Coded data generation or conversion – Digital code to digital code converters – To or from minimum d.c. level codes
Reexamination Certificate
1999-12-07
2001-07-03
Young, Brian (Department: 2819)
Coded data generation or conversion
Digital code to digital code converters
To or from minimum d.c. level codes
Reexamination Certificate
active
06255967
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to data communication over subscriber lines and, more particularly, to high speed data communication over such lines. A frame-based spectral shaping method and apparatus may be utilized to suppress very low frequency and DC energy in the communicated data.
A data communication system includes an encoder that is digitally connected to a digital portion of the general switched telephone network. A data source provides an input to the encoder. A subscriber is located at the opposite end of the communication system. The subscriber is typically connected to the general switched telephone network by a twisted pair of wires, commonly known as an analog loop.
In such a system, information may be communicated from the data source to the subscriber as follows. Information from the data source is converted into a series of digital codewords by the encoder. The digital codewords pass in digital form from the encoder through the digital portion of the telephone network. At an interface between the digital portion of the telephone network and the analog loop, the series of codewords is converted into an analog voltage waveform by a digital-to-analog converter. A decoder located at the subscriber end of the analog loop receives a distorted version of the analog voltage waveform and reconstructs the series of codewords from the waveform. The information from the data source may then be extracted from the reconstructed series of codewords.
For communication in the opposite direction, information at the subscriber end may be modulated and transmitted in analog form over the analog loop. At the interface between the analog loop and the digital portion of the telephone network, the analog signal is converted into a series of codewords by an analog-to-digital converter. The codewords are transmitted from the interface to the data source, where they are demodulated and the information is recovered.
Various standards have been adopted throughout the world for the analog-to-digital and digital-to-analog conversions performed by the telephone network. The United States, for example, uses a conversion scheme in which the analog-to-digital converter in the interface samples the analog signals at the rate of 8000 samples per second and maps the samples into one of 255 possible distinct codewords. The 255 codewords correspond to quantization levels defined by a non-linear mapping rule called the &mgr;-law companding rule, which is the Pulse Code Modulation (“PCM”) voice coding and companding standard in North America and Japan. In Europe, the A-law companding rule is used. The codeword chosen for each analog sample corresponds to the quantization level that is closest to the voltage of the analog sample. The digital-to-analog converter in the interface performs the inverse of this mapping, i.e. each codeword utilized by the digital portion of the telephone network is associated by the digital-to-analog converter with an analog voltage.
The codewords utilized by the digital portion of the general switched telephone network are typically eight bit codewords.
FIG. 1
shows a bit allocation map for a &mgr;-law codeword. In the eight bit codeword, the most significant bit, b
7
, is a sign bit. The next three bits, b
6
through b
4
, identify one of eight segments in the &mgr;-law quantization characteristic. The last four bits, b
3
through b
hd
0
, identify one of sixteen steps within that segment. The bit locations b
6
through b
0
may be referred to herein as the magnitude field of the eight bit codeword.
The general switched telephone network utilizes DC signals on analog loops to power telephone handsets and to signal when the handset, or other customer equipment such as a modem, goes off hook. The off hook signal indicates that the customer equipment is connected to the analog loop. Accordingly, it is desirable to design customer equipment that may be connected to analog loops, such as modems, answering machines and the like, to separate telephony signalling from DC network signalling. In modems, for example, it is common to use isolation transformers and capaitors to block low frequency signals.
With respect to telephony signalling, it has been observed that very low frequency signals suffer greater harmonic distortion in analog loop circuits than do higher frequencies. In addition, very low frequency components in a telephony signal may cause very long echo impulse response, thereby increasing the complexity of echo cancellation. It is therefore desirable to reduce DC and very low frequency components in telephony signals.
A device is known that compensates for a DC component in transmitted data by altering a stream of, for example, &mgr;-law codewords. In U.S. patent application Ser. No. 08/352,651, which is incorporated herein by reference, Townshend shows a DC eliminator for use in an encoder. The encoder converts a data stream into a stream of codes, which may be transmitted over a telephone network to a subscriber. The subscriber may be connected by an analog loop to a digital portion of the telephone network. The encoder has a digital connection to the digital portion of the telephone network. At an interface between the analog loop and the digital portion of the telephone network, a telephone network digital-to-analog converter converts the stream of codes into an analog voltage waveform. The DC eliminator in the encoder may function to alter the stream of codes, as described below, so that the analog voltage waveforn does not have a DC component.
A functional block diagram of the Townshend DC eliminator
50
is shown in FIG.
2
. In the DC eliminator shown by Townshend, the code stream
52
is converted by converter
54
to linear values, which are accumulated and negated by a summer
56
and a unit delay
58
to form a DC offset signal
60
. The DC offset signal is applied to a converter
62
that produces a DC restoration code
64
. A two-input selector
66
then chooses an output code from one of the code stream
52
and the DC restoration code
64
. In an operational mode of the DC eliminator shown by Townshend, the two-input selector
66
outputs seven sequential values from the code stream
52
followed by one value of the DC restoration code
64
.
A disadvantage of the DC eliminator shown by Townshend is the cost of the DC elimination in terms of data rate. For example, if each code is the same bit length and one in every eight transmitted codes is a DC restoration code, which carries no data, then the data rate for such a system may not exceed seven-eighths of its potential value. It is desirable to minimize the cost of DC elimination on the data rate of the system.
In accordance with an alternative known method for DC suppression, the sign bit, such as the bit b
7
shown in
FIG. 1
, of every nth codeword is commandeered to suppress the DC content of the analog voltage waveform. A stream of codewords may be converted to a series of linear values, which are accumulated by a summer. The encoder may then modify the stream by inserting a sign bit into every nth codeword, where the value of the sign bit (positive or negative) is selected to oppose the sign of the value accumulated by the summer. The value accumulated by the summer may then be reduced (if the value accumulated by the summer is positive) or increased (if the value accumulated by the summer is negative) by the value of the nth codeword. In terms of retrieving data at the decoder, the decoder simply ignores the sign bit of every nth codeword.
FIGS. 3A and 3B
show a simulated spectral output of an encoder that commandeers the sign bit of every sixth codeword for purposes of DC suppression. In
FIG. 3A
, the entire frequency band from 0 Hz to 4000 Hz is shown.
FIG. 3B
provides an expanded view of the 0 Hz to 200 Hz range from FIG.
3
A.
A disadvantage of this method is that the magnitude of the DC suppression codes, i.e. the magnitude of every nth codeword, depends upon the random value of the data bits in the magnitude field of every nth codeword. If the magnitude
Lery Scott A.
Norrell Andrew L.
Parizhsky Vladimir G.
Waldron Mark A.
3Com Corporation
McDonnell & Boehnen Hulbert & Berghoff
Young Brian
LandOfFree
Frame-based spectral shaping method and apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Frame-based spectral shaping method and apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Frame-based spectral shaping method and apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2552672