compression. Signals 10 and 24 8 8 349 QO are formed from signals IF "and QF", respectively, and are processed in the same manner as signal YO.

The signals YE, IE, QE and YO, 10, QO are combined by a side-to-center signal combiner 28, e.g., a time multiplexer 5, to produce signals YN, IN and QN with an NTSC-compatible bandwidth and a 4: 3 aspect ratio. These signals are in the form of the signal YN shown in Fig. 3. The combiner 28 also includes suitable signal delays to equalize the transmission times of the signals to be combined. Such signal delays are also included elsewhere in the system as needed to equalize signal transmission times.

The modulator 30, the bandpass filter 32, the HVT bandpass filter 34, and the combiner 36 form the encoder 31 of the enhanced NTSC signal. to produce. The modulator 30 is of conventional design and will be described with reference to Figure 9. The modulated signal CN is bandpass filtered in the vertical (V) and temporal (T) dimensions by a two-dimensional (V-T) filter 32 which removes crosstalk interference in the interleaved chrominance signal before it is applied to the chrominance of the combiner 36. 25 input signal as signal CP. The luminance signal YH is bandpass filtered in the horizontal (H), vertical (V) and temporal (T) dimensions by a three-dimensional H-V-T band-stop filter 34 before being applied as a signal to the luminance input of the YP combiner 36. Filtering of the Lu-30 minance signal YN and the chrominance color difference signals IN and QN ensures that the luminance-chrominance crosstalk is significantly reduced in the subsequent NTSC coding. Multidimensional spatial-temporal filters such as the H-V-T filter 34 and the V-T filter 32 in Figure 1 have the structure of Figure 10, which will be discussed next.

The H-V-T bandpass filter 34 in Fig. 1a has the structure of Fig. 10b and removes upwardly moving oblique-5 frequency components from the luminance signal YN. These frequency components are similar in appearance to the chrominance subcarrier components and are removed to make a gap in the frequency spectrum in which the modulated chrominance is placed. Removal of upwardly moving oblique frequency components from the luminance signal YN does not visibly degrade the displayed image, as it has been found that the human eye is substantially insensitive to these frequency components. The cut-off frequency of the filter 34 is approximately 1.5 MHz so as not to degrade the vertical luminance discrimination information.

The VT bandpass filter 32 reduces the chrominance bandwidth so that the modulated chrominance sideband information can be placed in an aperture formed in the luminance spectrum by the filter 34. The filter 32 reduces the vertical and temporal separation of the chrominance information so that static and moving , but this phenomenon has little or no significance due to the insensitivity of the human eye to such a phenomenon.

The outgoing mid / low signal C / SL combiner 36 includes NTSC-compliant information to be displayed as obtained from the center band of the widescreen signal, as well as compressed side band low (both luminance and chrominance) obtained from the side bands of the wide screen. in the left and right horizontal sweep areas not visible to the viewer of the NTSC receiver screen. The compressed sideband low in the overscan area represents one component in the widescreen sideband information. The second component, high sideband, is formed by processor 18, as will be discussed below. The sideband high signals YH (luminance high), IH (I high), and QH (Q high) are depicted in Figure 4. Figures 6, 7, and 8 show an apparatus for generating these signals, as will be processed. In Fig. 4, signals YH, IH, and QH include left-band high-frequency information related to left-band pixels 1 to 84, and right-band high-frequency information related to right-band pixels 671 to 754.

The middle band portion of the signal C / SL is processed by an adaptive in-frame processor 38 to produce a signal N which is applied to the input of adder 40. The in-frame processed signal N is substantially similar to the signal C / SL due to the visual correlation of the in-frame image information of the signal C / SL. The average circuit 38 averages the signal C / SL from approximately 1.5 MHz above and helps to reduce or eliminate vertical-temporal crosstalk between the main and auxiliary signals. The high pass frequency range of 1.5 MHz and 20 above which the in-frame average circuit 38 operates was chosen to ensure that full in-frame averaging is performed for information at 2 MHz and above to prevent degradation of the luminance vertical separation information in the in-frame center. 25 in the value process. Horizontal crosstalk is eliminated in the 200 kHz guard band between the filter of the in-frame average circuit 38 in the encoder 31 and the filter of the in-frame processor unit in the de-encoder of Fig. 13. Figure 11b shows details of the high frame in-frame processor 38. Figures 11b and 13 will be discussed later.

The signals IH, QH and YH are placed in the NTSC format by an NTSC encoder 60 similar to the encoder 31. More specifically, the encoder 60 includes a device of the type • 35 shown in Fig. 9, as well as a device for cross-modulating the sideband chrominance. high information for sideband luminance high information at 3.58 MHz to produce a signal NTSCH, NTSC format sideband high information. This sig-5 signal is depicted in Figure 5.

The use of multidimensional bandpass filtering in NTSC encoders 31 and 60 preferably allows the separation of luminance and chrominance components with virtually no crosstalk at the receiver when the receiver includes complementary multidimensional filtering to separate luminance and chrominance information. The use of complementary filters for luminance / chrominance coding and decoding is called co-processing and is described in detail in an article by 15 CH Strolle entitled "Cooperative Processing for Improved Chrominance / Luminance Separation" in SMPTE Journal, Vol. 95, No. 8, August. 1986, pp. 782-789. Even standard receivers using conventional suction and line comb filters 20 benefit from the use of multidimensional prefiltration in the encoder which reduces chrominance / luminance crosstalk.

The NTSCH signal is expanded with time by the unit 62 to produce an expanded height signal ESH; 25 having a 50 ps Active horizontal line period, i.e., less than a standard NTSC active line period of approximately 52 ps. More specifically, as shown in Figure 5, the expansion is performed by a "overlay" process that places the left sideband of the NTSCH signal: pixels 1-84 in the pixel positions 15-377 of the ESH signal, i.e., the high left side of the NTSCH signal is expanded to use approximately half the ESH line time of the signal. The right sideband portion of the NTSCH signal (pixels 671 to 754) is processed in the same manner. The time-expanding process reduces the horizontal bandwidth (compared to the NTSCH bandwidth) of the information constituting the signal ESH by a factor of 363/84. The placement process in which the expansion with respect to time is performed can be carried out with a device of the type shown and discussed with reference to Figures 12-12d. From the signal ESH, an in-frame average is taken by a circuit 64 of the type shown in Fig. 11a to produce the signal X described in Fig. 5. The in-frame average signal X is substantially similar to the ESH signal due to the strong visual correlation of the in-frame image information of the ESH signal. The signal X is applied to the signal input of the transverse modulator 80.

The signal YF 'is also filtered by a horizontal bandpass filter 70 with a passband of 5 MHz to 6.0 MHz. The output signal of filter 70, high in horizontal luminance, is applied to amplitude modulator 72, where it amplitude modulates the 5 MHz carrier signal fc. The modulator 72 has an output low pass filter having a cut-off frequency of approximately 1.0 MHz to provide a signal in the passband from 0 to 1.0 MHz to the output of the modulator 72. The Y-lower (false repetition) sideband (5.0 to 6.0 MHz) produced by the modulation process is removed with a 1.0 MHz low-pass filter. The high frequencies of horizontal luminance in the range of 5.0 MHz to 6.0 MHz have been effectively shifted to the range of 0-1.0 MHz as a result of the amplitude modulation process and subsequent low-pass filtering. The amplitude of the carrier should be large enough to preserve the original signal amplitudes when filtering with a 1.0 MHz low-pass filter. 30 mella have been completed. This means that a frequency shift is produced without affecting the amplitude.

The frequency shifted horizontal luminance high signal from unit 72 is encoded (compressed with time) by format encoder 74. This means that encoder 74 encodes the frequency shifted horizontal 29 8 8 34 9 high luminance highs such that this signal has an active line period 50 ps, less than constant Nts 52.6 ps, using the methods discussed in connection with Figures 6-8. When the input signal 5 to the encoder 74 is compressed with respect to time by the encoder 74, its bandwidth increases from approximately 1.0 MHz to 1.1 MHz at the output of the encoder 74. The signal from encoder 74 is adaptively processed in-frame by device 76, as previously discussed, in the same manner as illustrated in Figure 11a 11a before being applied to unit 80 as signal Z. The in-frame average signal Z is substantially similar to the signal from encoder 74. high visual correlation of image information. The modulation signal X, the combined signal containing the luminance and chrominance information, and the modulation signal Z have substantially the same bandwidth, approximately 0-1.1 MHz.

As will be discussed with reference to Figure 24, unit 80 performs amplitude compression of the nonlinear gamma function on the large amplitude deviations of said two additional signals X and Z before these signals cross-modulate the alternating subcarrier signal ASC. A gamma value of 0.7 is used, raising the abso-25 lute value of each sample to a power of 0.7 and multiplying by the sign of the original "·" sample value. Gamma compression reduces the visibility of potentially interfering high amplitude deviations of modulated signals in existing receivers, and allows predictive recovery - in a widescreen receiver, because the inverse function of the gamma function used in the encoder is predictable and can be easily implemented in the receiver decoder.

The amplitude compressed signals are then cross-modulated with 3.1075 MHz phase-controlled alternating auxiliaries. on the carrier ASC, which is an odd multiple of half of the horizontal line frequency (395 x H / 2). The lie of its alternating subcarrier is set to alternate 180 ° from one field to another, unlike the chrominance subcarrier phase.

The field-varying phase of the alternating subcarrier allows the additional modulation information of the signals X and 2 to overlap with the chrominance information and produces phase-complementary additional information components of the additional modulated signal A1, -A1 and A3, -A3, which simplifies the reception -you. The cross-modulated signal M is summed to the signal N in adder 40. The resulting signal NTSCF is a 4.2 MHz NTSC compatible signal.

The described nonlinear gamma function used in the encoder for high amplitude compression is an integral part of a nonlinear compression system (compression expansion) which also includes a complementary gamma function in the widescreen receiver decoder 20 for amplitude expansion, as will be discussed later. The described nonlinear companding system has been found to significantly reduce the effect of non-standard additional information on standard information, without causing visible image degradation due to interference phenomena. The compression system uses a nonlinear gamma function to momentarily compress high amplitude deviations in the non-standard widescreen high frequency supplementary information in the encoder, and a complementary nonlinear gamma function is used to expedite such high frequency information in the decoder, respectively. The result is a reduction in the amount of interaction in the existing standard video information caused by the high-amplitude high-frequency supplementary information in the described compatible widescreen system, in which the non-standard widescreen supplementary information is divided into low and 31 38 3; high frequency parts that are companded. In the decoder, the nonlinear amplitude expansion of the compressed high frequency information does not cause excessive detectable interference, because the high amplitude high frequency information 5 is typically associated with high contrast image edges, and the human eye is insensitive to interference at such edges. The described companding process also advantageously reduces the cross-modulation components between the alternating and chrominance subcarrier, which is also associated with a reduction in the visible inter-10 interference components.

The luminance separation signal YT has a bandwidth of 7.16 MHz and is encoded in a 4: 3 format by a format encoder 78 (e.g., as shown in Figure 6), and is low pass filtered horizontally to 750 kHz with a filter-79 to produce a signal YTN. The side portions are low pass filtered to 125 kHz before time compression in the low pass filter of the input signal of the format encoder 78, which corresponds to the input filter 610 of the apparatus shown in Figure 6, but at a cutoff frequency of 125 kHz. Side high highs are removed. Thus, the signal YTN spatially correlates with the main signal C / SL.

The signals YTN and NTSCF are converted from digital (binary) to analog by the DACs 53 and 54, respectively, before these signals are applied to the RF-25 transverse modulator 57 to modulate the TV RF carrier signal. The RF modulated signal is subsequently applied to transmitter 55 for transmission via antenna 56.

The alternating subcarrier ASC associated with the modulator 80 is synchronized horizontally, and its frequency * 30 is selected to provide sufficient separation of side and center information (e.g., 20-30 dB), and to be a negligible constant with respect to the image displayed on the NTSC receiver. The ASC frequency should preferably be an interleaving frequency that is an odd multiple of half of the horizontal line rate so that it does not cause an interaction that would degrade the quality of the displayed image.

Transverse modulation, such as that performed by unit 80, preferably allows two narrowband signals to be transmitted simultaneously. Expanding the modulating high signals with time results in a reduction in bandwidth that is compatible with the narrowband requirements for transverse modulation. The more the bandwidth is reduced, the less likely it is that the interaction between the carrier and the modulating signals will follow. Further, the high-energy DC component of the sideband information is typically compressed into the overscan range, rather than being used as a modulating signal. Thus, the energy of the modulating signal, and therefore the potential interaction of the modulating signal, is greatly reduced.

The encoded NTSC-compatible widescreen signal transmission by antenna 56 is intended to be received by both NTSC receivers and widescreen receivers, as shown in FIG.

In Fig. 13, a radio-compatible widescreen EDTV interlaced television signal is received by the antenna 1310 and applied to the antenna input of the NTSC receiver 1312. Receiver 1312 processes the compatible widescreen signal in the normal manner to produce a 4: 3 aspect ratio image display, the widescreen sideband information being partially compressed (i.e., "low") into horizontal over-scan areas invisible to the viewer, and partially (i.e., "modulating"). which does not interfere with the operation of the standard receiver.

The compatible widescreen EDTV signal received by the antenna 1310 is also applied to a widescreen progressive scan receiver 1320 capable of displaying a video image having an aspect ratio of, for example, 5: 3.

33 8 8 3 4 9

The received widescreen signal is processed by an input unit 1322 including radio frequency (RF) tuner and amplifier circuits, a synchronous video demodulator (transverse demodulator) that produces a baseband video signal to 5, and analog / digital converter (ADC) signals that produce ) in binary form. The ADC circuits operate at a sampling frequency that is four times the chrominance subcarrier frequency (4 x fsc).

The NTSCF signal is applied to an in-frame processor 1324, which processes the 262H spaced lines within the frames above 1.7 MHz to recover the main signal N and the cross-modulated auxiliary signal M substantially without V-T crosstalk. The 200 kHz 15 horizontal crosstalk guard band is set between the lower limit of the 1.7 MHz operating frequency of the unit 1324 and the lower limit of the 1.5 MHz operating frequency of the unit 38 in the encoder of Figure 1a. The returned signal N contains information substantially similar to the main signal C / SL image-20 formation due to the high visual in-frame image correlation of the original main signal C / SL as a result of the in-frame averaging of the encoder of Figure 1a.

The signal M is coupled to a transverse demodulator and an amplitude expansion unit 1326 for demodulating the additional signals X and Z under the control of a variable subcarrier ASC having a field-varying phase in the same manner as the signal ASC discussed in connection with Fig. 1a. The demodulated signals X and Z contain information substantially visually similar to the image information of the signal ESH and the output signal of the unit 74 of Figure 1a due to the high visual in-frame image correlation of these signals as a result of taking the in-frame average of the encoder of Figure 1a. Unit 35 1326 also includes a 1.5 MHz low-pass filter to remove unwanted 34 3 8 3 -: 9 high frequency demodulation components from twice the variable subcarrier frequency, and an amplitude expander for expoding demodulated signals (previously compressed) using inverse gammaf 0.7 (1.429), i.e. the inverse function of the nonlinear compression function performed by the unit 60 of Fig. 1a.

Unit 1328 compresses the time-color-coded sideband highs to use their original 10 time slices, thus restoring the NTSCH signal. Unit 1328 compresses the NTSCH signal with respect to time by the same amount as unit 62 of Figure 1a expands the signal NTSCH with respect to time.

The luminance (Y) high decoder 1330 decodes the luminance horizontal high signal Z into a widescreen display format by expanding this signal by the same amount as compressing the corresponding component of the encoder of Figure 1a over time, as shown in Figure 17, using the placement methods described herein.

Modulator 1332 amplitude modulates the signal from decoder 1330 on a 5.0 MHz carrier fc. The amplitude modulated signal is later high pass filtered by a filter 1334 having a 5.0 MHz cutoff frequency to remove the lower sideband. The output signal of filter 1334 returns the center. 25 band frequencies 5.0 to 6.0 MHz and the sideband frequencies 5.0 to 6.0 MHz are restored. The signal from filter 1334 is applied; for adder 1336.

The signal from the NTSCH compressor 1328 is applied to the unit 1340 to separate the high luminance from the high chrominance 30 to produce the signals YH, IH and QH. This can be done with the arrangement of Figure 18.

The signal N from the unit 1324 is separated into its constituent luminance and chrominance components YN, IN and QN by a luminance-chrominance separator 1342, which can 35 8 8 3 ,? be similar to the separator 1340, and may have a device of the type shown in Figure 18.

The signals YH, IH, QH and YN, IN, IN, QN are provided as input signals to the Y-I-Q format decoder 1344, which de-5 encodes the luminance and chrominance components into a widescreen format. The sideband lows are expanded with time, the middleband is compressed with time, the sideband highs are summed with the sideband low, and the sidebands are connected to the middleband 10 in the pixel transition region 10 using the principles of Figure 14. Details of decoder 1344 are shown in Figure 19.

The signal YF 'is connected to an adder 1336, where it is summed with the signal from the filter 1334. The extended high-frequency horizontal-15 luminance separation information recovered by this process is summed with the decoded luminance signal YF '.

The signals YF ', IF' and QF 'are converted from interleaved to progressive scan format by converters 1350, 1352 and 1354, respectively. The luminance progressive scan converter 1350 also processes the "auxiliary luminance signal" from the YT format decoder 1360, which decodes the encoded' signal '. Decoder 1360 decodes the signal YTN into a widescreen format, and its structure is similar to that of Fig. 17.

I and Q converters 1352 and 1354 convert signals from interlaced to progressive scan by taking temporal averages of lines one frame apart to produce missing progressive scan line information. This can be done. . 30 with a device of the type shown in Fig. 20.

; The progressive sweep conversion unit 1350 is similar to the unit shown in Fig. 20, except that the signal YT is summed as shown in the arrangement shown in Fig. 21. In this unit 35, the sample of the "auxiliary signal" YT is summed to the temporal mean 36 38 3- :? to assist in the reconstruction of the missing progressive scan pixel sample. Full temporal separation is restored in the horizontal frequency band included in the coded line separation signal (750 kHz, after coding). Above this horizontal frequency band, the signal YT is zero, so the missing sample is reconstructed by taking a temporal mean.

The widescreen progressive scan signals YF, IF, and QF are converted to analog by a digital / ana-10 logic converter 1362 before being applied to a video signal processor and matrix amplifier unit 1364. The peak signal control level, , and other conventional video signal processing circuits. The matrix amplifier 1364 combines the luminance signal YF with the color difference signals IF and QF to produce signals R, G and B representing the color image. These color signals are amplified by display card amplifiers in unit 1364 to a level adapted to directly control a widescreen color image display device 1370, e.g., a widescreen CRT.

Fig. 6 illustrates an apparatus included in the processor 18 of Fig. 1a for generating signals YE, YO and YH from a wideband wide signal YF. The signal: 25L YF "is low pass filtered horizontally by an input filter 610 having a cutoff frequency of 700 kHz to produce a low frequency luminance signal YL which is applied to one input of the reducing combiner 612. The signal YF" is applied to the second input of the combiner 612 and demultiplexed in time. after being delayed by unit 614 to compensate for the signal processing delay of filter 610. Combining the delayed signal YF "and the filtered signal YL produces a high frequency luminance signal YH at the output of the combiner 612.

37 3 8 3- ,?

The delayed signal YF "and the signals YH and YL are applied to different inputs of a demultiplexer 616 including demultiplexing units (DEMUX) 618, 620 and 621, respectively, for processing signals YF", YH and YL. The details of the de-5 multiplexer 616 will be discussed with reference to Figure 8. The demultiplexers 618, 620 and 621 receive the full bandwidth midband signal YC, the sideband high signal YH and the sideband low signal YL ', respectively, as illustrated in Figures 3 and 4.

The signal YC is expanded with time by a time expander 622 to produce a signal YE. The signal YC is expanded with respect to time by a midband expansion factor sufficient to leave space in the left and right over-15 scan areas. The center band expansion factor (1.19) is the ratio of the intended width of the signal Y (pixels 15 to 740) to the width of the signal YC (pixels 75 to 680), as shown in Fig. 3.

The signal YL 'is compressed by a side compression factor of 20 in a time compressor 628 to produce a signal YO. The lateral compression coefficient (6.0) is the ratio of the width of the corresponding part of the signal YL '(e.g. left pixels 1-84) sig-. the reference YO (e.g., left pixels 1-14) refers to that width, as shown in Fig. 3. Aikaekspan-. The processors 622, 624 and 626 and the time compressor 628 may be of the type shown in Figure 12, as will be discussed.

The signals IE, IH, 10 and QE, QH, Q0 are generated from the signals IF "and QF", respectively, in a manner similar to that in which the signals YE, Y and YO are generated by the device of Fig. 6. In this regard, reference is made to Fig. 7, which illustrates an apparatus for generating signals IE, IH and 10 from the signal IF ". The signals QE, QH and QO are generated from the signal QF" in the same manner.

38 3 8 33

In Fig. 7, the wideband wideband signal IF ", after being delayed by the unit 714, is connected to a demultiplexer 716 and is also reduced to a low frequency signal IL from a low pass filter 710 5 in a subtracting combiner 712. 720 and 721 associated with demultiplexer 716 to produce signals IC, IH and IL '. The signal IC 10 is time-expanded by an expander 722 to produce a signal IE, and the signal IL 'is time-compressed by a compressor 728 to produce a signal 10. The signal IC is expanded with a center part expansion coefficient similar to that used for signal YC, as discussed, and the signal IL 'is compressed with a side compression coefficient similar to that used for signal YL', as discussed.

Fig. 8 illustrates a demultiplexer 816 such as may be used in the device 616 of Fig. 6 and the device 716 of Fig. 20 7. The device of Fig. 8 is illustrated in connection with the demultiplexer 616 of Fig. 6. The input signal YF "contains 754 pixels defining image information. Pixels 1-84 define the left band, pixels 671-754 define the right band, and pixels 75-680 define the center band, which slightly overlaps the left band. and with the right band. The signals IF "and QF" have a similar overlap. As will be discussed, such a band crossing has been found to make it possible to combine (connect) the center and side bands at the receiver to substantially eliminate boundary interference.

The demultiplexer 816 includes first, second, and third demultiplexer units (DEMUX) 810, 812, and 814, respectively, associated with the left, center. 35 part and right band information. In each demulti- 39 3 8 3 -v? the plexiglass unit has an input "A" to which the signals YH, YF "and YL are applied, respectively, and an input" B "to which the blackout signal (BLK) is applied.The blackout signal may be a logic 0 level or ground, for example. 5 from the input signal YH of the left and right highs as long as the signal selection input (SEL) of the unit 810 receives the first control signal from the number comparator 817 indicating the image of the left sideband pixel elements 1 to 84 and the right band -10 the presence of element elements 671 to 754. At other times, the second control signal of the number comparator 817 causes a BLK signal at input B rather than a signal YH at input A to be connected to the output of unit 810. Unit 814 and number rate comparator 820 operate similarly to produce a sideband 15 low signal YL ' Unit 812 connects the signal YF "from its output A to its output to produce a midband signal YC only when n control signal from the number comparator 818 indicates the presence of midband pixels 75 to 680.

The number comparators 817, 818 and 820 are synchronized to the video signal YF with a "pulse output signal from the counter 822, which is controlled four times by a chrominance subcarrier frequency (4 x fse) clock signal, from the incoming video signal" H Each output pulse from counter 822 corresponds to a pixel position on a horizontal line. Counter 822 initially has a transition number value of -100 corresponding to 100 pixels counting from the beginning of the horizontal synchronization pulse at time TH3 horizontal; 30 to the end of the blackout period, at which point the pixel 1 appears at the beginning of the horizontal line display period. Thus, counter 822 has a number "1" at the beginning of the line display period. Other counter systems may also be formed. The principles used in the demultiplexer 816 may be. ’. 35 also applies to a multiplexing device performing an opposite signal combining operation, such as that performed by the side-center band combiner 28 in Fig. 1a.

* 0 883-: 9

Fig. 9 shows details of the modulator 5 in the encoders 31 and 60 of Fig. 1a. In Fig. 9 the signals IN

and QN occur four times at the chrominance subcarrier frequency (4 x fsc) and are applied to the signal inputs of latches 910 and 912, respectively. The latch circuits 910 and 912 also receive 4 x fsc clock signals to input the signals IN 10 and QN, and a 2 x fsc switching signal applied to the inverting switching signal input of the latch circuit 910 and the unmatched switching signal input of the holding circuit 912. The signal outputs of the holding circuits 910 and 912 are connected to one output line, where the signals I and Q occur alternately and are applied to the signal inputs of the non-inverting holding circuit 914 and the inverting holding circuit 916. These latches are clocked at a rate of 4 x fsc and receive a switching signal at the chrominance subcarrier frequency fsc to the inverting and non-inverting inputs, respectively. The non-inverting latch 914 produces a sequence of output I and Q signals of varying positive polarity, and the inverting latch 916 produces a sequence of I and Q signals of varying negative polarity, i. -I, -Q. The output signals 25 of the latch circuits 914 and 916 are combined into a single output line having an alternating sequence of paired I and Q signals having opposite polarities, i. I, Q, -I, -Q ... etc, which form the signal CN. This signal is filtered by a filter 32 prior to combining in unit 36 with a filtered version of the snow-30 nance signal YN to produce an NTSC-encoded signal C / SL in the form Y + I, Y + Q, YI, YQ, Y + I, Y + Q .. . and so on.

Figure 10 illustrates a vertical-temporal (V-T) filter that may have V-T bandpass, V-T bandpass-35, or V-T low-pass configurations for adjusting the weight 4i 383-,? coefficients ai - a9. The table in Figure 10a illustrates the weighting factors associated with the V-T bandpass and bandpass filter configurations used in the described system. An HVT bandpass filter, such as the filter 34 of Fig. 1a, and HVT bandpass filters, such as those included in the decoder system of Fig. 13, form a combination of a horizontal low-pass filter 1020 and a VT bandpass filter 1021, respectively, as shown in Fig. 10b. a combination of a bandpass filter 1031, as shown in Figure 10c.

In the H-V-T bandpass filter of Fig. 10b, a horizontal low-pass filter 1020 is given a cutoff frequency and produces a filtered low frequency signal component. This signal is degressively combined in combiner 1023 with a delay version of the signal from delay unit 1022 to produce a high frequency signal component. The low frequency component is delayed by one frame on circuit 1024 before being applied to sum-20 but to combiner 1025 to produce an H-V-T bandpass filtered output signal. The V-T filter 1021 has the coefficients of the V-T bandpass filter shown in Fig. 10a. The H-V-T - ·· '· bandpass filter, as included in the decoder of Fig. 13: V, is shown in Fig. 10c, including the horizontal V ·; A 25-way bandpass filter 1030 with a given ·: * cutoff frequency coupled in series with a V-T bandpass filter 1031 having V-T bandpass filter coefficients, as shown in the table of Figure 10a.

The filter of Figure 10 includes a plurality of memory units (M) 1010a to 1010h connected in series to produce successive signal delays to the respective outputs t1 to t9, and to produce the total filter delay. The signals carried by the outputs are applied to one input of the multipliers 1012a to 10121, respectively. The second input of each multiplier corresponds to /. 35 newly receives a given weighting ai - a9 depending on 42 8 8 3 ,? the nature of the filtering process to be performed. The nature of the filtering process also determines the delays caused by the memory units 1010a to 1010h. The horizontal dimension filters have pixel storage memory elements such that the total delay of the 5 filters is less than the time period (1H) of one horizontal image line. The vertical dimension filters have line recording memory elements alone, and the temporal dimension filters have frame storage memory elements alone. Thus, the H-V-T three-dimensional filter contains pixel (<1H), line (1H), and frame storage elements (> 1H), while the V-T filter contains only the latter two types of memory elements. The weighted branched (mutually delayed) signals from elements 1012a to 1012i are combined in adder 1015 to produce a filtered output signal.

Such filters are non-recursive, finite impulse response (FIR) filters. The nature of the delay produced by the memory elements depends on the type of signal to be filtered and the amount of crosstalk that can be allowed between the high luminance, chrominance and sideband high signals in this example. The sharpness of the cutting properties of the filter is improved by increasing the number of memory elements connected in series.

Fig. 10d illustrates one of the separate filters 25 of Fig. 1a in Fig. 1a, which includes memory units (delay units) 1040a to 1040d connected in series, associated multipliers 1042a to 1042e with corresponding weighting factors a1 to a5, for receiving signals from signal outputs t1 to t5, and a signal combiner 1045 that sums the weighted output signals from the multipliers a1 to a5 to produce an output signal.

Fig. 11a shows an in-frame averaging circuit suitable for use as the in-frame averaging circuits 64 and 76 of Fig. 1a. The incoming combined vi-35 deosignal signal is applied to a delay circuit containing 262H delay 43 38 3 · '? water elements 1102 and 1104, and also a multiplexer (MUX) 1115, which is switched at a field rate controlled by a 30 Hz switching signal. The 30 Hz MUX switching signal is vertically synchronized under the control of vertical sync-5 pulses associated with the incoming combined video signal. The second input of the MUX 1115 receives a signal from the output of the delay element 1112. The combiner 1118 sums the signals from the MUX 1115 and the average output between the delay elements 1110 and 1112, after these signals are weighted by an average factor of%. This weighting factor can be produced by a suitable matrixing circuit inside the combiner 1118, or by signal multipliers located in the input signal paths of the combiner 1118, respectively.

The signals "Y1 + Cl" and "Y2 + C2" are combined color video signals 262H apart in successive first and second image fields, and the signal "M1" is an in-frame average output signal, as exemplified in Fig. 1d. During the first 20 fields, the multiplexer 1115 is in the input position "1" and carries the signal Y2 + C2 to the combiner 1118, where it is summed to the center output signal Y1 + C1 to produce an average output signal M1. In the next field, the center output between delay elements 1110 and 1112 contains a signal value Y2 + C2 and MUX 1115 is in position "2" to select a signal path from the output of delay element 1112 containing signal value Y1 + C1, thus producing an average combiner of the same signal M1. 1118 departures. The described device generates identical pixel pairs 262H from each other. , 30 and is not limited to the use of the averaging process.

Any weighting values can be used to produce the desired weighted combination of pixel pairs, and delays other than 262H can be used (along with the associated change in MUX switching rate) depending on the requirements of the system in question.

44 8 8 3 -?

Fig. 11b shows a frequency-selective in-frame averaging circuit suitable for use as the in-frame averaging circuit 38 of Fig. 1a. Fig. 11b includes the arrangement of Fig. 11a, except that the subtracting signal combination 5 is associated with the combiner 1128 instead of the summing signal combination. thereto. Briefly, the output of combiner 1128 represents the difference between the image fields, rather than the average, as in the arrangement-10 of Figure 11a. This difference is essentially an elimination term summed back to signal Y1 + Cl by combiner 1134 to eliminate differences between successive fields to ensure that the contents of consecutive image fields are averaged in the same manner. Filter 1130 filters out the elimination term from the output of combiner 1128 to limit the averaging process to the desired frequency range. Gate 1132 is controlled to determine when your averaging process occurs in the frame period, in this case in the center band, excluding time-compressed areas of the sidebands.

More specifically, the input signal paths to combiner 1128 have signal weighting factors + ½ and -h, as shown, so that the output signal of combiner 1128 corresponds to the difference between the information content of the input signals of combiner 1128, which input signals are 262H apart in adjacent fields. Complementary weighting factors can be produced by using signal multipliers on the respective input paths of combiner 1128, or by arranging combiner 1128 as a difference amplifier. The output signal of combiner 1128 is filtered by a 1.5 MHz horizontal high pass filter 1130 before being applied to electronic transfer port 1132. Port 1132 responds to the switching control signal by passing the high frequency signal filter 1130 only during the center of the main signal (component 1) 35. Currently gate 1132 is open 45 3 8 3 -: 9 (leading). Gate 1132 is closed (non-conductive) with respect to the time of the main signal during compressed sideband portions, e.g., during the described positive pulse periods of the control signal. The output signal of port 1132 is summed at combiner 5 1134 to the combined video signal present in the center output between delay elements 1120 and 1122. The port control signal is vertically synchronized under the control of the vertical period synchronization pulses associated with the combined video signal. The gate control signal is also horizontally synchronized. Horizontal synchronization can be achieved with a device controlled by the horizontal line synchronization pulse component of the incoming composite video signal, which includes a pixel counter for determining the timing of the positive pulse components of the port control signal that follow each horizontal line synchronization pulse. A predetermined period of time can be easily set between the horizontal line synchronization pulse and the pixel of the first image.

Referring again to Fig. Id in conjunction with Fig. 20 11b, when the MUX 1125 is in position 1 as shown and the gate 1132 is closed, only the combined video signal Y1 + C1 is delayed by the delay output elements 1120 and 1122; the sample occurs at the output of combiner 1134. Thus, at present, the output signal of the combiner 1134 is the compressed sideband information of the combined video signal Y1 + C1 associated with the field 1. The output signal of the combiner 1134 is the constant compressed sideband information of the combined video signal Y2 + C2 associated with the next image field 2 when the MUX 1125 is in position 2.

When the MUX 1125 is in position 1 for field 1 and the gate 1132 is closed during the middle ν ': band period between the sideband periods, the output signal of the combiner 1134 includes the signal components Y1 + Cl and M1. The component Y1 + Cl ____ contains unchanged, i.e. not within the in-frame average, mid-band information at approximately 1.5 MHz 46 3 8 3 · :? and below. Component M1 contains the in-frame average midband information above approximately 1.5 MHz. When MYX 1125 is in position 2 during the next field 2 and gate 1132 is closed to the center band period channel ai-5, the output signal of combiner 1134 includes an intra-frame average component M1, as discussed above, and component Y2 + C2. The latter component contains unchanged (non-in-frame average) midband information at approximately 1.5 MHz and below-10.

Figure 12 illustrates a raster placement device that can be used for the time expander and compressors of Figures 6 and 7. In this regard, the waveforms are considered in Figure 12a, which illustrates the placement process. Fig. 12a 15 shows a waveform S of an input signal having a center portion between pixels 84 and 670 to be placed in pixel locations 1 to 754 of the outgoing waveform W by a time expansion process. The endpoint pixels 1 and 670 in waveform S are located directly as the endpoint pixels 1 and 754 of waveform 20. The intervening pixels are not placed directly on a 1: 1 basis due to expansion over time, and in many cases are not placed on an integer basis. The latter case is described when, for example, the pixel position 85.33 of the incoming waveform S corresponds to the position 3 of the outgoing waveform W. Thus, the pixel position 85.33 of the signal S includes a whole part (85) and a fraction DX (0.33). , and the pixel position 3 of the waveform W includes a whole part (3) and a fraction part (0).

In Fig. 12, a pixel counter operating at 4 x fse produces an outgoing WRITE ADDRESS signal M representing the pixel positions of the outgoing raster (1.... 754). The signal M is applied to the PROM (programmable read-only memory) 1212, which contains a look-up table containing the programmed values depending on the raster placement 47 8 8 3- ;? nature, e.g., compression or expansion. Controlled by the signal M, the PROM 1212 produces an outgoing READ ADDRESS signal N representing an integer and an outgoing signal DX representing a fraction greater than or equal to zero but less than one. In the case of a 6-bit signal DX (26 = 64), the signal DX has fractions 0, 1/64, 2/64, 3/64 ... 63/64.

The PROM 1212 allows the video input signal S as a function of expansion or compression of the signal N from the stored value-10. Thus, the programmed value of the READ ADDRESS signal N and the programmed value of the fractional signal DX are produced under the control of the total values of the pixel position signal M. To achieve signal over-expansion, for example, the PROM 1212 is arranged to produce a signal N at a slower rate than the signal M rate. Otherwise, to achieve signal compression, the PROM 1212 produces a signal N at a faster rate than the signal M rate.

The video signal S is delayed by the pixel delay elements 1214a, 1214b and 1214c connected in series to produce video-20 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 the video signal inputs of the respective dual port memories 1216a to 1216d, as is known. The signal M is applied to the write address input 25 of each memory 1216a to 1216d, and the signal N is applied to the read address input of each memory 1216a to 1216d. The signal M determines where the incoming video signal information is written to the memories, and the signal N determines which values are read from the memories. Memories can write to one address at the same time as they read from another address. The output signals S (N1), S (N), S (N + 1) and S (N + 2) from memories 1216a to 1216d have a time-expanded or time-compressed format depending on the read / write operation of memories 1216a-1216d, which is a function of how PROM 35 1212 is programmed.

48 8 8 3 -r 9

The signals S (N1), S (N), S (N + 1) and S (N + 2) from memories 1216a to 1216d are processed by a four-point linear interpolator including peak value filters 1220 and 1222, PROMrn 1225 and a two-point linear interpo-5 generator 1230 , the details of which are shown in Figures 12b and 12c. Peak filters 1220 and 1222 receive three signals from a group of signals including signals S (N-1), S (N), S (N + 1) and S (N + 2), as shown, and also receive peak signal PX. The PX value of the peak value signal-10 Iin varies from zero to one as a function of the value of the signal DX, as shown in Fig. 12d, and is produced by the PROM 1225 controlled by the signal DX. The PROM 1225 contains a look-up table and is programmed to produce a given PX value under the control of a given DX value.

Peak filters 1220 and 1222, respectively, produce peaks from mutually delayed video signals S '(N) and S' (N + 1) to a two-point linear interpolator 1230, which also receives signal DX. The interpolator 1230 produces (compressed or expanded) 20 video output signals, where the output signal W is defined by the representation W - S '(N) + DX [S' (N + 1) - S '(N)]

The described four-point interpolator and peak value function preferably approximate the (sin X) / X interpolation function 25 with good high-frequency resolution accuracy.

Fig. 12b shows details of the peak value filters 1220 and 1222, and the interpolator 1230. In Fig. 12b, the signals S (N-1), S (N) and S (N + 1) are applied to a weighting circuit 1240 in the peak value filter 1220, where these signals. 30 are weighted by peak value factors of -1/4, 1/2, and -1/4, respectively. As shown in Fig. 12c, the weighting circuit 1240 includes multipliers 1241a to 1241c for multiplying the signals S (N-1), S (N) and S (N + 1) by peak value coefficients of -1/4, 1/2 and -1/4, respectively. Output signals from multipliers 1241a. 35 1241c is summed in adder 1242 by the peak value signal P (N) 49 8 8 7>. 9 multiplied by the signal PX in the multiplier 1243 to produce a peak value signal summed to a signal S (N) in an adder 1244 to produce a peak value signal S '(N). The peak value filter 1222 has a similar structure and operation.

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

Figure 15 shows details of the in-frame processor 1324 of Figure 13. The decoder device of Figure 15 is essentially similar to the code-15 device of Figure 11b.

In Fig. 15, the combined video signal to the processor 1324 includes the signal components "Y1 + Cl" and "M1 + A1" in the first field. In the next second field, the input signal includes the components “Y2 + C2” and “M1-Al.” The components Y1 + Cl, M1, and Y2 + C2, M1 are components produced by the in-frame processor 38. Components + A1 and -AI represent the alternate subcarrier signal modulated component 2 and component 3 include in-frame processed information from units 64 and 76, respectively, to successive fields 25. In this regard, Figures 1, 1a and Figure Id are considered in particular.

The in-frame processor of Figure 15 operates in substantially the same manner as the arrangement of Figure 11b discussed earlier. With MUX 1525 in position 1, the field ·. * 30 separation components are obtained from the output of combiner 1528. After filtering by the high-pass filter 1530 and porting by the unit 1532, a component -AI results in which, combined with the signal Y1 + Cl, M1 + A1 in the combiner 1534, cancels the modulated additional subcarrier component (+ A1) to produce a recovered * · 35 main signal Y1 + Cl, M1. Returned main 3 8 3 -f? the signal component Y1 + Cl is unchanged below the 1.7 MHz cut-off frequency of the high-pass filter 1530, and the component M1 represents the in-frame processed midband information above approximately 1.7 MHz. The field difference elimination term (-AI) is the additional modulated signal A1 after the inversion performed by the unit gain amplifier 1535.

The recovered main signal Y1 + C1, M1 corresponds to the signal N in Fig. 13, and is further processed by the circuit 1342, 10 as discussed. The returned additional signal A1 corresponds to the signal M in Fig. 13 and is demodulated by the circuit 1326.

Fig. 16 illustrates the operation of circuit 1324, as shown in Fig. 15, in the next consecutive image field. In this case, the signal Y2 + C2, M1-A1 is generated between the delay elements 1520 and 1522, and the MUX 1528 uses position 2 to receive the signal Y1 + Cl, M1 + A1. The returned main signal Y2 + C2, M1 is produced at the output of the combiner 1534, and the opposite phase modulated additional signal -AI is restored.

In Fig. 18, an H-V-T bandpass filter 1810 having the configuration of Fig. 10c and a 3.58 ± 0.5 MHz passband passes a signal to the NTSCH reducing combiner 1814, which also receives the signal NTSCH after being passed through the transmission time equalizing delay 1812.

. 25 The high signal YH of the separated luminance signal appears at the output of the combiner 1814. The filtered NTSCH signal from filter 1810 is cross-demodulated by demodulator 1816 under the control of a chrominance subcarrier signal SC to produce chrominance high IH and QH.

In Fig. 19, the signals YN, IN, and QN are separated into compressed sideband low YO, 10, QO, and expanded centerband signals YE, IE, QE by a side-centerband signal separator (time multiplexer) 1940. The demultiplexer 1940 may use Fig. 8, which was processed: 35 previously, demultiplexer 816 principles.

51 R 8 3 -ί?

The signals YO, 10 and QO are time-expanded by a side expansion coefficient (corresponding to the side compression coefficient in the encoder of Fig. 1a) with a time expander 1942 to restore the sideband low initial spatial dependence in the widescreen signal, such as the returned sideband low signals YL. Similarly, to make space for the sidebands, the midband signals YE, IE, and QE are compressed by a time-proportional center compression coefficient (corresponding to the center expansion coefficient 10 in the encoder of Fig. 1a) by a time compressor 1944 to restore the original spatial dependence of the Compressor 1944 and expander 1942 may be of the type shown in Figure 12 discussed earlier.

The spatially restored sideband highs YH, IH, and QH are combined with the spatially restored sideband low YL, IL, and QL by a combiner 1946 to produce reconstructed sideband signals YS, IS, and QS. 20 These signals are coupled to a reconstructed centerband signal with a YC, IC, and QC float from a 1960 fully reconstructed. . to generate a widescreen luminance signal YF 'and fully reconstructed widescreen color difference signals IF' and QF '. The connection of the side and center band signal components is performed in a manner that virtually eliminates the visible seam at the boundary between the center and side bands, as can be seen from the following treatment of the float 1960 shown in Figure 14.

In Figure 20, the interleaved signals IF '(or QF'): 30 are delayed by 263H by element 2010 before being applied to the input of dual port memory 2020. An additional delay of 262H is applied to this delayed signal by element 2012 before it is summed to the input signal in adder 2014. The output signal of adder 2014 is connected to two: 35 divider circuit 2016 before being applied to the input of dual port memory. Memories 2020 and 2018 read data at 8 x f se and write data at 4 x fse. The output signals of the memories 2018 and 2020 are applied to a multiplexer (MUX) 2022 to produce the IF (QF) of the progressive scan signals 5. Also shown are waveforms representing an interlaced input signal (two lines, pixel samples C and X labeled) and a progressive scan output signal comprising pixel samples C and X.

Fig. 21 illustrates an apparatus suitable for use as a converter of the signal YF 'in Fig. 13. The interlaced signal YF' is delayed by the elements 2110 and 2112 before they are combined by an adder 2114, as shown. The delayed signal from element 2110 is applied to both port memory 2120. The output signal of adder 2114 is coupled to two splitter circuits 2116, the output of which is summed to signal YT in adder 2122. The output signal of adder 2118 is applied to dual port memory 2122. Memories 2120 and 2122 write 4 x fsc at 8 x fsc, and provide output signals to multiplexer 2124, which generates a progressive scan signal YF.

Fig. 14 shows a sideband-middle band threading device suitable for use as a threader, for example. 1960 in Fig. 19. Fig. 14 shows an internal circuit 1410 for producing a full bandwidth luminance signal YF 'from the sideband luminance signal component YS and the midband luminance signal component YC, as well as an I signal from the slider 1420 and Q signal p30 , which are similar in structure and operation to the circuit 1410. The center band and the side band are intentionally superimposed on a plurality of pixels, e.g., ten pixels. Thus, the middle and sideband signals have had several common re- 53 883- ,? dundant pixels in the signal encoding and transmission process prior to interconnection.

In a widescreen receiver, the middle and side bands are reconstructed from their respective signals, but due to the time expansion, time compression, and filtering performed on the band signals, several pixels at the side and center band boundaries are corrupted or distorted. The transition areas (OL) and the corrupted pixels (CP; slightly exaggerated for clarity) are indicated by the waveforms associated with the signals YS and YC in Fig. 14. If there were no transition area in the bands, the corrupted pixels would merge and the junction would be visible. A transition area wide of ten pixels has been found to be wide enough to compensate for three to 15 corrupted border pixels.

The redundant pixels preferably allow the pages to match the center bands in the transition region. The multiplier 1411 multiplies the sideband signal YS by the weighting function W in the transition regions, as described by the associated waveform 20, before the signal YS is applied to the signal combiner 1415. Similarly, the multiplier 1412 multiplies. . with a complementary weighting (1-W) function in the transition regions of the midband signal YC, as described by the associated waveforms, before the signal YC is applied to the ·: 25 combiner 1415. These weighting functions have a linear ramp-type characteristic in the transition regions and include 1 values and After weighting, the pages of the middle band pixels are summed by a combiner 1415 so that each reconstructed pixel is linear: a combination of side and middle band pixels.

The weighting functions should preferably approach one at the inner boundary of the transition region, and should approach zero at the outer boundary. This ensures that the damaged pixels have relatively little effect on the reconstructed band boundary. The described linear 54 88 3.1 ramp-type weighting function satisfies this requirement. However, the weighting functions do not have to be linear, and the nonlinear weighting function with its curved or rounded ends, i. near center of gravity 0 and 1 5 may also be used. Such a weighting function can be easily achieved by filtering a weighting function of a linear ramp of the type described.

The weighting functions W and 1-W can be easily formed by a circuit including a look-up table controlled by an input-10 signal representing pixel positions and a subtracting combiner. The transition points of the page-center pixels are known, and the Lookup Table is thus programmed to produce output values of 0-1, corresponding to a weighting function W controlled by the input signal. The input signal can be generated in a number of ways, such as by a counter that is synchronized with each horizontal line synchronization pulse. The complementary weighting function 1 - W can be produced by subtracting the weighting function W from one.

Fig. 22 shows an apparatus suitable for use as a progressive sweep to interleaved sweep 17c of the signal YF in Fig. 1a. Fig. 22 also shows a diagram of a portion of the progressive scan input signal YF in the vertical (V) and temporal (T) planes indicated by samples A, B, C, and X, as shown in Fig. 2a. The progressive scan signal YF is delayed by 525H elements 2210 and 2212 to produce relatively delayed samples X and A from sample B. Samples B and A are summed in adder 2214 before being applied to the two dividing circuits 2216.

: 30 The output signal of circuit 2216 is reduced in circuit 2218 to sample X to produce a signal YT. This signal is applied to one input of switch 2220, which switch operates at twice the rate of interlaced horizontal line sweep. The second input of the switch 2220 receives: 35 a delayed signal YF from the output of the delay 2210. Kyt- 55 3 8 3-i? the output signal of output 2220 is applied to a dual port memory 2222 which reads at 4 x fse and writes at 8 x fse to produce signals YF 'and YT in interleaved form at the respective outputs.

Fig. 23 shows a device suitable for use as converters 17a and 17b in Fig. 1a. In Fig. 23, the progressive scan signal IF (or QF) is applied to the 525H delay element 2310 before being applied to the dual port memory 2312, which reads at 4 x fse and writes at 8 x fse, the interlaced output signal IF '(or QF'). to produce. Also shown are waveforms depicting a progressive scan input signal with the first and second lines associated with samples C and X, and an interlaced output signal (the first line if the sample C is stretched at the H / 2 rate). The dual port memory 2312 transmits only the first line sample (C) of the input signal in stretched form.

Figure 24 shows details of unit 80. Signals X and Z are applied to the inputs of nonlinear amp-20 lithium compressors 2410 and 2412, respectively. Compressors 2410 and 2412 are programmable read only memory devices; - (PROM), each of which contains a look-up table containing programmed values corresponding to the desired nonlinear gamma compression function. This function is described by an immediate input-output response next to unit 2412. The compressed signals from the data outputs of units X10 and Z12 are applied to the signal inputs of signal multipliers 2414 and 2416, respectively. The comparison inputs of the multipliers 2414 and 2416 receive the corresponding alternating subcarrier wave signals ASC with each other in a 90® phase dependence, i. the signals ASC are in blue and cosine form. The output signals of multipliers 2414 and 2416 are summed in combiner 2420 to produce a cross-modulated signal M. In the decoder arrangement of Figure 13, the compressed signals X: 35 and Z are restored by a conventional transverse demodulator; ? and the complementary non-linear amplitude expansion of these signals is performed by associated PROMs whose look-up tables are programmed with values complementary to the values of PROMs 2410 and 2412.

Claims (9)

57 8 8 3- *? A system for processing a signal representative of a television-type widescreen image having sideband information and baseband information and a higher aspect ratio than a standard television image, the device comprising: a device controlled by said representative signal to form a first component containing sideband image information compressed with respect to time into the overscan area of said first component, and comprising baseband image information, characterized in that it comprises: a first means (38) for processing said baseband information within a frame separately from said time-compressed sideband information. The system of claim 1, characterized in that said first device (38) performs in-frame averaging of said main band information separately from said sideband information. The system of claim 1, further characterized in that said first device produces processed in-frame information that includes distinct groups of substantially identical in-frame pixel information. The system of claim 1, further characterized by: comprising: 30 a device controlled by said representative signal to form a second component (X) containing image content information, standard television image information ____; in addition; a second device (64) with said second composite. 35 nents are processed in-frame; and 58 3 8 3- + 9 a device for combining (80, 40) said processed first component (M1; M3) with a signal (M1 ± Al; M3) combined with the second components (A1, -AI; A3, -A3) complementary in phase ± A3) 5 to produce. The system of claim 4, further characterized in that: said first device (38) includes an intra-frame averaging device and generates processed in-frame information that includes separate groups of substantially identical in-frame pixel information; and said second device (64) includes an in-frame averaging device and produces processed in-frame information including distinct groups of substantially identical in-frame pixel information. The system of claim 1, characterized in that said time-compressed image information includes low frequency information separate from the high frequency information. A television receiver for a combined signal according to claim 4, comprising: a device (1324) controlled by said combined signal to separate said first and second components, characterized in that it comprises: a device (1940, 1942) controlled by said separated first component to produce baseband image-30 formation signals and time-expanded sideband image information signals with a substantially reduced amount of oblique interference; and a video signal processing apparatus (1944, 1946, 1960) for rearranging information about said separated main and sideband signals, and said second component describes to produce a representative signal 59 38349 having improved spatial separation uniformity between the main and sidebands. An apparatus for receiving a television-type signal representing a widescreen image having side portion image information and 5 mainband image information and having an aspect ratio greater than a standard television image, said representative signal including a first component including side portion image information compressed over time with said first component overscan area. , and including baseband image information processed in-frame separately from said time-compressed information, the apparatus comprising: a device (1940) for separating said time-compressed side portion image information and said main band image information; a first signal processing apparatus (1944) for producing a first output signal including signals defining the baseband image information; 20, characterized in that it comprises: a second signal processing device (1942) for producing a second output signal containing signals defining the image information of the time-expanded side portion having a substantially reduced amount of oblique interference; and. an apparatus (1960) for combining said first and second output signals to produce a representative signal having improved spatial separation uniformity between the main band and the side portion information. : 30 An apparatus for receiving a television-type signal representing a widescreen image having side portion image information and mainband image information and having an aspect ratio greater than vak ·· than in a standard television image, said representative signal comprising (a) a first component comprising a side portion image information compressed with respect to time into the overwrite area of said first component, and including baseband image information processed in-frame separately from said time-compressed information, and (b) a second component processed in-frame. includes additional image information, said second component modulating an additional subcarrier other than the chrominance subcarrier; the apparatus comprising: a device (1324) controlled by said television type signal to separate said first component 10 and said modulated auxiliary carrier; means (1326) for demodulating said separated auxiliary carrier to produce said second component; characterized in that it comprises: a device (1940, 1942) controlled by said first component to produce a first main band information output signal and a second time expanded side image information output signal having a substantially reduced amount of oblique interference; and a video signal processing apparatus (1960) controlled by said first and second output signals to produce an image representative signal with improved spatial separation uniformity between the main band and the side portion information. 61 3 8 3 -r 9