Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N5/00—Details of television systems
  • H04N5/14—Picture signal circuitry for video frequency region
  • H04N5/21—Circuitry for suppressing or minimising disturbance, e.g. moiré or halo
  • H04N5/211—Ghost signal cancellation
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N5/00—Details of television systems
  • H04N5/76—Television signal recording
  • H04N5/765—Interface circuits between an apparatus for recording and another apparatus
  • H04N5/775—Interface circuits between an apparatus for recording and another apparatus between a recording apparatus and a television receiver

Definitions

• the invention relates to video tape recorders and players, especially cassette types using helical-scan recording and playback- Background of the Invention
• Ghost images caused by multipath reception and commonly referred to as "ghosts", are a common occurrence in television pictures that have been broadcast over the air or have been transmitted by cable.
• the signal to which the television receiver synchronizes is called the reference signal, and the reference signal is usually the direct signal received over the shortest transmission path.
• the multipath signals received over other paths are thus usually delayed with respect to the reference signal and appear as trailing ghost images. It is possible, however, that the direct or shortest path signal is not the signal to which the receiver synchronizes.
• the receiver synchronizes to a reflected (longer path) signal
• there will be a leading ghost image caused by the direct signal or there will a plurality of leading ghosts caused by the direct signal and other reflected signals of lesser delay than the reflected signal to which the receiver synchronizes.
• the multipath signals vary in number, amplitude and delay time from location to location and from channel to channel at a given location.
• the parameters of a ghost signal may also be time-varying.
• the visual effects of multipath distortion can be broadly classified in two categories: multiple images and distortion of the frequency response characteristic of the channel. Both effects occur due to the time and amplitude variations among the multipath signals arriving at the reception site. When the relative delays of the multipath
• SUBSTITUTE SHEET signals with respect to the reference signal are sufficiently large, the visual effect is observed as multiple copies of the same image on the television display displaced horizontally from each other. These copies are sometimes referred to as “macroghosts” to distinguish them from “microghosts", which will be presently described.
• the direct signal predominates, and a receiver is synchronized to the direct signal.
• the ghost images are displaced to the right at varying position, intensity and polarity. These are known as trailing ghosts or "post-ghost” images.
• post-ghost ghost images displaced to the left of the reference image. These are known as leading ghosts or "pre-ghost” images.
• Multipath signals of relatively short delay with respect to the reference signal do not cause separately discernible copies of the predominant image, but do introduce distortion into the frequency response characteristic of the channel.
• the visual effect in this case is observed as increased or decreased sharpness of the image and iii some cases loss of some image information.
• These short-delay or close-in ghosts are most commonly caused by under inated or incorrectly terminated radio- frequency transmission lines such as antenna lead-ins or cable television drop cables.
• Short multipath effects, or microghosts are typically reduced by cancelation schemes.
• Short multipath effects, or microghosts are typically alleviated by waveform equalization, generally by peaking and/or group-delay compensation of the high frequency video response.
• Circuitry in the receiver can then examine the distorted GCR signal received and, with a priori knowledge of the waveform of a distortion-free GCR signal, can configure an adaptive filter to cancel, or at least significantly attenuate, the multipath distortion.
• a GCR signal should not take up too much time in the VBI (preferably no more than one TV line) , but should still contain sufficient information to permit circuitry in the receiver to analyze the multipath distortion and configure an compensating filter to cancel the distortion.
• the GCR signals are used in the television receiver for calculating the adjustable weighting coefficients of a ghost-cancelation filter through which the composite video signals from the video detector are passed to supply a response in which ghosts are suppressed.
• the weighting coefficients of this ghost-cancelation filter are adjusted so it ha-s a filter characteristic complementary to that of the transmission medium giving rise to the ghosts.
• the GCR signals can be further used for calculating the adjustable weighting coefficients of an equalization filter connected in cascade with the ghost-cancelation filter, for providing an essentially flat frequency spectrum response over the complete transmission path through the transmitter vestigial-sideband amplitude-modulator, the transmission medium, the television receiver front-end and the cascaded ghost-cancelation and equalization filters.
• a transmitted reference or GCR signal is used that is substantially the proposed BTA (Japanese) GCR signal and that utilizes as the main reference or deghosting signal a (sin x)/x waveform.
• This (sin x)/x waveform as received together with ghosts thereof is Fourier transformed to provide a set of Fourier coefficients.
• the Fourier transform of the ghosted GCR signal is then processed with an available Fourier transform of an unimpaired GCR to compute the deghosting filter parameters, that is, tap gain information for both an infinite-impulse-response (IIR) deghosting filter and a finite-impulse-response (FIR) waveform equalization filter.
• IIR infinite-impulse-response
• FIR finite-impulse-response
• U.S. Patent No. 4,896,213 issued 23 January 1990 to Kobo et alii and entitled "GHOST CANCELLING REFERENCE SIGNAL TRANSMISSION/RECEPTION SYSTEM” discloses a system with a built-in ghost cancelling device for reducing or eliminating ghost components attributable to group-delay distortion and frequency-versus-a plitude characteristic distortion generated in a signal transmission path.
• a digital signal composed of frame synchronizing signals, clock synchronizing signals, and data signals is generated and superposed on a television signal to be transmitted, during a VBI scan line thereof.
• the digital signal is utilized as a ghosted CGR signal in an arrangement that correlates that signal with its known non- ghosted CGR signal to control adaptive filtering of the video signal to reduce the ghost phenomenon.
• Bessel pulse chirp signals are the de facto standard
• SUBSTITUTE SHEET GCR signal for television broadcasting in the United States of America.
• the distribution of energy in the Bessel pulse chirp signal has a frequency spectrum extending continuously across the video frequency band. Th.e chirp starts at the lowest frequency and sweeps upward in frequency therefrom to the 4.1 MHz highest frequency.
• the chirps are inserted into the first halves of selected VBI lines, the 19th line of each field currently being preferred.
• the chirps which are on +30 IRE pedestals, swing from -10 to +70 IRE and begin at a prescribed time after the trailing edges of the preceding horizontal synchronizing pulses.
• the chirp signals appear in an eight-field cycle in which the first, third, fifth and seventh fields have a polarity of burst defined as being positive and the second, fourth, sixth and eighth fields have another polarity of burst defined as being negative.
• the initial lobe of a chirp signal ETP that appears in the first, third, sixth and eighth fields of an eight-field cycle swings upward from the +30 IRE pedestal to +70 IRE level.
• the initial lobe of a chirp signal ETR that appears in the second, fourth, fifth and seventh fields of the eight- field cycle swings downward from the +30 IRE pedestal to -10 IRE level and is the complement of the ETP chirp signal.
• the time-base stability of the GCR signal in the received television signal is critical in order for the procedure of determining the weights for the ghost cancelation and equalizing filters by analyzing the GCR signal to work well.
• the theoretical validity of a ghost-cancelation procedure using weighted summation of differentially delayed video signals depends on the same signal with different delays having given rise to the ghosted signal. If the length of scan lines is different during the GCR signal transmission than during other portions of the video signal, then the weights determined for generating ghost- free GCR signal by weighted summation of variously delayed GCR signals will not be appropriate for generating ghost- free video at other times by weighted summation of variously delayed video signals.
• a television receiver with included display device and ghost cancelation circuitry In a television receiver with included display device and ghost cancelation circuitry, the problem of time-base stability of the detected video signals is hot a problem when receiving - off-the-air broadcast signals or when receiving such signals as relayed by cable broadcasting or community antenna systems.
• r-f radio frequency
• VCR home video cassette recorder
• SUBSTITUTE SHEET detector in the television receiver with the display device, where the ghost-cancelation circuitry supposedly can be used for r-f signals received from off-the-air, cable or a video recording medium.
• the term "television set" is used in this specification to describe a television receiver front end with accompanying kinescope, power supply for a kinescope, deflection circuitry for a kinescope, portions of a television receiver associated with converting the composite video signal to the color signals for driving a kinescope, loudspeaker(s) , stereophonic sound detector or audio amplication circuitry.
• the conventional video cassette recorder includes a television receiver front end without those accompanying further items, which are termed a "television monitor" in this specification and the accompanying drawing. If in a VCR and TV-set combined into a single piece of apparatus called a "combo" one desires the capability simultaneously to record a program received on one channel and to display a program received on a different channel, two television receiver front ends have to be provided, one for the video tape machine with recording capability and one for the television receiver with image displaying capability.
• Ghost suppression circuitry is estimated to cost US$50 as a portion of manufacturing price, which appears as an increase of about
• SUBSTITUTE SHEET active video do not have the same actual time duration as the scan line in which GCR signal occurs.
• Good time-base stability is essential also where the 19th scan lines of several fields are differentially delayed thereafter to be linearly combined in order to separate a GCR signal component from accompanying horizontal sync pulse, front porch, back porch including color burst and +30 IRE GCR signal pedestal components. These accompanying components will not cancel out well if there be errors in the timing of the samples of the 19th scan lines when those lines are digitized to facilitate their being differentially delayed using temporary digital memory.
• Home VCRs generally are not capable of providing the requisite time-base stability for separating GCR signal this way. Summary of the Invention
• a basic precept taught by the inventors is that, when relatively unsophisticated video tape recording and playback apparatus are used, the satisfactory suppression of ghosts, particularly macroghosts, in composite video signal retrieved from recorded video tape is not possible in practice. ghost cancelation must be done before the tape recording and playback procedures that give rise to time-base instabilities that interfere with ghost cancelation. Otherwise, when a television signal with ghosts is recorded, a video tape results that when played back to supply television signals to a television set will result in ghosts appearing in the reproduced image, even if that set includes ghost cancelation circuitry that satisfactorily suppresses ghosts when receiving television signals off-the-air or off-cable.
• Ghost cancelation done before recording the tape results in a tape that can be played back to supply deghosted signals to a television set.
• the reproduced image will be ghost free even if the television set is one without ghost suppression circuitry.
• This aspect of the invention is embodied in a combination of: a television receiver front-end, including elements up
• SUBSTITUTE SHEET to and including the sound detector and video detector; a video tape machine including at least recording electronics, to which signals from said sound detector and said video detector are supplied for recording; and the improvement wherein ghost-suppression circuitry is connected for receiving composite video signal from the video detector of said television receiver front- end and is connected for supplying its response to that composite video signal, in which response at least one ghost is suppressed, to the recording electronics of said video tape machine instead of supplying the recording electronics of said video tape machine with the composite video signal as taken directly from the video detector of said television receiver front end.
• a further aspect of the invention is that in a combination of a television receiver and video tape machine
• a single computer can be used for calculating the parameters for the filters in both sets of ghost suppression circuitry.
• FIGURE I is a schematic diagram of a video tape machine with recording capability, in combination with a television receiver front end up to and including the sound detector and video detector used for supplying sound signal and composite video signal for recording, which combination in accordance with the invention includes ghostsuppression circuitry through which the composite video signal is passed before being supplied to the video tape recorder.
• FIGURE 2 is a schematic diagram of ghost-suppression circuitry suitable for inclusion in the FIGURE I combination.
• FIGURE 3 is a schematic diagram of circuitry for resetting a modulo-eight field counter in the FIGURE 2
• FIGURE 4 is a flow diagram of a deghosting method used with the FIGURE 2 deghosting circuitry.
• FIGURES 5, 6 and 7 is a schematic diagram of a combination of television receiver and video tape machine, which combination is called a "combo" and is constructed in accordance with the principles of the invention.
• FIGURE I shows a video tape machine 10 having recording capability, which by way of example may be a video cassette recorder (VCR) of VHS, super-VHS or Betamax type.
• VCR video cassette recorder
• the video tape machine 10 may also be an improved VHS recorder of the type described in U. S. patent No. 5,113,262 issuedl2 Mayl992 to C. H. Strolle et alii and entitled "VIDEO SIGNAL RECORDING SYSTEM ENABLING LIMITED BANDWIDTH RECORDING AND PLAYBACK.
• a television receiver front end 20 in response to a radio-frequency television signal received thereby, supplies a sound signal and a composite video signal for recording by the video tape machine 10.
• the radio-frequency television signal may be broadcast over the air and then captured by an aerial television antenna 30, as shown by way of example. Alternatively, the radio-frequency television signal can be provided over cable by community- antenna or other television cable service.
• the television receiver front end 20 includes the portions of a conventional television receiver normally employed in combination with a video tape machine having recording capability. These portions generally include a radio- frequency amplifier, a down converter or "first detector”, at least one intermediate-frequency amplifier, a video detector or " second detector", and a sound demodulator (frequently of intercarrier type) .
• the television receiver front end 20 further includes separation circuitry for horizontal synchronizing pulses and for vertical synchronizing pulses.
• SUBSTITUTE SHEET television receiver front end 20 is demodulated from a frequency-modulated sound carrier, as heterodyned to intermediate frequency by the down converter. Before its demodulation the frequency-modulated sound carrier is limited to remove amplitude variations therein, and the capture phenomenon suppresses responses to ghosts in the sound signal from the sound demodulator. Accordingly, the sound signal from the sound demodulator in the television receiver front end 20 is supplied directly to the video tape machine 10, there to be recorded in the conventional manner.
• the composite video signal from the video detector in the television receiver front end 20 is supplied to ghost- suppression circuitry 40 to have the accompanying ghosts removed or suppressed.
• the resulting "deghosted" composite video signal is supplied from the ghost-suppression circuitry 40 to the video tape machine 10, there to be recorded in the conventional manner.
• the ghost-suppression circuitry 40 can be any one of the types known to the art.
• FIGURE 2 illustrates one form the ghost-suppression circuitry 40 can take, which is suited for use with the Bessel-chirp GCR signals being inserted into the first halves of the 19th VBI lines of each field.
• Composite video signal, supplied to the FIGURE 2 ghost-suppression circuitry from the television receiver front end 20, is digitized by an analog-to-digital converter 50.
• the ADC 50 typically will supply eight-parallel-bit samples of digitized composite video signal.
• the digitized composite video signal is applied as input signal to a cascade connection of a post-ghost cancelation filter 51, which is an adaptive filter of IIR type; a preghost cancelation filter 52, which is an adaptive filter of FIR type; and an equalization filter 53, which is an adaptive filter of FIR type.
• the output signal of the filter cascade is a digital deghosted composite video signal, which is converted to an analog deghosted composite video signal by a digital-to- analog converter 54.
• the analog deghosted composite video signal is supplied to the video tape machine 10 for
• the digital-to-analog converter 54 is dispensed with in advanced designs where the video tape machine 10 does digital recording, rather than recording analog signals.
• a filter coefficient computer 55 computes the weighting coefficients for the adaptive filters 51,52 and 53. These weighting coefficients are binary numbers, which the filter-coefficient computer 55 writes into registers within the digital filters 51, 2 and 53.
• the weighting coefficients stored in registers thereof are used as multiplier signals for digital multipliers receiving the filter output signal with various amounts of delay as multiplicand signals.
• the product signals from the digital multipliers are combined by addition and subtraction in digital adder/subtractor circuitry to generate the IIR filter response.
• the weighting coefficients stored in registers thereof are used as multiplier signals for digital multipliers receiving the filter input signal with various amounts of delay as multiplicand signals.
• the product signals from the digital multipliers are combined by addition and subtraction in digital adder/subtractor circuitry to generate the weighted summation response characteristic of an FIR filter.
• Pre-ghosts occurring in off-the-air reception can be displaced as much as 6 microseconds displacement from the direct signal, but typically displacements are no longer than 2 microseconds.
• direct off-the-air pick-up can precede the cable-supplied signal as much as 30 microseconds.
• the in-band video response can be rolled off as much as 20 dB at 3.6 MHz, but roll-off at 3.6 MHz is usually less than 10 dB.
• the number of taps in the FIR filters 52 and 53 depends on the range over which ghost suppression is sought; to keep filter costs within commercial constraints, typically the FIR filter 52 has around 64 taps for suppressing ghosts with as much as 6 microseconds displacement from the direct signal.
• SUBSTITUTE SHEET filter 53 used for frequency equalization need only have 32 taps or so.
• the cascaded FIR filters 52 and 53 are replaced in some designs by a single FIR filter having about 80 taps.
• the range for post-ghosts extends to 40 microseconds displacement from the direct signal, with 70% or so of post-ghosts occuring in a subrange that extends to 10 microseconds.
• the IIR post-ghost cancelation filter 51 required for suppressing post-ghosts over the full range can be as many as 600 taps long. However, since post- ghosts occur at discrete displacements, the weighting coefficients for many of these taps of the filter 51 will be zero-valued or nearly so.
• the taps requiring weighting coefficients of value significantly more than zero are usually clumped together in groups of ten or less. It is desirable, from the standpoint of economy of hardware, to use only as many digital multipliers as there are expected to be weighting coefficients of value significantly more than zero. Accordingly, the tapped delay line in the IIR filter 51 is usually designed as a cascade connection of ten-tap-or-so delay lines interspersed with programmable delay devices, making filter 51 what is sometimes termed a "sparse-weighting" filter.
• the ten-tap-or-so delay lines furnish signals to the digital multipliers for weighting.
• the programmable bulk delay devices comprise various length delay lines the chaining together of which can be controlled in response to control signals expressed as binary numbers.
• Such a sparse-weighting filter will include registers for the binary numbers specifying the delays of the programmable delay devices, the contents of which registers are also controlled by the filter- coefficient computer 55.
• SUBSTITUTE SHEET vertical sync pulses are counted modulo-8 by a three-stage digital counter 57, denominated "field counter". These counts are available to the filter-coefficient computer 55 for use in timing its operations, although connections for 5 furnishing these counts to the computer 55 are left out of FIGURE 2 to reduce its complexity.
• a decoder 58 responds to the scan line count from the line counter 56 being nineteen, corresponding to the scan line in each field containing GCR signal, to condition the output signal of a 10 multiplexer 59 to correspond to the digitized composite video signal from ADC 50 supplied as a first input signal thereto, rather than to a wired zero supplied as a zeroeth input signal thereto.
• FIGURE 15 provides a temporary (scan) line store 60 in FIGURE 2, which store 60 may be replaced by serial memory in alternative embodiments of the ghost-suppression circuitry.
• This temporary line store 60 is connected in an arrangement for accumulating the 19th-VBI-line GCR signals on a per
• the separated Bessel-chirp information is serially loaded one pixel at a time into a register of the filter- coefficient computer 55 during any line of FIELD 000 after its lath and before the line store 60 is cleared of data.
• the line store 60 is cleared of data during the last line of the last field of the eight-field sequence, but this clearance can take place during any line of FIELD 000 after the separated Bessel-chirp information is written into a register of the filter-coefficient computer 55.
• SUBSTITUTE SHEET transfer of accumulated data from the line store 60 to the computer 55 and the subsequent clearing of the accumulated data from the line store 60 can also take place during any two of the 1st through 18th scan lines of FIELD 001.
• the temporary line store 60 has to have the capability of storing a full scan line of sixteen- parallel-bit samples, assuming that it is to accumulate on a signed basis eight lines of eight-parallel-bit samples of digitized composite video signal supplied from the ADC 50.
• the signed arithmetic is preferably two's complement arithmetic.
• a digital adder/subtractor 61 supplies a sixteen-parallel-bit output signal to the temporary line store 60 as its write input signal.
• the digital adder/subtractor 61 receives as a first input thereto the output signal of a multiplexer 62, which normally corresponds to the readout from the temporary line store 60 received as the zeroeth input of the multiplexer 62.
• the digital adder/subtractor 61 receives as a second input thereto the eight-parallel-bit output signal of the multiplexer 59 together with eight wired ZEROs as a sign- bit extension.
• a decoder 80 decodes the modulo-eight field count being one, three, six, or zero (i. e., eight) to furnish a logic ZERO to the digital adder/subtractor 61 to condition it to add its input signals.
• the decoder 80 decodes the modulo-eight field count being two, four, five, or seven to furnish a logic ONE to the digital adder/subtractor 61 to condition it subtract its second input signal (supplied from the multiplexer 59) from its first input signal
• SUBSTITUTE SHEET multiplexer 62 is caused to be a ONE. This ONE conditions the multiplexer 62 to furnish an output signal corresponding to a first input thereto, which is an arithmetic zero comprising sixteen parallel bits Of Wired ZEROS. This results in the resetting of the accumulation result in the temporary line store 60 to arithmetic zero.
• the control signal for the multiplexer 62 is shown in FIGURE 2 as being generated by a two-input AND gate 63.
• a decoder 64 decodes the count from the scan line counter 56 corresponding to the last line of the current field to generate one of the input signals to the AND gate 63.
• a decoder 65 decodes the modulo-eight field count from the counter 57 to generate the other of the input signals to the AND gate 63.
• the eighth of each sequence of eight fields generates a 000 modulo-eight count from the field counter 57.
• Both the input signals to the AND gate 63 are ONE only during the last line of the eighth of each sequence of eight fields, during which line the AND gate 63 supplies a ONE to the multiplexer 62 as its control signal, causing the accumulation result stored in the temporary line store 60 to be reset to arithmetic zero.
• a two-input AND gate 66 supplies a ONE to the filter- coefficient computer 55 when the accumulation result stored in the temporary line store 60 is available for transfer into a ghosted Bessel-chirp register within the internal memory of the computer 55.
• the output signal of the decoder 65 is one of the input signals to the AND gate 66 and is ONE only during the eighth of each sequence of eight fields.
• a two-input NOR gate 67 generates the other of the input signals to the AND gate 66.
• the NOR gate 67 responds to the output signal of the decoder 64, which detects the last line of a field in the count from line counter 56, and to the output signal of a decoder 68, which detects the vertical blanking interval proceeding from the count from line counter 56. Accordingly, the NOR gate 67 output signal is ONE except during the vertical blanking interval or the last line of a field. So, the accumulation result in the temporary line store 60 is available for transfer into the internal memory of the computer 55 any time during the
• An oscillator 70 that has automatic frequency and phase control (AFPC) generates sinusoidal oscillations at the second harmonic of color subcarrier frequency as a primary clocking signal.
• a zero-crossing detector 71 detects average axis crossings of the sinusoidal oscillations to generate pulses at a rate four times color subcarrier frequency. These pulses time the sampling of the composite video signal for digitization by the ADC 50; and they would time the advance of data in the temporary line store 60 if it were a serial memory.
• AFPC automatic frequency and phase control
• the temporary line store 60 is a random-access memory arranged for read-then- write operation as each of its storage locations is addressed.
• the addresses of its storage locations are recurrently scanned in accordance with the count of pixels supplied from a ten-stage digital counter 72 denominated as "pixel counter", which counts the pulses from the zero- crossing detector 71.
• pixel counter which counts the pulses from the zero- crossing detector 71.
• These same addresses are supplied to the filter-coefficient computer 55 to be used to address a line storage register therein when separated GCR signal is transfered thereto from the temporary line store 60.
• the color burst signal is the most stable frequency reference in a composite video signal and is the preferred reference signal for AFPC of the oscillator 70.
• the overflow signal from the second stage of the pixel counter 72 is presumably a 3.58 MHz square wave and is supplied as a feedback signal to a first AFPC detector 73 for comparison to a separated burst signal, in order to generate an error signal an AFPC signal multiplexer 74 selectively applies to the pixel counter 72 for controlling the frequency and phase of its oscillations.
• a burst gate 75 responds to pulses from a burst gate control signal generator 76 to separate from the analog composite video signal supplied from the TV receiver
• SUBSTITUTE SHEET front end 20 a color burst signal to be supplied to the first AFPC detector 73.
• the horizontal sync pulses from the television receiver front end 20 are supplied to the burst gate control signal generator 76 and their trailing edges are used to time the pulses that the generator 76 generates during color burst intervals.
• a cascade of astable flip-flops or "one-shots" are customarily employed in the generation of these pulses.
• the decoder circuitry 68 responds to the scan line counts that the line counter 56 provides which correspond to the VBI lines in each field to generate an inhibitory signal.
• This inhibitory signal is applied to the burst gate control signal generator 76 to inhibit its generating pulses, so that the burst gate 75 will select only those backporch intervals during a field which can have color burst.
• the burst gate control signal generator 76 is not inhibited from generating burst gate pulses during the vertical blanking interval and the time constant of the first AFPC detector is made longer than necessary in the FIGURE 2 circuitry.
• An amplitude detector 77 denominated the "color burst presence detector” detects when burst is present in the output signal from the burst gate 75 to supply a ONE that conditions the AFPC signal multiplexer 74 to select the output signal from the first AFPC detector 73 as a first error signal, for application to the controlled oscillator 70 as its AFPC signal.
• the amplitude detector 77 comprises a synchronous detector stage followed by a threshold detector stage followed by a short-pulse eliminator.
• Arrangements of the pixel counter 72 can be made for providing a pair of 3.58 MHz square waves in quadrature phase relationship with each other for application to the synchronous detection portions of the detectors 73 and 77.
• Arrangements of counters to provide square waves in quadrature phase relationship with each other are familiar to television circuit designers, being commonly used in television stereophonic sound decoders. Short-pulse eliminators are
• SUBSTITUTE SHEET known from radar and are commonly constructed using circuitry for ANDing differentially delayed input signal thereto thereby to generate output signal therefrom.
• the reference signal for AFPC of the oscillator 70 will have to be the separated horizontal sync pulses supplied to the AFPC circuitry from the TV receiver front end 20.
• the color burst presence detector 77 will supply a ZERO when the composite video signal supplied from the TV receiver front end 20 has no attendant color burst, conditioning the AFPC signal multiplexer 74 to select the output signal from a second AFPC detector 78 to controlled oscillator 70 as its AFPC signal.
• a sync decoder 79 responds with a ONE to the count( ⁇ ) of the pixel counter 72 theoretically corresponding to the occurence of the horizontal sync pulse or a prescribed portion thereof, such as an edge thereof.
• the output signal from the sync decoder 79 is supplied as feedback signal to the second AFPC detector 78, which compares that feedback pulse to an input reference signal taken from the horizontal sync pulses supplied from the horizontal sync separator in the TV receiver front end 20 and generates a second error signal for being selectively applied by the AFPC signal multiplexer 74 to controlled oscillator 70 as its AFPC signal.
• This AFPC arrangement is called "line-locked-clock" by television engineers.
• Stability of the oscillations of the controlled oscillator 70 is required over the number of fields from which the 19th scan lines are taken for accumulation in the temporary line store 60, in order that the accumulation procedure by which the Bessel chirp is separated from those lines' adequately suppresses horizontal sync pulse, front porch, back porch including color burst and +30 IRE pedestal.
• Crystal control of the frequency of the oscillations is a practical necessity; and the automatic phase control (APC) aspect of the AFPC should predominate, with the automatic frequency control (APC) aspect of the AFPC having a very long time constant — i. e. , several
• the circuits for resetting the counters 56, 57 and 72 are omitted from FIGURE 2 to avoid undue complexity.
• the scan line counter 56 can be simply reset by the leading edges of vertical sync pulses supplied from the vertical sync separator in the TV receiver front end 20.
• the pixel count from the pixel counter 72 is reset when necessary in order to re-synchronize it with the scan lines in the composite video signal supplied from the video detector of the TV receiver front end 20.
• the leading and trailing edges of the horizontal sync pulses supplied from the horizontal sync separator of the TV receiver front end 20 are detected, using a differentiator followed by a appropriate level comparators.
• the leading edge detector result is used to command the loading of a temporary storage register with the current pixel count.
• the pixel count is applied to a window comparator to determine if it is within its expected range and to generate an indication of error if it is not.
• the count of the pixel counter 72 is conditionally reset to zero responsive to the trailing edge detector result.
• the condition for reset may be a single indication of pixel count error. However, better noise immunity is obtained by counting the errors in an up/down counter configured so a given number of consecutive errors must be counted before pixel count is corrected.
• FIGURE 3 shows circuitry for resetting the modulo- eight field counter , 57 so its count either is correctly phased or is misphased by four fields.
• the temporary line store 31 is shown as being a random-access memory addressed by the pixel count supplied from the pixel counter 72.
• the line store 31 is arranged for read-then- write operation.
• the logic ONE issued by the decoder 58 only during the 19th scan line of each field is furnished to a multiplexer 310 to condition the updating of the temporary line store 31 with digitized 19th scan line samples supplied from the ADC 50.
• the logic ZERO issued by the decoder 58 conditions the
• SUBSTITUTE SHEET multiplexer 310 to apply the data read from the temporary line store 31 for writing back thereinto.
• the temporary line store 31 is provided with- pixel latches 32 and 33 clocked by the output signal from,the zero-crossing detector 71.
• the pixel latches 32 and 33 are used for temporily storing the last pixel written into the temporary line store 31 and the last pixel read out of the temporary line store 31, respectively, aligning those samples in time to be respective ones of the subtrahend and minuend input signals of a digital subtractor 34.
• the pixel samples of the difference signal from the subtractor 34 will all be zero valued except during 19th scan lines.
• the difference signal from the subtractor 34 is furnished to an absolute-value circuit 35, which can comprise a battery of two-input exclusive-OR gates each receiving the sign bit of the difference signal as a first input and receiving a respective other bit of the difference signal for selectively complementing, and which can then further comprise a digital adder for adding the sign bit of the difference signal to the selectively complemented remaining bits of the difference signal to generate as a sum output signal the absolute value of the difference signal.
• an absolute-value circuit 35 can comprise a battery of two-input exclusive-OR gates each receiving the sign bit of the difference signal as a first input and receiving a respective other bit of the difference signal for selectively complementing, and which can then further comprise a digital adder for adding the sign bit of the difference signal to the selectively complemented remaining bits of the difference signal to generate as a sum output signal the absolute value of the difference signal.
• An accumulator 36 for successive samples of the absolute-value circuit 35 output signal includes an output latch 361 for temporarily storing successive values of the accumulation result, a digital adder 362 for adding the successive samples of the output signal of the absolute- value circuit 35 to the accumulation result to augment its value, and a multiplexer 363 for selectively supplying the augmented accumulation result to the output latch 361 for updating its contents.
• the multiplexer 363 is wired for inserting arithmetic zero into the output latch 361 whenever the decoder 58 does not detect the counter 56 supplying a scan line count of nineteen.
• a decoder 364 responds to the pixel count from the counter 72 being descriptive of those portions of a scan line as may contain Bessel chirp information to furnish a ONE, which is ANDed with the output signal from the zero-crossing detector 71
• the output latch 361 is clocked to receive input data responsive only to a ONE being received from the AND gate 365.
• the accumulation result should have appreciable value if the current field is not FIELD 001 or FIELD 101.
• the 19th lines of FIELD 000 and of FIELD 001 both contain ETP signal, so their difference is zero-valued except for noise.
• the 19th lines of FIELD 100 and of FIELD 101 both contain ETR signal, so their difference is zero-valued except for noise.
• the output signal of a threshold detector 37 which is a ONE when the accumulation result is substantially more than arithmetic zero and is otherwise a ZERO, is complemented by a NOT gate 38 to supply one of the four input signals of an AND gate 39.
• a decoder 41 detects the field count from the counter 57 being other than 001 or 101 to furnish a ONE to the AND gate, which ONE is indicative that the field count is misphased and enables the resetting of the counter 57.
• the output signal of the decoder 58 which detects the occurence of the 19th line of a field, and the output signal of a decoder 42, which responds to the pixel count from the counter 72 to detect the end of a scan line, are the other two input signals to the AND gate 39.
• the AND gate 39 Providing that the field count is not 001 or 101, the AND gate 39 generates a ONE to reset the counter 57 to 001 field count at the end of the 19th line of a FIELD 000 or of a FIELD 100 in the television signal received by the TV receiver front end 20.
• the counter 57 could be reset to 101; or provision can be made for resetting only the two least significant bits of the field count, resetting them to 01.
• SUBSTITUTE SHEET accumulation will be eight times the ETP Bessel chirp signal devoid of accompanying horizontal sync pulse, front porch, back porch including color burst and +30 IRE pedestal.
• the modulo-eight field count provided by the field counter 57 is misphased by four fields, the accumulation result attained in the temporary line store 60 during FIELD 000, the last field in the cycle of accumulation will be eight times the ETR Bessel chirp signal devoid of accompanying horizontal sync pulse, front porch, back porch including color burst and +30 IRE pedestal.
• a wired three binary place shift in the direction towards reduced magnitude divides the accumulation results attained in the temporary line store 60 during FIELD 000 by eight, and the resulting quotients are supplied as the ETP or ETR signal to the filter- coefficient computer 55.
• the filter-coefficient computer 55 which is well- adapted to performing correlations against a ghost-free Bessel chirp function ETP or ETR stored in an internal register thereof, is programmed to perform a correlation substep that determines whether the input it receives from the temporary line store 60 during FIELD 000 is ETP signal, is ETR signal, or is unrelated to the ETP or ETR signal. This procedure enables the filter-coefficient computer 55 to determine when no GCR signals are included in the television signal received by the TV receiver front end 20.
• the computer 55 may then apply predetermined weighting coefficients as stored in registers therewithin to the filters 51, 52 and 53.
• the computer 55 may be arranged to compute weighting coefficients for the filters 51, 52 and 53 proceeding from data concerning received ghosts supplied by means that do not rely on GCR signals being included in the television signal received by the TV receiver front end 20.
• circuitry external to the computer 55 is provided for analyzing the GCR signal stored in the temporary line store 31 (during the scan line following its acquisition, for
• SUBSTITUTE SHEET example to determine whether it is an ET P or ET R signal and this determination is used to determine whether the most significant bit of the reset condition for the field counter 57 is a ZERO so reset is to 001 field count or is a ONE so reset is to 101 field count.
• the contents of the temporary line store 31 are scanned in accordance with the pixel count from the counter 72 during the analysis procedure.
• the portions of the pixel count corresponding to the initial lobe of the Bessel chirp are decoded to selectively generate a ONE that is used to enable accumulation by either of two accumulators.
• One accumulator further requires that the sign bit of the current GCR signal be ZERO in order to accumulate its magnitude (absolute-value) in excess of a threshold value T.
• the other accumulator further requires that the sign bit of the current GCR signal be ONE in order to accumulate its magnitude (absolute-value) in excess of a threshold value T.
• the magnitudes of the accumulator contents are each compared in respective comparators to a threshold value T that is almost as large as the integral of the absolute value of the initial lobe of the Bessel chirp. If the contents of the accumulator that requires that the sign bit of the current GCR signal be ZERO in order to accumulate exceeds this threshold T after the initial lobe of the Bessel chirp, the associated comparator furnishes a ONE to the filter-coefficient computer 55, which identifies the presence of an ETP signal.
• the associated comparator furnishes a ONE to the computer 55, which identifies the presence of an ETR signal. If this threshold T is not exceeded by the contents of either of these accumulators after the initial lobe of the Bessel chirp, the associated comparators furnish ZEROs to the computer 55, which determines that
• FIGURE 2 ghost-suppression circuitry are possible wherein, when data is being transferred from the temporary line store 60 to a line storage register in the filter-coefficient computer 55, the addressing of the
• FIGURE 2 ghost- suppression circuitry are also possible wherein a plurality of temporary line stores are used, instead of a single
• the decoder 58 may be replaced by a decoder for detecting the presence of the 19th - 20th scan lines to condition the multiplexer 59 for loading the temporary two-scan-line store.
• the temporary single-scan-line store 60 may be
• the decoder 58 may be replaced by a decoder for detecting the presence of the 19th through 21st scan lines to condition the multiplexer 59 for loading the temporary three-scan- line store.
• SUBSTITUTE SHEET ghost-suppression circuitry is the accumulation in the temporary line store 60 of the 19th scan lines from 16 consecutive fields, rather than S. This further correlates the separated Bessel chirp information, which significantly improves its signal-to-noise ratio as supplied to the filter-coefficient computer 55.
• the modulo-8 field counter 57 is replaced by a modulo-16 field counter. Further accumulation — e.g., of the 19th scan lines from 24 consecutive fields — provides little more improvement in the signal-to-noise ratio of the separated Bessel chirp information supplied to the filter-coefficient computer 55.
• FIGURE 4 shows the flow diagram of a procedure for establishing the operating parameters of the filters 51, 52 and 53, which procedure is carried out by the filter- coefficient computer 55.
• Entry to the START condition 81 of the procedure is at the time power is turned on in the television receiver, when a new channel is tuned, or when a prescribed time has elapsed since the last deghosting procedure.
• a RESET ALL DEGHOST FILTERS step 82 preferably sets the filter coefficients in the filters 51, 52 and 53 to values previously determined for the channel to which the TV receiver front end 20 is tuned and stored in a channel-addressed memory.
• the filter coefficients in the filters 51, 52 and 53 can be to values associated with a ghost-free signal; and during periodic deghosting previous values of the filter coefficients are retained during "reset".
• An ACQUIRE DATA step 83 then follows, which step 83 is completed after the number of fields elapse that the computer 55 must wait for accumulation in the temporary line store 60 to be completed, in order to generate a separated GCR signal that is suitable input data for the computer 55.
• the ACQUIRE DATA step 83 includes a correlation substep not shown in FIGURE 4 which substep determines whether the input the computer 55 receives from the temporary line store 60 during FIELD 000 is ETP signal, is ETR signal, or is unrelated to the ETP or ETR signal.
• a CHANNEL CHARACTERIZATION step 84 then takes place.
• SUBSTITUTE SHEET The location in time of the predominant response in the data supplied the computer 55 is detected, then the respective location in time of each successively smaller one of the significantly large ghost responses, up -,to the 5 number of post-ghosts that can be suppressed by the filter 5 1 and up to the number of pre-ghosts that can be suppressed by the filter 52.
• the respective locations in time of the predominant response and multipath responses in the data supplied the computer 55 are calculated to be used
• An UPDATE IIR COEFFICIENTS step 85 is performed after the CHANNEL CHARACTERIZATION step 84 is performed, in which step 85 the programmable delays and the non-zero weighting 20 coeffients of the IIR filter 51 are updated.
• An UPDATE FIR COEFFICIENTS step 86 is performed after the UPDATE IIR COEFFICIENTS step 85.
• the non-zero weighting coeffients of the FIR filter 52 are updated.
• the procedure loops back to the ACQUIRE DATA step 83.
• a threshold 30 dB down from the predominant image has been used in step 87. If the decision is YES, all significant ghosts have been canceled or the filters 51 and
• the procedure goes on to an EQUALIZATION step 88 in which weighting coefficients for the amplitude-equalization filter 53 are calculated.
• SUBSTITUTE SHEET cancelation filters here the post-ghost filter 51, gives rise to spurious ghosts of the type that cannot be suppressed by the final one of these filters. Since the weighting coefficients calculated in the channel characterization step 84 normally do not take these spurious ghosts into account, the weighting coefficients of the initial one of the cascaded ghost cancelation filters should be recalculated to introduce compensatory ghosts that will reduce the spurious ghosts in the initial filter response. Since this reduction may not be complete, recalculation of the weighting coefficients of the final one of the cascaded ghost cancelation filters is advisable. The decision loop around steps 83-86 implements these recalculations.
• the EQUALIZATION step 88 can be performed by taking the discrete Fourier transform (DFT) of the response of the cascade connection of the filters 51, 52 and 53 to the correlator response, then dividing it by the DFT of the ideal correlator response as stored in the memory of the computer 55, thereby to obtain the basis for calculating the adjustments necessary in the tap weights of the FIR filter 53. Since the number of taps for the FIR filter 53 is typically no more than thirty-two, the number of spectral bins in the DFT is reasonably small; however, the DFT calculations tend to be lengthy.
• DFT discrete Fourier transform
• An alternative, more rapid way to calculate equalization filter coefficients is to use a least-means- squares method to adjust the filter 53 weighting coefficients so that the response of the cascade connection of filters 51-53 accumulated in the temporary line store 60 best fits an ideal response stored in the memory of the computer 55.
• the FIGURE 3 procedure reaches the condition 89, DONE. It is preferred that the UPDATE FIR COEFFICIENTS step 85 and the EQUALIZATION step 88 be performed after the UPDATE IIR COEFFICIENTS step 85 is performed, because the higher-order ghosts generated in the IIR filtering can be accounted for before the FIR filtering coefficients are computed. Then,
• the FIGURE 5 combo includes the video tape machine 10, the television receiver front end 20 for that machine, and the television antenna 30 of FIGURE I together with adaptive ghost-suppression circuitry similar to the adaptive ghost-suppression circuitry 40 of FIGURE 1.
• This ghost-suppression circuitry consists of a filter— coefficient computer 90; GCR signal acquisition circuitry 91, which acquires a GCR signal for the computer 90 from the television receiver front end 20; and a ghost- suppression filter 92, which deghosts the composite video signal from the television receiver front end 20 for application to the video tape machine 10 as a deghosted composite video signal for recording.
• the video tape machine 10 sound signal for recording from the sound detector in the television receiver front end 20.
• the FIGURE 5 combo further includes another television receiver front end 93 to which the television antenna 30 is selectively connected by an input selection switch 94; further GCR signal acquisition circuitry 95, which acquires a GCR signal for the computer 90 from the television receiver front end 93; and another ghost-suppression filter 96, which deghosts the composite video signal from the television receiver front end 93.
• the further GCR signal acquisition circuitry 95 is similar to the GCR signal acquisition circuitry 91, which can be similar to that formed by the elements 56-78.
• the ghost-suppression filters 92 and 96 can each comprise a respective cascade connection of filters similar to the cascade connection of filters 51-53 shown in FIGURE 2.
• the deghosted composite video from the ghost- suppression filter 96 is supplied to a luma/chroma separator 97, which responds to the deghosted composite video to generate separated luminance and chrominance signals.
• Chroma demodulator circuitry 98 responds to the separated chrominance signal to generate a pair of color- difference signals —e. g. , the well-known. I and Q color- difference signals — for application to color matrix
• the SUBSTITUTE SHEET circuitry 99 together with the separated luminance signal.
• the color-matrix circuitry 99 generates red (R) , green (G) and blue (B) color signals for application to a television monitor 100.
• the television monitor 100 responds to the R, G and B signals from the color-matrix circuitry 99 and to sound signal, horizontal synchronizing signal and vertical synchronizing signal from the television receiver front end 93 to generate television images with accompanying sound for a human observer.
• a feature of the FIGURE 5 combo is that the single filter-coefficient computer 90 performs a dual function, computing on a time-division multiplexed basis the filter coefficients both for the ghost-suppression filter 92 and for the ghost-suppression filter 96. There is a significant cost saving realized by not using separate computers for computing the filter coefficients for the filters 92 and 96. There are some technical advantages as well, as saving in power consumption being one of them.
• FIGURES 5, 6 and 7 are similar, each computing on a time-division multiplexed basis the filter coefficients for both of the ghost-suppression filters 92 and 96.
• the differences between these combos is in the way that playback sound and video are routed to the television monitor 100.
• the color-under signal recovered from a recorded video tape is up-converted in the playback electronics of the video tape machine 10 to generate a quadrature amplitude-modulated(QAM) playback chrominance signal.
• QAM quadrature amplitude-modulated
• a circuit 101 additively combines this chrominance signal with a playback luminance signal recovered from the recorded video tape, to generate a playback composite video signal, which composite video signal is modulated onto a low-power radio-frequency picture carrier in the amplitude-modulator 102.
• the sound signal recovered from the recorded video tape is modulated onto a low-power radio-frequency sound carrier in the frequency modulator 103.
• a circuit 104 additively combines the output signals from the modulators 102 and 103 to generate a low-level television signal, which signal the
• SUBSTITUTE SHEET input selection switch. 94 can select as input signal to the front end 93 rather than signal from the television antenna 30 when playing back from a recorded video tape, and which is available for application to another television receiver.
• the television receiver front end 93 is partitioned into a leading portion, including a radio-frequency amplifier 105 and down converter 106, and a trailing portion, including a plural-stage intermediate- frequency amplifier 107, a sound detector 108, a video detector 109, and sync separation circuitry 110.
• a selector switch 111 selects the input signal for the intermediate-frequency amplifier 107, with the output signal from the down converter 106 being selected when the television monitor 100 is used for viewing signals received via antenna 30 or a cable substitute therefor.
• the color- under signal recovered from a recorded video tape is up- converted in the playback electronics of the video tape machine 10 to generate a quadrature amplitude- modulated(QAM) playback chrominance signal.
• the circuit 101 additively combines this chrominance signal with a playback luminance signal recovered from the recorded video tape, to generate a playback composite video signal, which composite video signal is modulated onto a low-power intermediate-frequency picture carrier in the amplitude- modulator 112.
• the sound signal recovered from the recorded video tape is modulated onto a lowpower intermediate-frequency sound carrier in the frequency modulator 113.
• a circuit 114 additively combines the output signals from the modulators 112 and 113 to generate a low-level television signal, which signal the selector switch 111 selects as input signal to the intermediate- frequency amplifier 107 during playback from a recorded video tape to the television monitor 100.
• An up converter 115 can be used to generate a low-level television signal for application to another television receiver. This up converter 115 is shown being located directly after the additive combining circuit 114, but may alternatively be located after the intermediate-frequency amplifier 107, so
• SUBSTITUTE SHEET acts as a vestigial sideband filter.
• the television receiver front end 93 is modified so that a selector switch 116 selects the input signal applied to the sync separation circuitry 110.
• the composite video signal from the video tape machine 10 is selected to the sync separation circuitry 110 during playback of a recorded video tape; otherwise, the composite video signal from the video detector 109 is selected as the input signal for the sync separation circuitry 110.
• a further selector switch 117 selects the chrominance signal from the video tape machine 10 to the chroma demodulator circuitry 98 as its input signal; otherwise, the selector switch 117 selects the chrominance signal from the luma/chroma separator 97 as the circuitry 98 input signal.
• a still further selector switch 118 selects the luminance signal from the video tape machine 10 to be applied to the color matrix circuitry 99 as its luminance signal input; otherwise, the selector switch 118 selects the luminance signal from the luma/chroma separator 97 as the circuitry 99 luminance signal input.
• yet another selector switch 119 selects the sound signal from the video tape machine 10 to be applied to the TV monitor 100; otherwise, the selector switch 119 selects the sound signal from the TV receiver front end 93 to be applied to the TV monitor 100.
• Video tapes recorded responsive to ghosted radio- frequency television signals using any of the combinations shown in FIGURES 11 5, 6 and 7 of the drawing can be played back on a conventional video tape player to supply deghosted radio-frequency television signals to a conventional television receiver having no ghost- suppression circuitry; and the television images so recovered will be free of ghosts. This makes the combinations shown in FIGURE 1, 5, 6 and 7 of the drawing still more attractive commercially.

Landscapes

• Engineering & Computer Science (AREA)
• Multimedia (AREA)
• Signal Processing (AREA)
• Picture Signal Circuits (AREA)

Priority Applications (2)

Applications Claiming Priority (2)

Publications (1)

Family

ID=25496262

Family Applications (1)

Country Status (4)

Cited By (1)

Citations (4)

Family Cites Families (3)

• 1993
  • 1993-02-26 JP JP06507563A patent/JP3142137B2/ja not_active Expired - Fee Related
  • 1993-02-26 KR KR1019940700598A patent/KR0169617B1/ko not_active Expired - Fee Related
  • 1993-02-26 WO PCT/KR1993/000014 patent/WO1994008426A1/en unknown
  • 1993-08-16 TW TW82106564A patent/TW224560B/zh active

Patent Citations (4)

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events