Error detection/correction and fault detection/recovery – Pulse or data error handling – Transmission facility testing
Reexamination Certificate
1999-11-04
2003-02-25
Tu, Christine T. (Department: 2784)
Error detection/correction and fault detection/recovery
Pulse or data error handling
Transmission facility testing
C714S704000
Reexamination Certificate
active
06526532
ABSTRACT:
FIELD OF INVENTION
The present invention relates to a channel error correction apparatus and method. In particular, the present invention relates to an a posteriori channel error correction apparatus using subband synthesis.
TECHNICAL BACKGROUND OF THE INVENTION
FIG. 1
illustrates a usual subband transmission scheme. Here, an input signal X is supplied to an analysis section
100
wherein the input signal X is split into two different subbands, e.g., a low-pass subband and a high-pass subband. This is achieved with a first low-pass filter
102
having an impulse response h
0
and a second high-pass filter
104
having an impulse response h
1
. The outputs X
0d
and X
1d
of the different subband filters
102
and
104
are supplied to related downsampling units
106
and
108
, respectively, wherein the sampling rate of these signals is lowered. The downsampled signals X
0f
and X
1f
are then transmitted via a transmission channel
110
.
As also shown in
FIG. 1
, at the output side of the transmission channel
110
the received signals X
0
and X
1
are supplied to a synthesis section
116
, where they are first upsampled by the respective up-sampling unit
112
and
114
, respectively. Furthermore, the synthesis section
116
comprises a synthesis filter (
118
and
120
) for each subband to which the up-sampled signals are supplied. Finally, the output signals of these synthesis filters
118
and
120
are added in an adding unit
122
in order to derive the final output signal X′. Being an estimate for the original input signal X, the synthesized signal X′ is supposed to be as close to X as possible.
The typical application for such a subband-oriented transmission of digital data is the area of image and video coding/decoding wherein most applications are based on a two-band scheme in which the original image signal is split into two subbands. The downsampled subband samples are then quantized, compressed, and transmitted. At the synthesis section
116
they are reconstructed by upsampling, filtering and adding.
FIG. 2
illustrates the result of such a two band subband transmission for an input signal X shown in FIG.
2
(
a
). Just as in
FIG. 1
the sender-side down-sampled subband signals are denoted X
0f
and X
1f
in FIGS.
2
(
b
) and (
c
). Also, FIGS.
2
(
b
) and
2
(
c
) show the subband signals X
0
and X
1
after transmission via the transmission channel. As shown, an error occurs between sample indices
2
and
4
, i.e. at sample index
3
, in signal X
0
and at sample index
6
in signal X
1
.
FIG.
2
(
d
) shows the output signal achieved through processing these subband signals X
0
and X
1
in the receiver. Typically, there exist errors in the output signal X′ due to transmission errors. For this reason channel or error control coding is used which, however, consumes a certain amount of available channel capacity so that the bandwidth for the actual transmission must be reduced correspondingly. Due to the reduced transmission bandwidth the quantization of the samples must be made coarser to achieve a better compression of the samples of the subband signals X
0f
and X
1f
. This in turn degrades the quality of the transmitted signals.
SUMMARY OF INVENTION
In view of the above, the object of the invention is to provide an approach to correction of residual errors in subband-based transmission schemes.
According to the present invention, this object is achieved through a channel error correction apparatus comprising a first synthesis means synthesizing a preliminary output signal from subband signals received via a transmission channel through subband synthesis, a subband correcting means with a subband reanalysis means analyzing the preliminary output signal through subband filtering without downsampling, error estimation means estimating errors in subband reanalysis output signals, using subband specific error patterns, and an error compensation means correcting the subband signals using error estimation results, a second synthesis means synthesizing a final output signal from output signals of the error compensation means through subband synthesis.
An important advantage of this channel error correction apparatus is that it allows to minimize channel coding or even to drop it completely and to fully use the available bandwidth of the transmission channel for transmission of more source coded data. The reason for this is the application of an a posteriori channel error correction. In particular, using no downsampled subband signals the error positions and amplitudes in the subband signals may be estimated. This allows for a correction of these signals and a resynthesis of a nearly error-free signal.
Further, the channel error correction apparatus may be extended to multiple dimensions, e.g., to image transmission. The present invention does not require an image restoration or post-processing method and blurring—inherent to most of these methods—is not present.
Since the channel error correction apparatus is used at the receiver side only it relocates the error correction to the receiver solely. This is of particular advantage in applications that have to be enhanced without altering the sender which may already be well defined by, e.g., transmission standards.
According to a preferred embodiment, the subband correcting means may further comprise a subband interpolation means receiving the subband signals and generating related up-sampled subband interpolation signals supplied to the error estimation means. This interpolation enables an improved calculation precision of error amplitudes during error correction.
According to yet another preferred embodiment, the error estimation means comprises at least two error pattern correlation means, at least two error position estimating means and at least two error amplitude estimation means. Here, the error pattern correlation means correlates appropriate combinations of subband reanalysis output signals and upsampled subband interpolation signals with subband specific error patterns for the identification of error positions. In particular, subband specific error patterns are used to identify the error positions.
According to yet another preferred embodiment pairwise convolutions of analysis filter impulse responses and synthesis filter impulse responses are used as such subband specific error patterns during the error estimation. Therefore, according to the present invention particular use is made of the fact that through resynthesizing the received subband signals without downsampling, error indication information is available which is not the case for standard subband transmission schemes. In particular, through carrying out a first synthesis to generate a preliminary signal and then carrying out an additional reanalysis of this preliminary signal the reanalyzed subband signals have patterns comprised therein that allow to identify the error positions.
According to further preferred embodiments of the present invention, in each subband a distinction is made between the inband characteristic error and cross-band characteristic errors. The former kind of error stems from an error occuring during transmission of the related subband while the latter kind stems from errors occuring during the transmission of neighbouring subbands that have an impact on the considered subband due to aliasing or cross redundancy. To identify inband characteristic errors, resynthesized subband signals should be considered. To detect cross-band characteristic errors, differences between resynthesized subband signals and up-sampled interpolated subband signals should be considered. The difference signals allow to achieve a normalization to the resynthesized subband signals and thus facilitate the detection of cross-band characteristic errors having only very small amplitudes.
According to yet another preferred embodiment of the present invention, two subbands are used and the error amplitude estimation means in the error estimation means use different approaches for the estimation of the error amplitude in a low-pass sub
Cinkler Kalman
Kammeyer Karl Dirk
Jenkens & Gilchrist P.C.
Telefonaktiebolaget LM Ericsson (publ)
Tu Christine T.
LandOfFree
Channel error correction apparatus and method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Channel error correction apparatus and method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Channel error correction apparatus and method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3151691