Television – Image signal processing circuitry specific to television – Noise or undesired signal reduction
Reexamination Certificate
2001-07-11
2004-07-27
Lee, Michael H. (Department: 2614)
Television
Image signal processing circuitry specific to television
Noise or undesired signal reduction
C348S611000, C375S342000, C375S343000
Reexamination Certificate
active
06768517
ABSTRACT:
The invention relates to digital television (DTV) signals for over-the-air broadcasting, transmitters for such broadcast DTV signals, and receivers for such broadcast DTV signals, which broadcast DTV signals include novel echo-cancellation reference (ECR) signal components or initializing the parameters of adaptive filters used in the DTV receivers for channel-equalization and echo-cancellation.
BACKGROUND OF THE INVENTION
The Advanced Television Systems Committee (ATSC) published a Digital Television Standard in 1995 as Document A/53, hereinafter referred to simply as “A/53” for sake of brevity. Annex D of A/53 titled “RF/Transmission Systems Characteristics” is particularly incorporated by reference into this specification. Annex D specifies that the data frame shall be composed of two data fields, each data field composed of 313 data segments, and each data segment composed of 832 symbols. Annex D specifies that each data segment shall begin with a 4-symbol data-segment-synchronization (DSS) sequence. Annex D specifies that the initial data segment of each data field shall contain a data-field synchronization (DFS) signal following the 4-symbol DSS sequence therein. The fifth through 515
th
symbols in each A/53 DFS signal are a specified PN511 sequence—that is a pseudo-random noise sequence composed of 511 symbols capable of being rendered as +5 or −5 values. The 516
th
through 704
th
symbols in each A/53 DFS signal are a triple-PN63 sequence composed of a total of 189 symbols capable of being rendered as +5 or −5 values. The middle PN63 sequence is inverted in polarity every other data field. The 705
th
through 728
th
symbols in each A/53 DFS signal contain a VSB mode code specifying the nature of the vestigial-sideband (VSB) signal being transmitted. The remaining 104 symbols in the each A/53 DFS signal are reserved, with the last twelve of these symbols being a precode signal that repeats the last twelve symbols of the data in the last data segment of the previous data field. A/53 specifies such precode signal to implement trellis coding and decoding procedures being able to resume in the second data segment of each field proceeding from where those procedures left off processing the data in the preceding data field.
The broadcast TV signal to which the receiver synchronizes its operations is called the principal signal, and the principal signal is usually the direct signal received over the shortest transmission path. Thus, the multipath signals received over other paths are usually delayed with respect to the principal signal and appear as lagging ghost signals. It is possible however, that the direct or shortest path signal is not the signal to which the receiver synchronizes. When the receiver synchronizes its operations to a (longer path) signal that is delayed respective to the direct signal, there will be a leading multipath signal caused by the direct signal, or there will a plurality of leading multipath signals caused by the direct signal and other reflected signals of lesser delay than the reflected signal to which the receiver synchronizes. In the analog TV art multipath signals are referred to as “ghosts”, but in the DTV art multipath signals are customarily referred to as “echoes”. The multipath signals that lead the principal signal are referred to as “pre-echoes”, and the multipath signals that lag the principal signal are referred to as “post-echoes”. The echoes vary in number, amplitude and delay time from location to location and from channel to channel at a given location. Post-echoes with significant energy have been reported as being delayed from the reference signal by as many as sixty microseconds. Pre-echoes with significant energy have been reported leading the reference signal by as many as thirty microseconds. This 90-microsecond or so possible range of echoes of is appreciably more extensive than was generally supposed before spring 2000.
The transmission of the digital television (DTV) signal to the receiver is considered to be through a transmission channel that has the characteristics of a sampled-data time-domain filter that provides weighted summation of variously delayed responses to the transmitted signal. In the DTV signal receiver the received signal is passed through equalization and echo-cancellation filtering that compensates at least partially for the time-domain filtering effects that originate in the transmission channel. This equalization and echo-cancellation filtering is customarily sampled-data filtering performed in the digital domain. Time-domain filtering effects differ for the channels through which broadcast digital television signals are received from various transmitters. Furthermore, time-domain filtering effects change over time for the broadcast digital television signals received from each particular transmitter. Changes referred to as “dynamic multipath” are introduced while receiving from a single transmitter when the lengths of reflective transmission paths change, owing to the reflections being from moving objects. Accordingly adaptive filter procedures are required for adjusting the weighting coefficients of the sampled-data filtering that provides echo-cancellation and equalization.
Determination of the weighting coefficients of the sampled-data filtering that provides equalization and echo-cancellation is customarily attempted using a method of one of two general types. A method of the first general type relies on analysis of the effects of multipath just on an echo-cancellation reference (ECR) signal included in the transmitted signal specifically to facilitate such analysis. A method of the second general type relies on analysis of the effects of multipath on all portions of the transmitted signal. While the PN511 and triple-PN63 sequences in the initial data segments of the data fields in the ATSC standard DTV signal were originally proposed for use as ECR signals, the VSB receiver performance in actual field environments has demonstrated that these sequences are inadequate ECR signals, considered separately or in combination. So, most DTV manufacturers have used decision-feedback methods that rely on analysis of the effects of multipath on all portions of the transmitted signal for adapting the weighting coefficients of the sampled-data filtering. Decision-feedback methods that utilize least-mean-squares (LMS) method or block LMS method can be implemented in an integrated circuit of reasonable size. These decision-feedback methods provide for tracking dynamic multipath conditions reasonably well after the equalization and echo-cancellation filtering has initially been converged to substantially optimal response, providing that the sampling rate through the filtering is appreciably higher than symbol rate and providing that the rate of change of the dynamic multipath does not exceed the slewing rate of the decision-feedback loop,
However, these decision-feedback methods tend to be unacceptably slow in converging the equalization and echo-cancellation filtering to nearly optimal response when initially receiving a DTV signal that has bad multipath distortion. Bad multipath distortion conditions include cases where echoes of substantial energy lead or lag the principal received signal by more than ten or twenty microseconds, cases where there is an ensemble of many echoes with differing timings relative to the principal received signal, cases where multipath distortion changes rapidly, and cases where it is difficult to distinguish principal received signal from echo(es) because of similarity in energy level.
Worse yet, convergence is too slow when tracking of dynamic multipath conditions must be regained after the slewing rate of the decision-feedback loop has not been fast enough to keep up with rapid change in the multipath conditions. Data dependent equalization and echo cancellation methods that provide faster convergence than LMS or block-LMS decision-feedback methods are known, but there is difficulty in implementing them in an integrated circuit of reasonable size.
Acco
Limberg Allen Le Roy
McDonald James Douglas
Patel Chandrakant Bhailalbhai
Lee Michael H.
Tran Trang U.
LandOfFree
Repetitive-PN1023-sequence echo-cancellation reference... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Repetitive-PN1023-sequence echo-cancellation reference..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Repetitive-PN1023-sequence echo-cancellation reference... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3201589