KR100557122B1 - A receiver for dtv signals subject to co-channel ntsc interference and method of operating therefor - Google Patents

Abstract

The DTV signal receiver of the present invention includes an adaptive channel equalization filter, a comb filter for suppressing co-channel NTSC interference artifacts, and an intersymbol interference suppression filter for compensating intersymbol interference introduced by the comb filter. The adaptation of the filter coefficients of the equalization filter is made to be insensitive to the presence or absence of co-channel NTSC interference artifacts. Accordingly, it is possible to prevent the channel equalization filter from attempting to compensate for the intersymbol interference of the comb filter, which may not be made completely but may make the intersymbol interference introduced into the comb filter unpredictable. The adaptation of the equalization filter coefficient is done without affecting the intersymbol interference introduced by the comb filter. As a result, the intersymbol interference can be properly compensated by the intersymbol interference suppression filter.

Description

A receiver for a digital television signal subject to co-channel interference and a method of operation thereof {A RECEIVER FOR DTV SIGNALS SUBJECT TO CO-CHANNEL NTSC INTERFERENCE AND METHOD OF OPERATING THEREFOR}

The present invention relates to a digital television system, and more particularly to a method for adjusting a channel equalization circuit used in a receiver of a digital television signal that is wirelessly broadcast using the same channel as the broadcast channel of NTSC analog television signals.

Digital Television Standards, published by the Advanced Television Systems Committee (ATSC) on September 16, 1995, include residual sideband (VSB) signals used for transmission of digital television (DTV) signals on television channels with 6 MHz bandwidth. Are specified. The DTV signals will be transmitted on some of the microwave channels currently used in wireless broadcasting of NTSC (National Television System Committe) analog television signals in the United States. The VSB DTV signal is designed so that its spectrum is easily interleaved with the spectrum of the co-channel interfering NTSC analog TV signal. The symbol frequency of the DTV signal corresponds to three times the NTSC color subcarrier frequency, and the 3.58 MHz subcarrier frequency corresponds to 455/2 times the NTSC scan line rate. The pilot carrier and principal amplitude-modulated sideband frequencies of the DTV signal are located at odd multiples of one-quarter of the horizontal scan line speed of the NTSC analog TV signal. As a result, the DTV signal components are located at an even multiple of 1/4 of the horizontal scan line speed of the NTSC analog TV signal, and most of the energy of the luminance and chromatic components of the co-channel coherent NTSC analog TV signal is present at the even multiples. It is. The video carrier of the NTSC analog TV signal is offset by 1.25 MHz from the lower limit frequency of the television channel. Further, the carrier of the DTV signal is offset from the video carrier of the NTSC analog TV signal as described above by 59.75 times the horizontal scanning line speed of the NTSC analog TV signal, and is located about 309,877.6 kHz away from the lower limit frequency of the television channel. Thus, the carrier of the DTV signal is located about 2,690,122.4 Hz away from the center frequency of the television channel.

The exact symbol rate according to the digital television standard is 684/286 times the sound carrier offset by 4.5 MHz from the video carrier of the NTSC analog TV signal. Here, "684" represents the number of symbols per horizontal scan line of the NTSC analog TV signal, and "286" is a factor multiplied by the NTSC horizontal scan line speed to obtain a sound carrier offset by 4.5 MHz from the video carrier of the NTSC analog TV signal. Indicates. The symbol rate is a symbol rate corresponding to 10.762238 * 10 ⁶ symbols per second, and the symbol rate may be included in a VSB signal extended by 5.381119 MHz from the DTV signal carrier. That is, the VSB signal can be limited to a band extending by 5.690997 MHz from the lower limit frequency of the television channel.

According to the ATSC standard for terrestrial broadcasting of DTV signals in the United States, either of two high definition television (HDTV) formats having a 16: 9 aspect ratio can be transmitted. One HDTV format is a 2: 1 field interlaced scan, which uses 1,920 samples per scan line and 1,080 effective horizontal scan lines per 30 Hz frame. Another HDTV format is sequential scanning, which uses 1,280 luminance samples per scan line and 720 sequential scan lines of television images per 60 Hz frame. The ATSC standard also permits transmission of DTV display formats other than HDTV display formats, such as parallel transmission of four television signals with normal clarity compared to NTSC analog television signals.

The DTV signal transmitted by Residual Sideband (VSB) Amplitude Modulation (AM) for terrestrial broadcasting in the United States is a sequence of data fields with continuity in time, including 313 data segments each having continuity in time. It contains them. These data fields may be consecutively numbered modulo 2. Thus, each odd data field and subsequent even data fields form a data frame. Each data segment has a duration of 77.3 microseconds (ms). Therefore, if the symbol rate is 10.76 MHz, there are 832 symbols in each data segment. Each data segment starts with a four-segment data-segment-synchronization (DSS) code group that has consecutive + S, -S, -S, and + S values. The value + S is one level below the maximum positive data regression point, and the value -S is one level above the maximum negative data regression point. The initial data segment of each data field contains a data-field-synchronization (DFS) code group that encodes the training signal used for channel equalization and multipath suppression. The training signal consists of one 511-sample pseudonoise sequence ("PN sequence") followed by three 63-sample PN sequences. The intermediate ones of the 63-sample PN sequences of DFS codes are transmitted according to the first logic specification in the first line of each odd-numbered data field and in accordance with the second logic definition in the first line of each even-numbered data field. The first and second logic definitions have a complementary relationship to each other (ie, opposite polarities).

The data in the data lines are trellis using twelve interleaved trellis codes, which are two-third rate trellis codes with one uncoded bit each precoded. Is coded. The interleaved trellis codes are processed in a Reed-Solomon forward error correction coding scheme, which provides for correction of burst errors from noise sources such as automotive ignition systems that are almost non-blocking in terms of noise. Will be. The Reed-Solomon coding result is transmitted as symbol coded data of 8-level (3 bits / symbol) one-dimensional structure in the case of radio transmission. In addition, the Reed-Solomon coding result is transmitted as symbol-coded data of 16-level (4-bit / symbol) one-dimensional structure for cable broadcasting, in which case the transmission is performed without precoding after symbol generation. The VSB signal has a unique carrier whose amplitude will vary with the percentage of suppressed modulation.

The unique carrier is replaced with a pilot carrier of constant amplitude corresponding to a predetermined modulation percentage. This constant amplitude pilot carrier is generated by shifting, i.e., shifting, the direct current component of the modulation voltage applied to the balance modulator that generates the amplitude modulated sideband signal. The amplitude modulated sideband signal is provided to a filter which supplies a VSB signal as a response signal. If eight levels of the 4-bit symbol code have normalization values of -7, -5, -3, -1, +1, +3, +5, and +7 in the carrier modulated signal, the pilot carrier is It has a normalized value. In this case, the normalized value of + S is +5, and the normalized value of -S is -5.

VSB using a comb filter for suppressing artifacts of co-channel NTSC interference accompanying baseband symbol coded data and an intersymbol interference suppression filter that compensates for the intersymbol interference introduced by the comb filter DTV signal receivers are known. With respect to such a receiver, a patent, issued February 11, 1992, entitled "VSB HDTV TRANSMISSION SYSTEM WITH REDUCED NTSC CO-CHANNEL INTERFERENCE". W. United States Patent No. 5,087,975 by Mr. R. W. Citta. A. Patented on May 5, 1998 under the name "DIGITAL TELEVISION RECEIVER WITH ADAPTIVE FILTER CIRCUITRY FOR SUPPRESSING NTSC CO-CHANNEL INTERFERENCE". L. egg. US Patent No. 5,748,226 to A. L. R. Limberg. FIG. 16 shows an ISI suppression filter installed between the NTSC rejection comb filter and the data slicer to compensate for the precoding effect of the NTSC rejection comb filter. Figure 1 of the US Patent No. 5,748,226975 shows an ISI suppression filter installed at the rear end of the NTSC rejection comb filter and data slicer to compensate for the precoding effect of the NTSC rejection comb filter.

Co-channel NTSC interference artifacts occur during synchronous detection of a digital television signal to recover baseband symbol coded data. The artifacts of the video carrier of the co-channel coherent NTSC color TV signal have a frequency of 59.75f _H. Where f _H is the horizontal scanning frequency of the signal. The artifact of the color subcarrier has a frequency of 287.25f _H , and the artifact of an unmodulated NTSC audio carrier has a frequency of 345.75f _H.

The ISI suppression filter is a comb filter designed to match the NTSC cancellation comb filter to remove intersymbol interference introduced by the NTSC cancellation comb filter. The proper operation of the ISI suppression filter depends on the intersymbol interference whose characteristics are known. DTV signal receivers typically include an adaptive channel equalization filter designed to provide matched filtering to suppress intersymbol interference occurring in a transmission channel up to the demodulator, including a demodulator used to recover baseband symbol coded data. Doing. The filter coefficients of the adaptive channel equalization filter are generally initialized according to a training signal method using training signals extracted from data field synchronization (DFS) signals in the initial data segments of the data fields. Co-channel NTSC interference artifacts accompanying the DFS signals act to affect the filter coefficients of the adaptive channel equalization filter to reduce the presence of the filter coefficients. As a result, adversely affects matched filtering for the transmission channel that the adaptive channel equalization filter is expected to perform. Thus, the output signal of the NTSC cancellation comb filter is no longer matched filtered by a comb filter designed to suppress the intersymbol interference introduced by the NTSC cancellation comb filter. Therefore, the bit error rate associated with the conversion of the baseband symbol coded data into the error correction coded data is increased by the unsuppressed intersymbol interference, which is not desirable.

It is an object of the present invention to prevent the initialization of the filter coefficients of an adaptive channel equalization filter in response to a training signal affected by co-channel NTSC interference artifacts accompanying the DFS signals.

The present invention is implemented in a method of operating a channel equalizer of a receiver for a digital television signal subject to co-channel interference from analog television signals. Digital television signals are often demodulated to generate a baseband symbol code signal that will be accompanied by interference artifacts of the co-channel analog television signal. The baseband symbol code signal is comb filtered and symbol decoded to suppress interference artifacts of the co-channel analog television signal. The baseband symbol code signal is also subjected to channel equalization filtering prior to its symbol decoding. The channel equalization filtering causes the entire channel response signal resulting from the comb filtering and channel equalization filtering steps to match the comb filter response signal for the matched filtered transmission channel.

Another aspect of the invention resides in a receiver adapted to operate according to the invention for digital television signals that are co-channel interference from analog television signals. The receiver includes a demodulator device responsive to a received digital television signal to supply a digitized baseband demodulator response signal comprising symbol coded data accompanied by demodulation artifacts of any co-channel interference from the analog television signal. The receiver includes a cascade filter device for supplying a cascade filter response signal to the digitized baseband modulator response signal. The cascade filter device includes an adaptive channel equalization filter having adjustable filtering coefficients with comb filters for suppressing demodulation artifacts of interference from co-channel analog television signals. The receiver also includes a symbol decoder for supplying data in response to the cascade filter response signal and an intersymbol interference suppression filter for processing the data to compensate for intersymbol interference introduced by the comb filter. The receiver also includes an apparatus for extracting a received training signal from the cascade filter response signal during a period in which data field sync signals are generated in the digital television signal. The receiver includes a computer, which computes discrete Fourier transform terms of the training signal. The computer divides the terms into a computer's memory and divides them by corresponding discrete Fourier transform terms of a comb filtered and matched filtered response signal for a training signal that does not include a ghost to characterize the channel. Generate a transformation. The computer also calculates adjustable filtering coefficients of the adaptive channel equalization filter to complement the channel characterization.

According to another aspect of the invention, the invention comprises a detector for determining whether there is significant interference from a co-channel analog television signal. If there is significant interference from the co-channel analog television signal, the computer calculates discrete Fourier transform terms of the training signal as above. However, in this case, the computer stores the terms in the memory of the computer and with corresponding discrete Fourier transform terms of the matched but comb filtered response signal for the training signal that does not contain a ghost. Division produces a discrete Fourier transform that characterizes the channel. The computer also calculates adjustable filtering coefficients of the adaptive channel equalization filter to complement the channel characterization.

1 illustrates a portion of a DTV signal receiver including a symbol decoder having a co-channel NTSC interference detector responsive to a baseband I-channel signal and having a co-channel NTSC interference suppression circuit selectively operating in accordance with an aspect of the present invention. A block diagram showing.

FIG. 2 is an operational flow diagram illustrating a portion of the digital television signal receiver of FIG. 1 showing how the equalization process is modified depending on whether comb filtering is used to suppress co-channel NTSC interference. FIG.

3 illustrates a portion of a DTV signal receiver including a symbol decoder having a co-channel NTSC interference detector responsive to a baseband Q-channel signal and having a co-channel NTSC interference suppression circuit selectively operating in accordance with an aspect of the present invention. A block diagram showing.

4 is an operational flow diagram of the digital television signal receiver of FIG. 3 showing how the equalization process is modified depending on whether comb filtering is used to suppress co-channel NTSC interference.

5 is a flow chart of a routine used in the method of FIG. 2 or FIG. 4 for adjusting the coefficients of the channel equalization filter in response to a training signal when not using comb filtering to suppress co-channel NTSC interference.

6 is a flow chart of a routine used in the method of FIG. 2 or FIG. 4 for adjusting the coefficients of the channel equalization filter in response to a training signal when using comb filtering to suppress co-channel NTSC interference.

FIG. 7 is a block diagram illustrating a specific configuration of the portion of the DTV signal receiver shown in FIG. 1 or 3 related to the circuit for executing the routines of FIGS. 5 and 6.

8 and 9 are block diagrams illustrating a general configuration that the co-channel NTSC interference decoder of FIGS. 1 and 3 may optionally take.

Several parts of the circuits shown in FIGS. 1, 3, 7, 8, and 9 include shimming delay elements for accuracy in the order of operation, as will be appreciated by those skilled in the art of electronic design. You must install it. However, a separate description of the seaming delay device as described above is omitted in the following description unless a separate specific condition is required in relation to the need for a specific seaming delay device.

Figure 1 shows a digital television (DTV) signal receiver for use in restoring error corrected data suitable for recording by a digital video cassette recorder (DVCR) or for MPEG-2 decoding and display in a television set. Although the DTV signal receiver of FIG. 1 is shown as receiving television broadcast signals from the receiving antenna 8, it may also receive signals from the cable network. The television broadcast signals are supplied to the DTV signal receiver front end 10 as an input signal. The "shear" electronics 10 generally converts high frequency television signals into intermediate frequency television signals supplied as input signals to an intermediate frequency (IF) amplifier chain 12 for residual side band DTV (VSB DTV) signals. It includes a high frequency amplifier and a first detector. The DTV signal receiver preferably comprises an intermediate frequency amplifier for amplifying the DTV signals converted by the IF amplifier chain into the ultra high frequency (VHF) band by the first detector, and the amplified DTV signals by ultra high frequency (UHF). And a second detector for converting into a band and another IF amplifier for amplifying the DTV signals converted into the VHF band. When baseband demodulation is done in a digital scheme, the IF amplifier chain 12 also includes a third detector for converting the amplified DTV signals into a final intermediate frequency band closer to the baseband.

Advantageously, a surface-acoustic-wave (SAW) filter is used in the IF amplifier for the UHF band to form a channel select response and to remove adjacent channels. This SAW filter blocks and removes components just beyond 5.38 MHz from the pilot carrier with a similar and suppressed carrier frequency of the VSB DTV signal. Thus, this SAW filter eliminates much of the frequency modulated sound carrier of any cochannel coherent analog TV signal. The elimination of the FM sound carrier of any co-channel coherent analog TV signal in the IF amplifier chain 12 prevents the occurrence of artifacts of the carrier when the final IF signal is detected to recover baseband symbols, Artifacts do not interfere with the data slice processing of the baseband symbols during symbol decoding. Avoiding such artifacts from interfering with the data slicing process of the baseband symbols during symbol decoding can yield better results than comb filtering before the data slicing process. If it is larger than the symbol period, it is prominent.

The final intermediate frequency (IF) output signal from the IF amplifier chain 12 is supplied to the complex demodulator 14, and the DTV signal complex demodulator 14 is finally used to recover the real baseband signal and the imaginary baseband signal. Demodulate the VSB AM DTV signal in the intermediate frequency band. This demodulation may be done in a digital scheme after analog / digital conversion of the final intermediate frequency band in the multi-megacycle range, as described, for example, in US Pat. No. 5,479,449. Alternatively, the above demodulation may be performed in an analog system. In this case, analog / digital conversion processing is always performed on the demodulation result for the convenience of subsequent processing. The complex demodulation preferably comprises in-phase (I) synchronous demodulation and quadrature (Q) synchronous demodulation. Conventionally, the digital results of the demodulation processes above represent 2N level symbols with 8 bits or more of accuracy and which hatch N bits of data. Currently, 2N is 8 when the DTV signal receiver of FIG. 1 receives the radio broadcast signal through the antenna 12, and 16 when the DTV signal receiver of FIG. 1 receives the cable broadcast signal. The present invention relates to the reception of terrestrial radio broadcasting, and FIG. 1 also does not show a portion of the DTV signal receiver that provides symbol decoding and error correction decoding for the received cable broadcast signal.

At least the digitized real samples of the in-phase (I-channel) baseband signals output from the complex demodulator 14 are applied to a symbol synchronization and equalization circuit 16. In the DTV signal receiver of FIG. ) Is also shown to receive digitized imaginary imaginary samples of quadrature (Q-channel) baseband signals. The circuit 16 includes a digital filter with adjustable weighting factors that compensate for the ghost and tilt of the signal received thereon. The symbol synchronization and equalization circuit 16 provides symbol synchronization or "de-rotation" as well as amplitude equalization and ghost cancellation. A symbol synchronization and equalization circuit that is supposed to achieve symbol synchronization before amplitude equalization is known from US Pat. No. 5,479,449. In this configuration, the demodulator 14 will supply an oversampled demodulator output signal including real and imaginary baseband signals to the symbol synchronization and equalization circuit 16. After symbol synchronization, the oversampled data is decimated, thereby extracting baseband I-channel signals at normal symbol rates, and also sampling rates during digital filtering used for amplitude equalization and ghost rejection. (sample rate) can be reduced. Symbol synchronization and equalization circuits in which amplitude equalization precedes symbol synchronization, i.e., "derotation" or "phase tracking," are known to those skilled in the art of digital signal receivers.

Each sample of the output signal of the circuit 16 is decomposed into 10 or more bits, in fact each sample is a digital representation of an analog symbol representing one of 2N (2N = 8) levels. The output signal of the circuit 16 is finely gain controlled using any of several known methods, so the ideal step levels for the symbols are known. In the case of a gain control method selected among the known gain control methods because the gain control response speed is exceptionally fast, the DC components of the real baseband signal supplied from the complex demodulator 14 are adjusted to a nominal level of +1.25. This gain control method is generally described in US Pat. No. 5,479,449. More specifically, C.B., incorporated herein by reference on June 3, 1997, entitled "AUTOMATIC GAIN CONTROL OF RADIO RECEIVER FOR RECEIVING DIGITAL HIGH-DEFINITION TELEVISION SIGNALS". US Pat. No. 5,573,454 to C.B. Patel.

The output signal from the circuit 16 is supplied as an input signal to the data synchronization detector 18, which synchronizes the data field synchronization information (DFS) and data segment synchronization from the equalized baseband I-channel signal. Restore the information (DSS). Alternatively, an input signal to the data synchronization detector 18 may be obtained prior to equalization.                 

Equalized I-channel signal samples supplied at the normal symbol rate as the output signal from the circuit 16 are applied as an input signal to the NTSC cancellation comb filter 20. The comb filter 20 linearly combines the delayed symbol streams to generate a response signal of the comb filter 20 and a first delayer 201 for generating a pair of delayed pairs of 2N-level symbols. The first linear coupler 202 is included. As described in US Pat. No. 5,260,793, the first delay 201 may provide a delay equal to the duration of twelve 2N-level symbols, and the first linear combiner 202 is a subtractor date. Can be. Each sample of the comb filter 20's output signal is decomposed into 10 or more bits, in fact each sample is a digital representation of an analog symbol representing one of "4N-1" (4N-1 = 15) levels. .

The symbol synchronization and equalization circuit 16 is expected to be designed to suppress the direct current bias component of its input signal (i.e., the direct current term of the system function represented as digital samples). Accordingly, each sample of the output signal of the symbol synchronization and equalization circuit 16 applied as an input signal of the comb filter 20 has the following nominal levels: -7, -5, -3, -1, +1. It will be a digital representation of an analog symbol representing one of +3, +5, or +7. These symbol levels are referred to as "odd" symbol levels and are detected by the odd level data slicer 22 to generate intermediate symbol decoding results corresponding to 000, 001, 010, 011, 100, 101, 110, and 111, respectively. do.

Each sample of the output signal of the comb filter 20 actually has the following nominal levels: -14, -12, -10, -8, -6, -4, -2, 0, +2, +4, Digital representation of an analog symbol representing one of +6, +8, +10, +12, or +14. These symbol levels are referred to as "even" symbol levels and comb filtered symbols corresponding to 001, 010, 011, 100, 101, 110, 111, 000, 001, 010, 011, 100, 101, 110, 111, respectively. It is detected by the even level data slicer 24 to generate decoding results.

The data slicers 22 and 24 may be configured in a so-called "hard-decision" form, as assumed heretofore, or used to implement the Viterbi decoding method. It may also be of the so-called "soft-decision" type. The odd-level data slicer 22 and the even-level data slicer 24 are single data slicers that use multiplexer concatenation to shift positions within the circuit and to provide bias for varying slice ranges. It is also possible to replace a), but such a configuration is not preferable because it complicates operation.

In the description so far, the symbol synchronization and equalization circuit 16 is expected to be designed to suppress the DC bias component of the input signal (the DC term of the system function expressed as digital samples). The DC bias component has a nominal level of +1.25 and appears in the real baseband signal supplied from the complex demodulator 14 due to the detection of the pilot carrier. In practice, the symbol synchronization and equalization circuit 16 is designed to at least partially preserve the direct current bias component of its input signal, which can simplify the equalization filter of the symbol synchronization and equalization circuit 16 somewhat. Thus, the data slice levels in the odd level data slicer 22 are offset to account for the DC bias component that accompanies the data steps in its input signals. As long as the first linear coupler 202 is a subtractor, whether the circuit 16 is designed to suppress the DC term of the system function of its input signal or to preserve such DC term is determined by an even-level data slicer ( It does not matter with regard to the data slice levels in 24). However, assuming that the differential delay provided by the first delayer 201 is selected so that the first linear combiner 202 is an adder instead, the data slice levels in the even-level data slicer 24 are equal to their input signals. It must be offset to take into account the direct current term accompanying the data steps within.

An intersymbol interference suppression comb filter 26 is provided at the rear of the data slicers 22 and 24 to generate a filter response signal in which intersymbol interference (ISI) is introduced by the comb filter 20. The ISI suppression comb filter 26 is a second having a three-input multiplexer 261, a second linear combiner 262, and the same delay as the first delayer 201 in the NTSC rejection comb filter 20. And a retarder 263. The second linear coupler 262 is a modulo-8 adder if the first linear coupler 202 is a subtractor, and a modulo-8 subtractor if the first linear coupler 202 is an adder. The first linear combiner 202 and the second linear combiner 262 may be configured as ROMs, respectively, to allow the linear combining operation speed to be high enough to support the associated sample rates. The output signal from the multiplexer 261 is supplied as a response signal of the ISI suppression comb filter 26 and is delayed by the second delay unit 263. The second linear combiner 262 combines the precoded symbol decoding results of the even-level data slicer 24 with the output signal of the second delayer 263.

The output signal of the multiplexer 261 is applied to the multiplexer 261, which is selected in response to the first, second, and third states of the multiplexer control signal supplied from the controller 28 to the multiplexer 261. Play one of the input signals. The first input port of the multiplexer 261 has a controller (d) in which data field synchronization information (DFS) and data segment synchronization information (DSS) are recovered from the baseband I-channel signal equalized by the data synchronization detection circuit 18. Receive ideal symbol decoding results supplied from memory in 28). The controller 28 supplies the multiplexer control signal in the first state to the multiplexer 261 during the period, so that the ideal symbol decoding results supplied from the memory in the multiplexer 261 to the controller 28 are the final coding as their output signals. Control to feed as a result. The odd-level data slicer 22 supplies the intermediate symbol decoding results, which are its output signals, to the second input port of the multiplexer 261. The multiplexer 261 is controlled by the multiplexer control signal in the second state to reproduce intermediate symbol decoding results of the final coding results supplied therefrom. The second linear combiner 262 supplies the output signal of the ISI suppression filtered symbol decoding results thereof to the third input port of the multiplexer 261. The multiplexer 261 is controlled by the multiplexer control signal in the third state to reproduce the ISI suppression filtered symbol decoding results in the final coding results supplied from the multiplexer 261. Running errors in the ISI suppression filtered symbol decoding results output from the ISI suppression comb filter 26 may cause memory errors in the controller 28 during the period in which the data synchronization detection circuit 18 recovers DSS or DFS synchronization information. Reduced by feedback the ideal symbol decoding results supplied from

The output signal of the multiplexer 261 of the ISI suppression comb filter 26 includes the final symbol decoding results of the 3-parallel bit groups assembled by the data assembler 30 for application to the trellis decoder circuit 32. do. Typically, the trellis decoder circuit 32 uses twelve trellis decoders. The trellis decoding results are supplied from the trellis decoder circuit 32 to the data de-interleaver circuit 34 for de-commutation. The output signal of the data deinterleaver circuit 34 is converted into bytes of Reed Solomon error correction coded data for application to the Reed Solomon decoder circuit 38 by the byte construction circuit 36. The Reed Solomon detector circuit 38 performs Reed Solomon decoding so as to generate an error-corrected byte stream supplied to the data derandomizer 40. The data derandomizer 40 supplies the reproduced data to the rest of the receiver (not shown). The rest of the complete DTV signal receiver will include a packet sorter, audio decoder, MPEG-2 decoder, and so forth. The remaining portions of the DTV signal receiver used in the digital tape recorder / player may include circuitry for converting the data into a recordable form.

Co-channel NTSC interference detector 44 is used, which is insensitive to the DC bias component of its input signal and is an artifact caused by co-channel NTSC interference in the input signal. It is used to detect their intensity. The co-channel NTSC interference detector 44 sends a signal to the controller 28 indicating whether the co-channel NTSC interference has sufficient strength to cause an uncorrectable error in the data slice processing performed by the data slicer 22. Supply. If detector 44 indicates that co-channel NTSC interference does not have the strength as described above, controller 28 will supply a multiplexer control signal in a second state to multiplexer 261 for most of the time period. If the period is not within this period, the data field synchronization information DFS and the data segment synchronization information DSS are restored by the data synchronization detection circuit 18 so that the controller 28 sends the multiplexer control signal in the first state to the multiplexer 261. Is the period to apply. The multiplexer 261 is controlled to reproduce intermediate symbol decoding results supplied from the odd level data slicer 22 as its output signal by the multiplexer control signal in the second state.

If detector 44 indicates that co-channel NTSC interference has a strength sufficient to cause an uncorrectable error in the data slice processing performed by data slicer 22, controller 28 is for the most part The multiplexer control signal in the third state will be supplied to the multiplexer 261 during the process. If the period is not within this period, the data field synchronization information DFS and the data segment synchronization information DSS are restored by the data synchronization detection circuit 18 so that the controller 28 sends the multiplexer control signal in the first state to the multiplexer 261. Is the period to apply. The multiplexer 261 reproduces the ISI suppression filtered symbol decoding results provided as the second linear combination results from the second linear combiner 262 as its output signal by the multiplexer control signal in the third state.

FIG. 2 is a flowchart illustrating how equalization processes of the DTV signal receiver of FIG. 1 vary depending on whether comb filtering is used to suppress co-channel NTSC interference. The inventors have found that if co-channel NTSC interference artifacts are present in the baseband symbol coded data, an error will occur in the calculation without special measures to invalidate such artifacts in the calculation of the equalization filter kernel coefficients. Pointing out.

In the initial step S1, the complex demodulator 14 provided in the DTV signal receiver of Fig. 1 performs complex demodulation of the digital television signals, and is orthogonal to the received I-channel baseband signal and the received I-channel baseband signal. Separate received Q-channel baseband signals with respect. Similarly to the initial step S1, in the determination step S2, which is successively performed by the co-channel NTSC interference detector 44 provided in the DTV signal receiver of FIG. 1, a substantial amount of the same channel for the received Q-channel baseband signal. It is determined whether NTSC interference is accompanied.

Here, a significant amount of co-channel NTSC interference to the received Q-channel baseband signal is such that the error is generated during trellis decoding to a significant extent that reduces the error correction capability of the two-dimensional Reed Solomon decoding following trellis decoding. It means co-channel NTSC interference. In a reception state where normal background noise is present, a significant number of bit errors are ultimately included in the recovered data. In the case of a DTV signal receiver designed with a specific configuration, such a significant amount of co-channel NTSC interference can be easily determined by experimenting with the prototype.

If it is determined at step S2 that the received I-channel baseband signal is not accompanied by a significant amount of co-channel NTSC interference, adjusting the kernel weight of the digital equalization filter (S3) and (S3). A subsequent step S4 of symbol decoding the equalization filter response signal obtained in step S4 is performed. When the step S3 of adjusting the kernel weight is performed, the digital equalization filter provides a matched response signal for the I-channel baseband signal. Symbol decoding of the response signal of the equalization filter (S4) generates a symbol decoding result, which is trellis decoded in a subsequent step (S5) to correct the error. Resolomon decoding the trellis decoding result to correct errors included in the trellis decoding result (S6) following the trellis decoding step (S5), and deformatting the reedsolomon decoding result ( Deformatting is performed (S7).

If it is determined at step S2 that the received I-channel baseband signal is accompanied by a significant amount of co-channel NTSC interference, then the received I-channel baseband signal is generated to generate a comb filtered receive I-channel baseband signal. The comb filtering step S8 is performed using a suitable comb filter. In step S9, channel equalization filtering is performed so as to provide an ideal comb filter response signal for the I-channel baseband symbol code matched and filtered by the overall channel characteristics. That is, the kernel weights of the digital equalization filter are adjusted to match the response signals of the cascaded digital equalization filter and comb filter to the ideal response signal for such a filter cascade. Subsequently, symbol decoding of the filter cascade response signal as described above is performed (S10), and then the symbol decoding response signal is post-processed to obtain a corrected symbol decoding result for use in the trellis decoding step (S5). Coding is performed (S11). The post coding of step S11 compensates for the precoding resulting from the comb filtering of step S8 and suppresses the intersymbol interference associated with that precoding. In this case as well, following the trellis decoding step S5, the Reed Solomon decoding step S6 for correcting the errors included in the trellis decoding result, and the step of reformatting the Reed Solomon decoding result S7 are of course possible. Do it.

The secondary method used to adjust the kernel weights of the digital equalization filter in the step of equalizing the digital equalization filter response signal is similar to the weight adjustment method of the digital equalization filter used in the prior art. The above adjustment is performed by calculating the discrete Fourier transform (DFT) of the received data field sync code or predetermined portion thereof and dividing the calculated DFT by the ideal data field sync code or DFT of the predetermined portion thereof. Can be determined. The DFT of the DTV transport channel is normalized with respect to the maximum term (s) to characterize the channel, and the kernel weights of the digital equalization filter are selected to complement the normalized DFT characterizing the channel. This adjustment is an example of a C.B. patent issued on July 19, 1994 under the name "METHODS FOR OPERATING GHOST-CANCELATION CIRCUITRY FOR TV RECEIVER OR VIDEO RECORDER". It is described in detail in US Pat. No. 5,331,416 to C.B. Patel. The method is suitable for the initial manipulation of the kernel weights of the digital equalization filter, since this initial adjustment is easier than the usual case of using adaptive equalization. It is preferable to perform an adaptive equalization method after the initial adjustment of the kernel weights of the digital equalization filter. J. Patented July 15, 1997 under the name "RAPID-UPDATE ADAPTIVE CHANNEL-EQUALIZATION FILTERING FOR DIGITAL RADIO RECEIVERS, SUCH AS HDTV RECEIVERS." US Patent No. 5,648,987 to J. Yang, describes a block LMS method for performing adaptive equalization. Filed April 4, 1997 under the name "DYNAMICALLY ADAPTIVE EQUALIZER SYSTEM AND METHOD." L. egg. US patent application Ser. No. 08 / 832,674 to A. L. R. Limberg describes a continuous block LMS method for performing adaptive equalization.

In step S9, implement the secondary method of adjusting the kernel weights of the digital equalization filter to match the response signals of the cascaded digital equalization filter and comb filter to the ideal response signal for such a filter cascade. You can use DFT to do that. The DFT is particularly useful for fast initial equalization using the Data Field Synchronization (DFS) code or some portion thereof as a training signal prior to switching to adaptive equalization. The equalization filter coefficients are set to predetermined values during initialization, and thus the input signal thereof is reproduced by the filter response signal. Compute the Discrete Fourier Transform (DFT) of the received data field sync code or predetermined portion thereof in the comb filtered state by the comb filter 20 to remove NTSC artifacts. The DFT of the DTV transmission channel can be determined by dividing this DFT by the ideal DFS code in a comb filtered state or the DFT of a predetermined portion thereof. The DFT of the DTV transport channel is normalized with respect to the maximum term (s) to characterize the channel, and the kernel weights of the digital equalization filter are selected to complement the normalized DFT characterizing the channel. After initial adjustment of the kernel weights of the digital equalization filter, an adaptive equalization method is preferably used. These adaptive equalization methods double the number of valid signal states possible by using a comb filter 20 to remove NTSC artifacts compared to the adaptive equalization methods used when co-channel NTSC interference artifacts are not important. It is different in that it is.

3 shows a DTV signal receiver different from the DTV signal receiver of FIG. 1 in that a baseband Q-channel signal is applied to the co-channel NTSC interference detector 44 as its input signal instead of the baseband I-channel signal. The co-channel NTSC interference detector 44 is used to detect the strength of artifacts due to co-channel NTSC interference in the baseband Q-channel signal. The detection response signal of the co-channel NTSC interference detector 44 may appear in the baseband Q-channel signal during a period in which the phase-lock of the synchronous detectors provided in the complex demodulator 14 should continue to be established. It is insensitive to a DC bias component. Thus, there is no switching between the baseband signal and the comb filtered baseband signal when calculating the weighting factors for equalization filtering in the circuit 16. Following the DTV signal acquisition of the DTV signal receiver, any DC bias component that may appear in the baseband Q-channel signal (e.g. due to a poor phase lock occurring during the reception of a weak strength signal) may cause a co-channel NTSC interference detector ( 44) will not affect the detection response. In the case of the DTV signal receiver of FIG. 3, whether the received I-channel baseband signal is accompanied by a significant amount of co-channel NTSC interference or not, whether the received Q-channel baseband signal is accompanied by a significant amount of co-channel NTSC interference. To make a deduction.

FIG. 4 is a flowchart illustrating how the equalization process is modified in the digital television signal receiver of FIG. 3 according to whether comb filtering to suppress co-channel NTSC interference is used. 4, which relates to the DTV signal receiver of FIG. 3, includes a decision step S2 for determining whether a significant amount of co-channel NTSC interference is accompanied by the received I-channel baseband signal to the received Q-channel baseband signal. There is a difference from the flowchart of FIG. 2 with respect to the DTV signal receiver of FIG. 1 in that it is replaced by a decision step S20 that determines whether a significant amount of co-channel NTSC interference is accompanied.

5 is a flowchart of a known routine used in step S3 of equalizing the channel response signal in the method of FIG. 2 or FIG. The step S3 comprises a series of substeps starting from the substep 31 of extracting a training signal from the DFS signal at the beginning of each data field. The DFS signal from which the training signal is extracted is a signal generated by the complex demodulation step S1. Following the substep 31, the substep of accumulating the training signal extracted to generate a received ghost-cancellation reference (GCR) signal including the sub-ghosts over a predetermined even number of data fields ( 32) is executed. When the GCR signal is to be an intermediate PN63 sequence of an ATSC standard DFS signal, the polarity of the DFS signal is kept the same every other accumulation step. When the GCR signal is to be a PN511 sequence of an ATSC standard DFS signal, the polarity of the DFS signal is always kept the same in each accumulation step. Subsequently, a substep S33 of calculating a discrete Fourier transform (DFT) of the received GCR signal with subordinate ghosts is executed. Subsequently, in the substep S34 of characterizing the transmission channel by the DFT, the DFT of the matched filter response signal for the ideal GCR signal from which the accessory ghosts have been removed is extracted from the memory of the DTV receiver. Subsequently, the terms of the DFT of the received GCR signal including subsidiary ghosts in the sub-step S34 may be used to generate respective terms of the DFT characterizing the transmission channel. Are divided by the corresponding terms of the DFT. Finally, in step S35, channel equalization filtering coefficients are calculated to complement the terms of the DFT characterizing the transport channel.

6 is a flowchart of a modification routine of the routine of FIG. 5 used in step S3 of equalizing the channel response signal. This modification routine is used to execute the channel equalization filtering step S9 to generate an ideal comb filter response signal for the matched filtered transmission channel to be supplied for symbol decoding in step 10 in accordance with the present invention. The step S9 comprises a series of substeps starting from the substep 91 of extracting a training signal from the comb filtered DFS signal at the beginning of each data field. The DFS signal from which the training signal is extracted is not a DFS signal directly generated by the complex demodulation step S1 but a signal generated by the comb filtering step S8. Subsequent to step 91, the training signal extracted to generate a comb-filtered received ghost-cancellation reference (GCR) signal including sub-comb-filtered ghosts is defined as a predetermined even number of data fields. Sub-steps 92 of accumulating over the fields are executed. When the GCR signal is to be an intermediate PN63 sequence of an ATSC standard DFS signal, the polarity of the DFS signal is kept the same every other accumulation step. When the GCR signal is to be a PN511 sequence of an ATSC standard DFS signal, the polarity of the DFS signal is always kept the same in each accumulation step. Subsequently, a substep S93 of calculating the DFT of the received GCR signal with the subordinate ghosts is executed. Subsequently, in the substep S94 of characterizing the transmission channel by the DFT, the DFT of the comb filtered matched filter response signal for the ideal GCR signal from which the accessory ghosts have been removed is extracted from the memory of the DTV receiver. Subsequently, the terms of the DFT of the received GCR signal including the subsidiary ghosts in the sub-step S94 to generate respective terms of the DFT characterizing the transport channel are comb filtered matched filters for the ideal GCR signal. The corresponding terms of the DFT of the response signal are divided. Finally, in step S95 the channel equalization filtering coefficients are calculated from the inverses of the terms of the DFT characterizing the transport channel.

7 shows a specific configuration of a circuit used to perform the routines shown in FIGS. 5 and 6. The GCR signal received with the included ghosts is generated by an accumulator 50 that accumulates corresponding symbols of the DFS signals in the initial data segments of an even number of data fields. The polarity of the DFS signal is kept the same every other accumulation step. When the GCR signal is to be a PN511 sequence of an ATSC standard DFS signal, the polarity of the DFS signal is always kept the same in each accumulation step.                 

If the co-channel NTSC interference detector 44 determines that the received I-channel baseband signal is not accompanied by a significant amount of co-channel NTSC interference, then the DFS signal accumulated by the adder 50 is taken out of the NTSC cancellation comb filter 20. Extracted from the signal input to the via DFS gate circuit 51. If the co-channel NTSC interference detector 44 determines that the received I-channel baseband signal is not accompanied by a significant amount of co-channel NTSC interference, then the DFS signal accumulated by the adder 50 is taken out of the NTSC cancellation comb filter 20. Extracted from the signal output through the DFS gate circuit 52. The data segment counter 53 generates a coefficient value indicating which data segments in the frame group are being received so as to facilitate these processes; The decoder 54 then generates a response signal to the count value. The decoder 54 generates a logic " 1 " output in response to the count value from the counter 53 indicating that the data segment currently being received is the initial data segment of the data field. The decoder 54 also generates a logic "0" output in response to the count value from the counter 53 indicating that the data segment currently being received is a subsequent data segment of the data field.

The co-channel NTSC interference detector 44 supplies a logical " 0 " output signal if it determines that the received I-channel baseband signal is not accompanied by a significant amount of co-channel NTSC interference. A shift resist stage 56 responds to the " O " supplied at the beginning of the initial segment of the data field to apply a logic " 0 " to the input of the logic inverter 55 over the entire initial data segment. . The logic inverter 55 supplies logic " 1 " to the first input terminal of the two-input AND gate 57 in response to " 0 ". This " 1 " causes the AND gate 57 to output a logic " 1 " when the output signal of the decoder 54 received at its second input terminal is a logic " 1. " This condition occurs in response to the count value from the counter 53 indicating that the currently received data segment is the initial data segment of the data field. The DFS gate circuit 51 is controlled to supply the accumulator 50 with the DFS signal extracted from the input signal of the NTSC cancellation comb filter 20 by the output signal of the AND gate 57 which is "1".

The co-channel NTSC interference detector 44 supplies a logical " 1 " output signal if it determines that the received I-channel baseband signal is accompanied by a significant amount of co-channel NTSC interference. The shift resist stage 56 responds to the " 1 " supplied at the beginning of the initial segment of the data field to apply a logic " 1 " to the input of the inverter 55 across the initial data segment. The logic inverter 55 generates a logic "0" output signal as a response signal, and by this signal, the AND gate 57 is controlled to supply a logic "0" as its output signal. Due to this " 0 " AND gate 57 output signal, the DFS gate circuit 51 does not supply the DFS signal extracted from the input signal of the NTSC cancellation comb filter 20 to the accumulator 50. Controlled to indicate source impedance. Due to the high source impedance represented by the DFS gate circuit 51, the DFS gate circuit 52 supplies the accumulator 50 with the DFS signal extracted from the output signal of the NTSC cancellation comb filter 20.

The output signal of the shift resist stage 56 is applied to the first input terminal of the two-input AND gate 58. When the shift resist stage 56 supplies a logic " 1 " output signal to the first input terminal of the AND gate 58, the AND gate 58 is an output signal of the decoder 54 that is received at its second input terminal. Is controlled to output a logic "1" when is a logic "1". This condition is generated in response to the count value of the counter 53, which indicates that the currently received data segment is the initial data segment of the data field. The DFS gate circuit 52 is controlled to supply the accumulator 50 with the DFS signal extracted from the output signal of the NTSC removal comb filter 20 by the output signal of the AND gate 58 which is "1".

When the shift resist stage 56 supplies a logic " 0 " output signal to the first input terminal of the AND gate 58, the " 0 " causes the AND gate 58 to convert logic " 0 " as its output signal. Controlled to feed. This output signal from the AND gate 58 causes the DFS gate circuit 52 to provide a high source impedance such that the DFS signal extracted from the output signal of the NTSC rejection comb filter 20 is not supplied to the accumulator 50. Controlled to indicate. Due to the high source impedance represented by the DFS gate circuit 52, the DFS gate circuit 51 supplies the accumulator 50 with the DFS signal extracted from the input signal of the NTSC cancellation comb filter 20.

The shift resist stage 56 performs a shift in response to a shift command generated by the two-input AND gate 59 at the beginning of the initial segment of each data field. The AND gate 59 receives the output signal of the decoder 54 at its first input, which has a state of logic " 1 " entirely during the initial data segment of each data field. The second input of the AND gate 59 receives a pulse supplied to the data segment counter 53 at the end of each data segment. The pulse is supplied by a decoder responsive to the sample counter, which is not clearly shown in FIG.

When accumulator 50 generates update data of the received GCR signal with accompanying ghosts, the update data is loaded into a first input register of the small computer 60 embedded in the DTV signal receiver. The computer 60 is programmed to calculate the DFT of the received GCR signal and its associated ghosts. The computer 60 receives via the input line an output signal of the decoder 54 indicating when the initial data segment of the current data field is being received. Computer 60 receives the shift command generated by AND gate 59 via another input line. Although not explicitly shown in FIG. 7, the computer 60 also receives sample clocking information. 7 provides the computer 60 upon request a DFT of a comb filtered ghost canceled matched GCR signal used to characterize the transmission channel when it is determined that co-channel NTSC interference is not significant. A memory 61 is shown for this purpose. Figure 7 also shows the DFT of the comb filtered ghost canceled matched GCR signal used to characterize the transport channel when cochannel NTSC interference is determined to be significant for supplying the computer 60 upon request. Memory 62 is shown. The memories 61, 62 may conveniently be configured as a single ROM addressed by a sample count value and by an additional bit indicating which DFT is being requested.

In order for the received GCR signal with accessory ghosts generated by accumulator 50 to be useful, the received GCR signal accumulates initial data segments over 2N consecutive data fields with the same specific attributes. Must be obtained. In addition, it must be in a state that is determined to be equivalent to all of the 2N consecutive fields in which co-channel NTSC interference has finally occurred, or not significant anywhere in the continuous fields. 7 shows a circuit for determining whether the received GCR signal with accessory ghosts generated by accumulator 50 is useful. This circuit consists of a number of 63 (2N-1) additional serial-in / parallel-out (SPIO) type shift register stages, 2N-input NOR gate 64, and 2N. An input AND gate 65 and a two input OR gate 66 are included.

The NOR gate 64 receives the contents of each of the shift register stage 56 and the 63 plurality of additional SIPO type shift register stages as input signals, respectively. The NOR gate 64 supplies a logical " 1 " output signal to the OR gate 66 exclusively as a first input signal only in the state where no co-channel NTSC interference is determined to be significant in any of the most recent 2N consecutive fields. do. The OR gate 66 receives the output signal of the NOR gate 64 as a first input signal, and receives auxiliary ghosts generated by the accumulator 50 in response to the first input signal of "1". Informs the computer 60 that the received GCR signal it has is useful. The output signal of the NOR gate 64 is also applied to the computer 60. By the output signal of the NOR gate 64, which is " 1 ", the computer 60 is instructed to use the DFT read from the memory 61 in the calculation of the DFT characterizing the transmission channel.

The AND gate 65 also receives respective contents of the shift register stage 56 and the plurality of 63 additional SIPO type shift register stages as input signals, respectively. The AND gate 65 supplies a logical " 1 " output signal to the OR gate 66 exclusively as a first input signal only in a state where co-channel NTSC interference is determined to be significant in all of the most recent 2N consecutive fields. The NOR gate 65 receives the output signal of the AND gate 65 as a second input signal and receives auxiliary ghosts generated by the accumulator 50 in response to the second input signal of "1". Informs the computer 60 that the received GCR signal it has is useful. The output signal of the AND gate 65 is also applied to the computer 60. The output signal of the AND gate 65, which is " 1 ", causes the computer 60 to instruct the use of the DFT read from the memory 62 in the calculation of the DFT characterizing the transmission channel.

More sophisticated methods may be used as a method of validating the output signal of the accumulator 50, in which the receive GCR signal having the accumulator 50 having the accessory ghosts generated by the accumulator 50 Available by incorporating into the computer 60 together with circuitry to determine whether or not it is useful. The single-bit output signal from the shift register 56 is supplied to a number of additional SIPO shift register stages embedded in the computer 60. The computer 60 may be programmed to determine a value of 2N that should be used to determine whether the received GCR signal with accessory ghosts generated by the accumulator 50 is useful. Computer 60 may also be programmed to retreat to a previous cumulative result when determining that the current cumulative result is not useful.                 

8 shows the general form that the co-channel NTSC interference detector 44 may take in the DTV signal receivers of FIGS. 1 and 3. An input signal to the detector 44 is received at node 440. This input signal may be an equalized I-channel or Q-channel baseband signal supplied from the symbol synchronization circuit 16 as in each of the DTV signal receivers of FIGS. 1 and 3. Alternatively, the input signal may be an I-channel or Q-channel baseband signal supplied without equalization from the complex demodulator 14 in the modified embodiment of the DTV signal receiver of FIG. 1 or 3. In the NTSC cancellation comb filter embedded in the detector 44, the third delayer 441 receives an input signal applied to the node 440 to generate the subtracted and subtracted input signals to the digital subtractor 442. Delay the derivative. The difference output signal from the subtractor 442 is the NTSC cancellation comb filter response signal R with artifacts suppressed by synchronous detection of the co-channel coherent analog television signal suppressed. For example, the third delayer 441 includes 12 symbol periods, 1368 symbol periods (corresponding to two NTSC video scanning lines), and 179,208 symbol periods (corresponding to 262 NTSC video scanning lines). Period), or a delay corresponding to 718,200 symbol periods (a period corresponding to two NTSC video frames) may be introduced. In the NTSC selective comb filter built into the detector 44, the fourth delayer 443 differentiates the input signal applied to the node 440 to generate the subtracted and subtracted input signals to the digital subtractor 444. Delay. The difference output signal from subtractor 444 is an NTSC select comb filter response signal S with enhanced artifacts caused by synchronous detection of co-channel coherent analog television signals. For example, the fourth delayer 443 may introduce a delay corresponding to six symbol periods. The DC term of the system characteristic resulting from the synchronous detection of the pilot carrier is suppressed in both the NTSC removal comb filter response signal R and the NTSC selection comb filter response signal S.

The amplitude of the NTSC removal comb filter response signal R output from the subtractor 442 is detected by an amplitude detector 445, and the amplitude of the NTSC selection comb filter response signal S output from the subtractor 444 is It is detected by the amplitude detector 446. The amplitude detection results of the amplitude detectors 445 and 446 are compared with each other by the amplitude comparator 447. As a result, the amplitude comparator 447 outputs the amplitude detector 446 to the amplitude detector 445. An output bit indicating whether or not the output signal is substantially larger than the output signal is generated. This output bit is used to select one of the second and third operating states of the multiplexer 261. For example, the output bit output from the amplitude comparator 447 may be one of two control bits supplied by the controller 28 to the multiplexer 261 of the ISI suppression comb filter 26 of FIG. 1 or 3. Can be. The other control bit indicates whether the signal supplied from the controller 28 should be reproduced in the output signal of the multiplexer 261.

The amplitude detectors 445 and 446 may be, for example, envelope detectors having a time constant equal to several data sample periods, in which case the difference between the data components of their input signals is assumed to be a low value assumed to be random. Tend to be averaged. The amplitude difference between the random noises accompanying the difference output signals of the subtractors 442 and 444 tends to be averaged up to zero. Thus, if the amplitude comparator 447 indicates that the amplitude detection response signals of the amplitude detectors 445 and 446 differ from each other by more than a predetermined value, this indicates that the artifacts of any co-channel coherent analog television signal are not affected by the node 440. It is higher than the significant level (significant level) of the baseband signal supplied to. This significance level corresponds to the significance level of the equalized I-channel baseband signal applied to the odd level data slicer 22. As long as the artifacts of the co-channel coherent NTSC signal remain below the significance level, errors generated during symbol decoding by simply data-slicing the I-channel baseband signal are trellis-coded and Reedsolomon error-corrected coding. I can correct it.

The co-channel NTSC interference artifacts are removed from the comb filter response signal R output from the subtractor 442 and selected from the comb filter response signal S output from the subtractor 444. If the amplitude of the comb filter response signal (S) is substantially greater than the amplitude of the comb filter response signal (R), this difference is expected to be caused by the presence of co-channel NTSC interference artifacts in the signal at node 440. can do. The output bit supplied by amplitude comparator 447 in this state controls the multiplexer 261 inoperable in its second state, resulting in the selection of intermediate symbol decoding results output from the odd level data slicer 22. The intermediate symbol decoding results are not output from the multiplexer 261 as the final symbol decoding results.

If the amplitude of the comb filter response signal S is not substantially greater than the amplitude of the comb filter response signal R, no such amplitude difference indicates that there are no co-channel NTSC interference artifacts in the signal at node 440. It can be expected to indicate. The output bits supplied by amplitude comparator 447 in this state control the multiplexer 261 inoperable in its third state, resulting in an ISI suppression filtered symbol decoding output from the second linear combiner 262. Deselect the results so that the intermediate symbol decoding results are not output from the multiplexer 261 as the final symbol decoding results.

9 shows an alternative form 44 'that the co-channel NTSC interference detector 44 may take in the DTV signal receivers of FIGS. 1 and 3. The subtractors 442 and 444 are replaced by adders 448 and 449. In this variant, the third delayer 441 ′ can introduce a shorter delay corresponding to six symbol periods, for example. In addition, a delay corresponding to, for example, 12 symbol periods, 1368 symbol periods, 179,208 symbol periods, or 718,200 symbol periods may be introduced by the fourth delay unit 443 '.

In the case of alternative embodiments of the present invention as described above, equalization filtering is provided to facilitate the selection switching between the outputs of the data slicers 22 and 24 when the co-channel NTSC interference gradually increases or decreases periodically. This is done prior to NTSC removal comb filtering. Embodiments of the invention in which equalization filtering is performed after NTSC removal comb filtering are also conceivable.

The present invention is usefully used in DTV signal receivers using adaptive channel equalizers configured to initialize the filter coefficients using a training signal method and then convert them using a decision feedback method to adjust the filter coefficients. As long as the decision feedback error signal is generated correctly, the effect of the co-channel NTSC interference artifacts on the decision feedback method is smaller than that of the co-channel NTSC interference artifacts on the training signal method. However, if the training signal method is modified according to the present invention, the filter coefficient initialization using the training signal method is always performed in a much shorter time than the filter coefficient initialization using the decision feedback method.

The present invention has been described so far as using symbol decoders configured to make decisions in a "strict decision" manner that directly depends on the output of the data slicer. However, there are other embodiments of the present invention, for example, in which symbol decoding is performed in a "light decision" manner using a Viterbi algorithm. Such embodiments of the invention should also be included in the following claims.

The present invention has been described with reference to alternative embodiments in which an ISI suppression filter is provided at the rear of the data slicer to compensate for the precoding effect of the NTSC removal comb filter. In some embodiments of the present invention, the ISI suppression filter is provided between the NTSC elimination comb filter and the data slicer to compensate for the precoding effect of the NTSC elimination comb filter. These embodiments use a configuration similar to that shown in FIG. 16 of the drawing of US Pat. No. 5,087,975. Such embodiments of the invention should also be included in the following claims.

The present invention has been described in the context of alternative embodiments where channel equalization filtering is to be performed at the base. However, one of ordinary skill in the art of digital receiver design will be able to design embodiments of the invention in which passband channel equalization filtering is performed at a low intermediate frequency, in light of the above description. These embodiments should also be included within the scope of the following claims, and are not intended to clearly indicate that channel equalization filtering is performed at each baseline within each claim.

The present invention is used for adjusting channel equalization circuits used in receivers of digital television signals that are wirelessly broadcast using the same channel as that of NTSC analog television signals.

Claims (14)

A method of operating a receiver for a digital television signal subject to co-channel interference from an analog television signal,  Demodulating the digital television signal to generate a baseband symbol code signal, often accompanied by interference artifacts of the co-channel analog television signal; Symbol decoding the baseband symbol code signal; Comb filtering the baseband symbol code signal prior to the symbol decoding to suppress the interference artifacts of the co-channel analog television signal in the baseband symbol code signal; Channel equalization filtering the baseband symbol code signal prior to the symbol decoding to match the entire channel response signal resulting from the comb filtering and channel equalization filtering steps with the comb filter response signal for the matched filtered transmission channel. A method of operating a receiver for a digital television signal subject to co-channel interference, characterized in that it comprises. The method of claim 1, The channel equalization filtering step is performed by an adaptive channel equalization filter having adjustable filter coefficients, and the comb filtering step is performed by an NTSC cancellation comb filter cascaded with the channel equalization filter. A method of operating a receiver for digital television signals subject to channel interference. The method of claim 2, The channel equalization filtering step includes a routine for adjusting the adjustable filter coefficients of the adaptive channel equalization filter. Extracting, from the comb filter response signal, a training signal included in a data field synchronization signal intervening in the baseband symbol code signal at the beginning of each data field; Accumulating training signals over an even number of data fields to generate a ghost erase reference signal having accessory ghosts; Calculating a discrete Fourier transform of the ghost cancellation reference signal with associated ghosts; The terms of the discrete Fourier transform of the ghost cancellation reference signal with associated ghosts to the corresponding terms of the discrete Fourier transform of a matched and comb filtered ghost cancellation reference signal supplied from a memory of the digital television signal receiver. Dividing, generating discrete Fourier transform terms for characterizing the transport channel; Calculating channel equalization filter coefficients from the reciprocals of the terms of the digital television signal receiver. The method of claim 2, The symbol decoding step A data slicer for supplying a preliminary symbol decoding result in response to an output signal from the cascade connection between the channel equalization filter and the comb filter; And an additional comb filter for suppressing intersymbol interference introduced by the NTSC canceling comb filter in response to the preliminary symbol decoding result. The method of claim 4, wherein The channel equalization filtering step includes a routine for adjusting the adjustable filter coefficients of the adaptive channel equalization filter. Extracting, from the comb filter response signal, a training signal included in a data field synchronization signal intervening in the baseband symbol code signal at the beginning of each data field; Accumulating training signals over an even number of data fields to generate a ghost erase reference signal having accessory ghosts; Calculating a discrete Fourier transform of the ghost cancellation reference signal with associated ghosts; The terms of the discrete Fourier transform of the ghost cancellation reference signal with associated ghosts to the corresponding terms of the discrete Fourier transform of a matched and comb filtered ghost cancellation reference signal supplied from a memory of the digital television signal receiver. Dividing, generating discrete Fourier transform terms for characterizing the transport channel; Calculating channel equalization filter coefficients from the reciprocals of the terms of the digital television signal receiver. The method of claim 1, The channel equalization filtering step is performed by an adaptive channel equalization filter having adjustable filter coefficients, and the comb filtering step is performed by an NTSC cancellation filter cascaded to the rear of the channel equalization filter. A method of operating a receiver for digital television signals subject to co-channel interference. The method of claim 6, The channel equalization filtering step includes a routine for adjusting the adjustable filter coefficients of the adaptive channel equalization filter. Extracting, from the comb filter response signal, a training signal included in a data field synchronization signal intervening in the baseband symbol code signal at the beginning of each data field; Accumulating training signals over an even number of data fields to generate a ghost erase reference signal having accessory ghosts; Calculating a discrete Fourier transform of the ghost cancellation reference signal with associated ghosts; The terms of the discrete Fourier transform of the ghost cancellation reference signal with associated ghosts to the corresponding terms of the discrete Fourier transform of a matched and comb filtered ghost cancellation reference signal supplied from a memory of the digital television signal receiver. Dividing, generating discrete Fourier transform terms for characterizing the transport channel; Calculating channel equalization filter coefficients from the reciprocals of the terms of the digital television signal receiver. The method of claim 6, The symbol decoding step A data slicer for supplying a preliminary symbol decoding result in response to an output signal from the cascade connection between the channel equalization filter and the comb filter; And an additional comb filter for suppressing intersymbol interference introduced by the NTSC canceling comb filter in response to the preliminary symbol decoding result. The method of claim 8, The channel equalization filtering step includes a routine for adjusting the adjustable filter coefficients of the adaptive channel equalization filter. Extracting, from the comb filter response signal, a training signal included in a data field synchronization signal intervening in the baseband symbol code signal at the beginning of each data field; Accumulating training signals over an even number of data fields to generate a ghost erase reference signal having accessory ghosts; Calculating a discrete Fourier transform of the ghost cancellation reference signal with associated ghosts; The terms of the discrete Fourier transform of the ghost cancellation reference signal with associated ghosts to the corresponding terms of the discrete Fourier transform of a matched and comb filtered ghost cancellation reference signal supplied from a memory of the digital television signal receiver. Dividing, generating discrete Fourier transform terms for characterizing the transport channel; Calculating channel equalization filter coefficients from the reciprocals of the terms of the digital television signal receiver. A receiver for digital television signals receiving co-channel interference from analog television signals, In response to the received digital television signal of the digital television signals, supplying a digitized baseband demodulator response signal comprising symbol coded data accompanied by demodulation artifacts of any cochannel interference from a cochannel analog television signal; A demodulator device; A cascade filter device for supplying a cascade filter response signal to said digitized baseband modulator response signal; An adaptive channel equalization filter embedded in the cascade filter device and having adjustable filtering coefficients; A comb filter embedded in the cascade filter device for suppressing demodulation artifacts of interference from a co-channel analog television signal; A symbol decoder for supplying data in response to the cascade filter response signal; An intersymbol interference suppression filter processing the data to compensate for intersymbol interference introduced by the comb filter; An apparatus for extracting a received training signal from a response signal of the cascade filter during a period in which data field synchronization signals are generated in the digital television signals; Computing the discrete Fourier transform terms of the training signal, dividing the terms by corresponding discrete Fourier transform terms of the comb filtered and matched filtered response signal for the training signal stored in the internal memory and not including the ghost. And a computer for generating a discrete Fourier transform characterizing a channel and calculating said adjustable filtering coefficients of said adaptive channel equalization filter to complement said channel characterization. . The method of claim 10, And the channel equalization filter is installed at the front end side of the comb filter in the cascade filter device. 12. The receiver of claim 10, wherein the intersymbol interference suppression filter has an input for receiving decoded data output from the symbol decoder. A receiver for digital television signals receiving co-channel interference from analog television signals, In response to the received digital television signal of the digital television signals, supplying a digitized baseband demodulator response signal comprising symbol coded data accompanied by demodulation artifacts of any cochannel interference from a cochannel analog television signal; A demodulator device; A co-channel interference detector for determining whether there is significant interference from the co-channel analog television signal; A channel equalization that is connected to receive the digitized baseband demodulator response signal as an input signal, has adjustable filtering coefficients, and supplies a channel equalization filter response signal adapted to respond to changes in the adjustable filtering coefficients. A filter; A comb filter connected to receive the channel equalization filter response signal as an input signal and supplying a comb filter response signal in which the demodulation artifacts of interference from the same channel analog television signal are suppressed; A first symbol decoder for supplying data in response to the channel equalization response signal; A second symbol decoder configured to supply data in response to the comb filter response signal; An intersymbol interference suppression filter for processing the data to compensate for intersymbol interference introduced by the comb filter; The first symbol decoder in response to a determination of the co-channel interference detector having a function of selecting data for continuous processing at the receiver and determining that function is such that interference from the co-channel analog television signal is not significant. Compensating for intersymbol interference introduced by the comb filter in response to the determination of the co-channel interference detector where data from is selected for continuous processing at the receiver while determining that the interference from the co-channel analog television signal is significant. A multiplexer for performing data from the second symbol decoder in a state processed by the intersymbol interference suppression filter to select for continuous processing at the receiver; The same channel interference having a function of extracting a training signal received during a period in which data field synchronizing signals are generated from the digital television signals, and determining that the interference from the same channel analog television signal is not significant. In response to the determination of the detector, the received training signal is extracted from the channel equalization filter response signal while the co-filter response signal is determined in response to the determination of the co-channel interference detector, which determines that the interference from the same channel analog television signal is significant. An apparatus for extracting the received training signal; The discrete Fourier transform terms of the received training signal are calculated, and the discrete Fourier transform terms of the received training signal are transferred to internal memory in response to the determination of the co-channel interference detector which determines that the interference from the same channel analog television signal is not significant. The interference from the co-channel analog television signal is significant, while dividing by the matched discrete Fourier transform terms of the matched filtered response signal for the stored and non-ghosted training signal to characterize the channel. In response to the determination of the co-channel interference detector, the discrete Fourier transform terms of the received training signal are stored in an internal memory, and the corresponding discrete Fourier transform of the matched-filtered response signal is stored in an internal memory. Generating a Discrete Fourier Transform that divides the term into the terms to characterize the channel and computes the adjustable filtering coefficients of the adaptive channel equalization filter to complement the channel characterization. Receiver for receiving digital television signals. 14. The receiver of claim 13, wherein the intersymbol interference suppression filter has an input for receiving decoded data from the symbol decoder.

Images