CN1231789A - Device, system and method for spectrally shaping transmitted data signals - Google Patents
Device, system and method for spectrally shaping transmitted data signals Download PDFInfo
- Publication number
- CN1231789A CN1231789A CN 97198256 CN97198256A CN1231789A CN 1231789 A CN1231789 A CN 1231789A CN 97198256 CN97198256 CN 97198256 CN 97198256 A CN97198256 A CN 97198256A CN 1231789 A CN1231789 A CN 1231789A
- Authority
- CN
- China
- Prior art keywords
- level
- sent
- group
- digital network
- equivalence class
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/38—Synchronous or start-stop systems, e.g. for Baudot code
- H04L25/40—Transmitting circuits; Receiving circuits
- H04L25/49—Transmitting circuits; Receiving circuits using code conversion at the transmitter; using predistortion; using insertion of idle bits for obtaining a desired frequency spectrum; using three or more amplitude levels ; Baseband coding techniques specific to data transmission systems
- H04L25/4917—Transmitting circuits; Receiving circuits using code conversion at the transmitter; using predistortion; using insertion of idle bits for obtaining a desired frequency spectrum; using three or more amplitude levels ; Baseband coding techniques specific to data transmission systems using multilevel codes
- H04L25/4927—Transmitting circuits; Receiving circuits using code conversion at the transmitter; using predistortion; using insertion of idle bits for obtaining a desired frequency spectrum; using three or more amplitude levels ; Baseband coding techniques specific to data transmission systems using multilevel codes using levels matched to the quantisation levels of the channel
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Dc Digital Transmission (AREA)
- Small-Scale Networks (AREA)
Abstract
The transmitter (56) for transmitting a sequence of octets over a digital network which digital network converts the octets into a sequence of levels, the transmitter including: mapping device (74) outputting for each group of data bits an equivalence class of one or more levels; a level selector (76), for selecting the level in the equivalence class to represent the group of data bits to be transmitted; a filter device (78), which receives at its input previously selected levels and provides its output to the level selector, wherein the level selector (76) selects the level to be transmitted based on the output of the filter device (78); and an octet converter (80), which receives the selected level at its input and transmits over the digital network an octet corresponding to the selected level to produce the analog signal having the predefined spectral shape when the transmitted octets are converted to levels in the digital network.
Description
The present invention is the continuation application of the u.s. patent application serial number 08/730,434 of application on October 15th, 1996, and the full text of above-mentioned patent application is by income this paper, as a reference.
The present invention relates to the high-speed data communication on a PSTN (PSTN), particularly relate on frequency spectrum the System and method for that forms all analog data signals, above-mentioned all analog data signals are that a telephone central office from this PSTN mails to the long-distance user's by an analogue loop.
This PSTN comprises a digital backbone network and all simulation local loops that all users are connected to this backbone network.In typical call once, the analog signal of being sent by this local user is digitized in this local central office, and be converted into the bit stream of one group of 64kbit/s, the latter is transmitted on this numeral backbone network, and be inverted in this office of remote center and be changed to analog signal, so that be sent to this long-distance user by this long-range local loop.By means of digital information being modulated on one group of analog signal that is used to transmit, just can be implemented in the dialing modem communication on this PSTN.Digital to analogy transfer process in the entrance that enters this numeral backbone network has been introduced quantizing noise, and this just is limited in message transmission rate about the 30kbit/s.
ISDN is a kind of circuit switching public network, and it allows to use the various circuit of 64kbit/s to carry out the communication of terminal-to-terminal service.When an isdn user communicates with a PSTN user, the analog information that produces in this ISDN one side is converted into the bit stream of one group of 64kbit/s in an isdn terminal adapter, this central office will originate from one group of analog signal digital of PSTN one side in an identical manner.Except ISDN, can also use other various digital access stencils to arrive a PSTN end points.For example, remove to visit this PSTN at medium to the common T1 line that uses between each major company.Do not consider this numeral access medium, this traditional method that this modem signal is converted to one group of 64kbit/s bit stream is limited in accessible modem rate about the 30kbit/s.
Therefore need set up a kind of system and method, so that avoiding using under the condition of special and infrastructure costliness, realize communication having with this telephone network between all users that numeral connects and all users that are connected to this network by an analogue loop with higher transmission rate.
Fig. 1 is the simplified block diagram of the central office of typical telephone operator an of family;
Fig. 2 is the spectral shape of the low pass filter of the spectrogram of all yk signals of restraining converter,linear output of this m from Fig. 1 and Fig. 1;
Fig. 3 is the figure of the part of two frequency spectrums that respectively have a zero point on flip-flop, and one of them frequency spectrum drops to zero very precipitously on flip-flop, and another frequency spectrum descends relatively gentlyer;
Fig. 4 is the graphic representation of the part of a typical m rule constellation;
Fig. 5 is the block diagram that a modem data on this telephone system connects, and this telephone system comprises a transmitter, and it forms all signals on frequency spectrum according to the present invention.
Fig. 6 is a block diagram of encoder shown in Figure 6, and described all analog signals that this encoder ad hoc is used for an analogue loop on this PSTN generate direct current zero point; And
Fig. 7 is a block diagram of encoder shown in Figure 6, and this encoder usually can be used to revise as required the frequency spectrum that outputs to all signals of this terminal use from this analogue loop.
When a user once directly connects with a digital network foundation, a kind of technology allows to transmit with the speed that is higher than 30kbit/s significantly, for example, via ISDN or the T1 line that has been developed and in following co-pending patent application, has been illustrated transmit, these patent applications comprise the following patent application of transferring the possession of in assignee of the present invention: the u.s. patent application serial number of on October 15th, 1996 application: unsettled case (agent make a summary catalog number (Cat.No.) No.CX096045) is entitled as " being adjudicated the device that the self adaptation self-noise is offset that is used for that the timing of domination recovers; system and method "; The u.s. patent application serial number of on October 15th, 1996 application: unsettled case (agent make a summary catalog number (Cat.No.) No.CX096046) is entitled as " being used to handle all baseband signals to overcome intersymbol interference and nonlinear system, devices and methods therefor in a communication system "; And the u.s. patent application serial number of on October 15th, 1996 application: No.08/730,433 (agents make a summary catalog number (Cat.No.) No.CX096047) are entitled as " detecting, characterize and alleviate system, the devices and methods therefor of prejudgementing character distortion in a communication network ".These co-pending patent applications form is in full taken in this paper, as a reference.Adopt this technology, use a channel encoder, according to residing region in the world, with coding digital information at random is all bytes of μ rule or A rule, subsequently all bytes directly are mapped to the central office that is arranged in this long-distance user digital to analogy (D/A) transducer various level (below, except as otherwise noted, we will be assumed to the μ rule.It is simple and clear extending to the A rule).This mapping can be used whole or any subclass of 255 kinds of level of this D/A converter, and this submits to the restriction in the adjusting of average power.
Owing in this digital network, come transmission information with the form of all bytes, this coded data at first be mapped as m and restrain all bytes, so that the speed with per second 8000 bytes is transmitted, in long-distance user's central office, these bytes are converted into required various amplitude leyels in this D/A converter then.Then, the sequence of all level of resulting 8kHz is sent to this long-distance user by a low pass filter (LPF) and via this analogue loop.The output of this D/A converter can be regarded as a pulse train, and wherein each pulse all has an amplitude corresponding to a kind of level of this D/A.At the remote terminal place, a receiving modem is sent out by at first detecting which D/A level, then these level is carried out reverse mapping to obtain that an estimation of this original digital information is recovered this raw information.Can realize the transfer of data of 64kbps on this technical know-how; But because the various restrictions and the noise of real system, acquisition 48kbps is more real to the transmission rate of 56kbps.Such transmission rate is once great improvement for the theoretic limit of about 32kbps that people were once once imagining.
When this information that is sent out was random information, the frequency spectrum that after D/A conversion a spectrum analysis of all signals has been disclosed this sequence of being exported by this D/A converter was smooth basically.Therefore, when this sequence was passed through this low pass filter in this central office, the frequency spectrum of this signal was taked the shape of the frequency spectrum of this LPF.Regrettably, this frequency spectrum has a significant energy values near on the frequency of direct current (f=0), and the transformer that it can order about in this system enters saturation condition, and the signal that is being sent out is introduced undesirable nonlinear distortion.In some applications, for example, in high-rate data transmission techniques, this distortion can not be allowed, therefore need be eliminated.
More generally, need a kind of like this scheme, it can be adjusted from the shape of the frequency spectrum of this signal of this D/A converter transmission.In addition, also need such frequency spectrum formation is combined with chnnel coding.
Under current scientific and technological level, existing multiple known method can be used for numerical data is mapped in the sequence of the spectral shape with a kind of regulation goes.For example, the United States Patent (USP) that is entitled as " partial response channels signaling system " has just been described a kind of like this method the 5th, 040, No. 191.Yet these methods suppose that all signals that are sent out all are selected from a signal constellation (in digital modulation), and the each point of this signal constellation (in digital modulation) all is positioned on the translation of a grid.Because all level on a D/A converter of a central office are disposed unevenly, thereby can not be on the translation of a grid, so these methods can not directly be applied to present patent application.And, adopting the present invention, this cataloged procedure is finished in this modulator-demodulator the inside, is mapped to various level and then takes place in this network internal.
Fig. 1 and 2 explanation is passed through an analogue loop in all signals of a long-distance user's modulator-demodulator transmission, in the existence near energy on the frequency of direct current.Figure 1 illustrates the part of a typical telephone central office 10 in a PSTN, this central office 10 (sends the modulator-demodulator of a side at its input 12 from a modulator-demodulator that directly is connected to the numerical portion of this telephone system, not shown) all m rule byte of being sent, that is narrated in for example superincumbent common unsettled all patent applications is a kind of, and it directly is all bytes that are used to transmit with this digital data coding.These bytes are also referred to as a m and restrain converter,linear 14 by a D/A converter, are converted to the sequence of all voltage level yk, and wherein each level all is a kind of in 255 m rule level.All level are output to a low pass filter 18 by circuit 16, and the latter exports one group of analog signal s (t) through filtering by analogue loop 20 to this long-distance user's receiver, and it is an analog representation of all level.This analog signal is received the modulator-demodulator demodulation sign indicating number of a side by this, this modulator-demodulator output set of number bit stream, and it is an estimation of the initial data that is sent out.
In Fig. 2, appear at sequence on the circuit 16, that restrain all level of converter,linear 14 from m and have a flat frequency response 22.As put shown in 26, the spectral shape 24 of low pass filter 18 has a significant energy values near on the frequency of direct current (f=0).Because this sequences y k has a flat frequency response, so the frequency spectrum by this signal s (t) of filter 18 output has the spectral shape 24 that is same as this filter 18, thereby this signal s (t) is also having a significant energy values near on the frequency of direct current.As mentioned above, trending towards making all transformers in this system to enter saturation condition near this energy on the frequency of direct current, this just produces unwanted nonlinear distortion in being addressed to this signal s (t) of this recipient's modulator-demodulator.
In some applications, must reduce this distortion.By reducing the signal energy on the close frequency of direct current in the signal that this is sent out, just can realize this point.Such direct current 0. 28 has been described in Fig. 3.From technology development situation as can be known, generate this spectral null in order to be sent out at this in signal, the mobile digital of all level yk that are sent out and (RDS, that is, the algebraical sum of all level that all before had been sent out) must be retained as and approach zero.Near this direct current 0. 28 spectral shape can change to a frequency spectrum 32 that descends in flip-flop from a mild relatively ramped shaped frequency spectrum 30 precipitously.The steepness at this zero point depends on the controlled strict degree of this RDS.
The present invention correspondingly encodes to being sent to this numerical data that m restrains all bytes near a kind of mode that makes this RDS value remain on zero, so that on flip-flop, generate required spectral null, thereby reduce the nonlinear distortion that produces because of transformer is saturated.
For the method that generates a spectral null is described, we consider to send 6 such examples of bit with each symbol yk.Concerning the professional person, clearly, the present invention can be used to send the signal of the bit number of every symbol for any other numerical value, perhaps is used to the bit number situation different with the difference of symbol of every symbol of being sent out.In a system that does not have a spectral null, people at first select a subclass that contains 64 kinds of level from available 255 kinds of m rule level, make to keep a minimum range d between all level
MinThese 64 kinds of level are symmetrical on following meaning,, concerning each positive level, the negative level of an equal size are arranged all that is.For example, can just in time below the limit-12dBm0 of this adjusting, be that an average energy acquisition numerical value is 32 d
Min
The part that figure 4 illustrates whole 255 kinds of m rule level 34 (128 kinds positive with 127 kinds negative) is represented.These level are followed a kind of logarithmic parabola, that is, 64 kinds the level of close initial point be spaced apart between-63 and 63 with 2 equably.The section of next positive and negative starts from+/-66, and each section all contains 16 points, between points be spaced apart 4.This scale continues according to per 16 sections, and interval between points takes 2
nForm, the interval in front of each section between section of face then is 0.75*2
nLast all sections from+/-2112 extend to+/-4032, be spaced apart 128.This set 35 is set of 64 kinds of level selecting from 255 kinds of level, in order to represent each combination of 6 bits, that is, and 2
6=64.
In transmitter one side, all bits of arrival are that one group mode is collected according to per 6, are mapped as all m rule bytes of the required level of representative then.In this inside, central office, all m rule bytes are converted into all level, subsequently resulting all level are sent.In receiver one side.Equalizer is to because the distortion that this low pass filter and this local loop cause compensates, and a judgment device is estimated this transmission level by the level of selecting the most close this acceptance point place subsequently.
In order in above-mentioned example, to obtain (required) spectral shape, also used additional all level, but the minimum range between all level still is retained as 32.For example, consider to use the situation of 92 kinds of level.At first, these 92 kinds of level are divided into all equivalence classes.The distinct methods that produces these equivalence classes has a variety of.Wherein a kind of useful especially method is described below: we come all level of mark with integer 0 to 91, for example label 0 are distributed to minimum (the most negative) level, and label 1 is distributed to time minimum level, or the like.Then, we organize together and define 64 " equivalence classes " by its label being differed exactly all level of 64.A kind of like this tissue causes 36 equivalence classes, and a kind of corresponding in innermost all level of 36 kinds of minimum radiuses of a kind of level wherein only arranged, and also causes having 28 equivalence classes that label differs two level of 64.Can also use other the whole bag of tricks that are used to produce all equivalence classes.So, just can represent each possible combination of 6 bits to be sent with an equivalence class.
For example, bit combination 000000 may be corresponding to the 1st equivalence class, and it contains two kinds of level, and each level wherein all uses a different byte to represent.Be noted that the whole dynamic range that does not need to go to use this D/A converter.As long as use the level more than 64 kinds, the present invention just can work on the level of arbitrary number.Certainly, the number of levels that is used is many more, and is can getable required spectral shape also just good more.Our experiment shows, in order to produce a direct current zero point with quite precipitous trap point, only need consider that additional levels seldom gets final product.
In the above example, owing to represent each combination of 6 information bits with an equivalence class, and the level that more than one are usually arranged in an equivalence class usually, so, before a byte representing this level was sent out, these information bits must be mapped on the level in all level in a chosen equivalence class.Below with reference to Fig. 5-7 this function is described.
The receiver of face described in common unsettled all applications compared in fact and do not changed in front of the operation of receiver 66.Unique difference is exactly that this receiver need consider that now a bigger set and this reverse mapping of possible all level need determine equivalence class.As described in common unsettled all applications, 70 pairs of nonlinear distortions owing to low pass filter 64 and 65 introducings of this local loop of equalizer compensate.For example, when using a linear equalizer, the output of this equalizer can be expressed as follows:
rk=vk+nk (1)
In the formula, nk is the summation that the noise that appears at this equalizer output adds distortion.Then, decoder 72 selects all level yk of the most close rk as judgement, determines its equivalence class, then recovers this 6 information bits by reverse mapping once.
If this equalizer comprises a maximum likelihood degree sequential estimator (for example, viterbi equalizer), then this received signal can be expressed as following form:
rk=Syk-jfj+nk (2)
At this moment, this decoder use a Viterbi decoder to select immediate sequence { yk}, this decoder is determined its equivalence class for each estimative symbol yk, and finds out this 6 information bits via reverse mapping once subsequently.
This judgement is made according to following step: at first, encoder 56 feeds back to functional block 78 by the output with electrical level selector 76, keeps following the tracks of to the mobile digital of all level yk of having sent with (RDS).Functional block 78 is calculated the RDS of this weighting, z from the level yk that has before sent
kThe RDS of=-(1-b), in the formula, b is a weighted factor, 0≤b＜1.Because D/A converter is non-linear, in encoder 56, may not know the accurate numerical value of all level yk; Yet this should not produce great influence.Might determine this error and this information is returned to encoder 56, so that make these calculating more accurate.
Provide after the 6 bit groups to be sent, electrical level selector 76 is selected level near the RDS of this weighting as level yk from this equivalence class { y (i, 1), y (i, 2) }.As can be seen, when this RDS is timing, z
kWill be for negative, vice versa.This makes this encoder select a level yk from each equivalence class, makes that it is compared with other level in this equivalence class when its numerical value is added to this RDS, can make this RDS more approach zero.After selected this level yk, represent the byte of this level yk just to be determined, and on this digital network, transmit by this byte converter 80.From a look-up table, can obtain the numerical value of this byte that has sent.
This variable b is a weighted factor, in order to control compromise between the steepness of this spectral null and this average energy that is sent out signal.Our analysis shows, when the number of level was in a ratio of enough greatly with the number of all equivalence classes, this sequences y k will have can be with filter response h (D)=(1-D)/(1-b D)) frequency spectrum that is similar to.Clearly, when b=0, we find h (D)=(1-D), the well-known I class partial response that has a sinusoidal spectral shape and have a zero point on flip-flop that Here it is.On the other hand, along with b approaches 1, except having on the flip-flop the very precipitous spectral null, this frequency spectrum becomes smooth on the major part of this frequency band.As can be seen, under the condition of b=0, the average energy of yk will be the twice under the situation of a smooth spectral shape.Yet along with b approaches 1, the increase of average energy will be tending towards disappearing.In some applications, may wish that the ratio of the number of number by all level and all equivalence classes is weighed, thus the expansion that keeps this constellation.The present invention can the constellation of number of levels be used with having arbitrarily, also can be used with the number of less all equivalence classes arbitrarily, and this is conspicuous to the professional person.
The present invention can be applied even more extensively as required and carry out shape in all analog signals of on frequency spectrum the m in this central office being restrained converter,linear output and modify.Example described above is to use the present invention to go to reduce a near special case of the energy of this transmission signal flip-flop, but employed principle of the present invention can be by vague generalization in that example, make it to modify the spectral shape of various signals by number of ways, for example, various signals are carried out preequalization.
A kind of general pattern of encoder of the present invention, encoder 56 is shown in Fig. 7.Unique difference that this general situation is caught up with the special case of a spectral null of stating is how this sequence or this frequency spectrum function zk produce.Make h (D) be a first term be 1 (progression, h (D) they are the cause and effect impulse response of a filter of the required spectral shape of representative, and D is a delay operator here.Suppose we with the D conversion that is expressed as y (D) and z (D) respectively represent all sequences yk} with zk}, then, this sequence z (D) can be expressed as:
z(D)=(1-1/h(D))y(D)
This equational detailed observation is disclosed, and at a given time k, zk only depends on the numerical value in the past of yk, thereby can recursively be determined.Therefore, for the group of each 6 bit, encoder 56a will determine come which level of autocorrelative equivalence class numerically to approach most zk, and select this level.Represent this byte of this level to be sent out then.Have, we the analysis showed that again, for enough big number of levels, { yk} will have a frequency spectrum to this sequence that is sent by this central office 60, and it very closely is similar to the frequency spectrum of this filter with response h (D).
Here the technology of being narrated also can be used in conjunction with a kind of more complicated scheme in order to all information bits are mapped to all equivalence classes.For example, it can shine upon in conjunction with shell and use, and this is a kind of mapping techniques, is used for V.34 among the high speed modem specification.
All examples described above all do not have a coded system at a kind of.Yet described every principle also can easily be applied to the system of a coding, for example a trellis coding system.In this case, unique in this example difference is exactly, and all equivalence classes are divided into some subclass further, and the latter is used to constitute this trellis coding.
For example, when the signal constellation (in digital modulation) that utilizes a kind of one dimension trellis codings of dividing based on a kind of 4 tunnel set with 64 kinds of identical level comes along when sending each symbol that is made of 5 bits, all equivalence classes are divided into following subclass: a1, b1, c1, d1, a2, b2, c2, d2 ..., a
n, b
n, c
n, d
nIn above-mentioned example, these 64 equivalence classes can be divided into 4 subclass, and wherein, each subclass house has 16 equivalence classes.Then, the output of ratio-1/2 convolution coder, for example, in 6 bits in a group two determine this subclass, and these remaining 4 " uncoded " bits are chosen in this specific equivalence class of this subclass the inside.From the system of selection of the actual level of this chosen equivalence class in this chosen subclass as mentioned above.Under other different situations, the work of this encoder is constant.
Certainly, when using trellis coding, this receiver will use a decoder to go to select the most similar sequence.This trellis decoder also can be an equalizer, and it is jointly decoded and balanced intersymbol interference to this grid code.
It also is possible using the present invention to detect losing of frame synchronization in a receiver.Select the rule of this signaling point just can accomplish this point in a given equivalence class by occasionally, still periodically violating, here, this cycle is selected as the desired frame integral multiple in (cycle).In this receiver, just can detect losing of frame synchronization by monitoring such violations of rules and regulations.This receiver can also regain frame synchronization, perhaps can be simply to a synchronous mode of this transmitter requests (training sequence).
Under the prerequisite that does not deviate from its spirit or fundamental characteristics, can implement the present invention with other particular forms.It is illustrative and not restrictive that all embodiment that described are regarded as merely in all respects.Therefore, scope of the present invention is by appended claims, rather than represent by the description of front.All changes of making in the implication of the equivalent of these claims and ken all are included within their scope.
Claims (29)
1. transmitter that is used on a digital network sending a sequence of all bytes, this digital network is a sequence of all level with all byte conversion, so that send to a receiver by an analog channel, the sequence of all level produces one group of analog signal with a kind of predetermined spectral shape, and this transmitter comprises:
A mapping device is used for all data bits to be sent are mapped to a sequence of all equivalence classes, and wherein each equivalence class contains one or more level, and has at least an equivalence class to contain more than a kind of level;
An electrical level selector operationally is connected to this mapping device, in order to select a kind of level in each equivalence class, to represent all data bits to be sent;
A filter apparatus, operationally be connected this electrical level selector and have a kind of predetermined filter response, this filter apparatus receives previous selected all level and this electrical level selector is sent in its output at its input, and wherein this electrical level selector is selected this level to be sent according to the output of this filter apparatus; And
A byte converter, operationally be connected to this electrical level selector, the latter receives this selected level at its input, and on this digital network, send a byte corresponding to this selected level, when all bytes that sent with box lunch are converted into all level in this digital network, produce analog signal with predetermined spectral shape.
2. transmitter as claimed in claim 1, wherein all level are separated unevenly.
3. transmitter as claimed in claim 2, all level that wherein separate unevenly all are selected from this A rule or mu rule specification.
4. transmitter as claimed in claim 1, wherein this electrical level selector selects to approach most the level of the output of this filter apparatus in this equivalence class.
5. transmitter as claimed in claim 1 wherein should be characterised in that its response h (D)=(1-D)/(1-b D) by predetermined filter, and in the formula, b is a weighted factor, 0＜b≤1.
6. transmitter as claimed in claim 1, wherein said mapping device comprises a convolution coder, and described all equivalence classes are divided into all subclass further, make all output of this convolution coder select this subclass, and remaining all bits are selected this equivalence class.
7. data communication system that is used for uploading a sequence sending all bytes at a digital network, this digital network is a sequence of all level with all byte conversion, so that send to a receiver by an analog channel, all level produce one group of analog signal with a kind of predetermined spectral shape, and this system comprises:
A transmitter, it comprises a numerical data gatherer, the latter collects numerical data to be sent among all data group, also comprise an encoder, it is operably connected to this numerical data gatherer, this encoder is encoded to a byte with (data) group that each is collected into, and sends each byte on this digital network;
A transducer that interconnects with this digital network, being used for this all byte conversion that sent is analog quantity, is used for the analog signal of transmitting on an analog channel with generation; And
Receiver with the interconnection of this analog channel is used for all group of this analog signal conversion for the numerical data of having encoded;
This encoder comprises:
A mapping device, it receives all groups of all data bits to be sent of having collected at its input, and for each group exports an equivalence class that contains one or more level, wherein has at least an equivalence class to contain more than a kind of level;
An electrical level selector operationally is connected to this mapping device, in order to select this level in this equivalence class, to represent this group of all data bits to be sent;
A filter apparatus, operationally be connected this electrical level selector and have a kind of filter response corresponding to this predetermined spectral shape, this filter apparatus receives previous selected all level and this electrical level selector is sent in its output at its input, and wherein this electrical level selector is selected this level to be sent according to the output of this filter apparatus; And
A byte converter, operationally be connected to this electrical level selector, the latter receives this selected level at its input, and on this digital network, send a byte corresponding to this selected level, when all bytes that sent with box lunch are converted into all level in this digital network, produce analog signal with predetermined spectral shape.
8. transmitter as claimed in claim 7, wherein all level are separated unevenly.
9. transmitter as claimed in claim 8, all level that wherein separate unevenly all are selected from this A rule or mu rule specification.
10. transmitter as claimed in claim 7, wherein this electrical level selector selects to approach most the level of the output of this filter apparatus in this equivalence class.
11. transmitter as claimed in claim 7 wherein should be characterised in that its response h (D)=(1-D)/(1-b D) by predetermined filter, in the formula, b is a weighted factor, 0＜b≤1.
12. transmitter that is used on a digital network, sending a sequence of all bytes, this digital network is a sequence of all level with all byte conversion, so that send to a receiver by an analog channel, all level produce one group of analog signal that has a spectral null on flip-flop, and this transmitter comprises:
A mapping device receives all groups of all data bits to be sent at its input, and for equivalence class with one or more level of each group's output, wherein has at least an equivalence class to contain more than a kind of level;
An electrical level selector operationally is connected to this mapping device, in order to select this level in this equivalence class, to represent this group of all data bits to be sent;
A calculation element operationally is connected to this electrical level selector, in order to the mobile digital of determining previous selected all level and, wherein this electrical level selector is selected this level to be sent like this, it makes this mobile digital and approach zero most; And
A byte converter, operationally be connected to this electrical level selector, the latter receives this selected level at its input, and on this digital network, send a byte corresponding to this selected level, when all bytes that sent with box lunch are converted into all level in this digital network, produce one group and have this direct current analog signal at zero point.
13. transmitter as claimed in claim 12, wherein this mobile digital and be a weighting mobile digital and, and this selected level be numerically approach most this weighting mobile digital and level.
14. transmitter as claimed in claim 12, wherein all level are separated unevenly.
15. transmitter as claimed in claim 14, all level that wherein separate unevenly all are selected from this A rule or mu rule specification.
16. method that is used on a digital network, sending a sequence of all bytes, this digital network is a sequence of all level with all byte conversion, so that send to a receiver by an analog channel, all level produce one group of analog signal with a kind of predetermined spectral shape, and this method comprises:
Receive all groups of all data bits to be sent;
Each group of all data bits of having received is mapped to an equivalence class of one or more level, wherein has at least an equivalence class to contain more than a kind of level;
In this equivalence class, select this level, to represent this group of all data bits to be sent:
Previous selected all level are carried out filtering to produce one group of filtered output, and wherein this filter response is corresponding to this predetermined spectral shape, and wherein should select step according to should filtered output selection this level to be sent; And
Transmission is corresponding to this byte of this selected level, and all bytes that are sent out with box lunch produce the analog signal with this predetermined spectral shape when being converted into all level in this digital network.
17. method as claimed in claim 16, wherein all level are separated unevenly.
18. method as claimed in claim 17, all level that wherein separate unevenly all are selected from this A rule or mu rule specification.
19. method as claimed in claim 16, wherein this selection step is included in the level of selecting to approach most this filtered output in this equivalence class.
20. method as claimed in claim 16, wherein this filter response is characterised in that this response h (D)=(1-D)/(1-b D), and in the formula, b is a weighted factor, 0＜b≤1.
21. data transferring method that is used for uploading a sequence sending all bytes at a digital network, this digital network is a sequence of all level with all byte conversion, so that send to a receiver by an analog channel, all level produce one group of analog signal with a kind of predetermined spectral shape, and this method comprises:
Numerical data to be sent is collected among all data group;
Each group of having collected is encoded to a byte;
On this digital network, send each byte;
With all byte conversion that sent is analog quantity, to produce this analog signal, so that transmit on an analog channel; And
With all group of this analog signal conversion for the numerical data of having encoded;
The step of coding comprises:
Each group of all data bits is mapped to an equivalence class of one or more level, wherein has at least an equivalence class to contain level more than one;
In this equivalence class, select this level, to represent this group of all data bits to be sent;
Previous selected all level were carried out filtering to produce one group of filtered output, and wherein this filter response is corresponding to this predetermined spectral shape, and wherein should select step according to this output of filtering select this level to be sent;
With selected level conversion is a byte corresponding to this level of having selected; And
Send this byte, produce analog signal when all bytes that are sent out with box lunch are converted into all level in this digital network with this predetermined spectral shape.
22. method as claimed in claim 21, wherein all level are separated unevenly.
23. method as claimed in claim 22, all level that wherein separate unevenly all are selected from this A rule or mu rule specification.
24. method as claimed in claim 21, wherein this selection step selects to approach most this level of the output of this filter apparatus in this equivalence class.
25. method as claimed in claim 21, wherein this filter response is characterised in that this response h (D)=(1-D)/(1-b D), and in the formula, b is a weighted factor, 0＜b≤1.
26. method that is used on a digital network, sending a sequence of all bytes, this digital network is a sequence of all level with all byte conversion, so that send to a receiver by an analog channel, all level produce one group of analog signal that has a spectral null on flip-flop, and this method comprises:
Receive all groups of all data bits to be sent;
Each group of all data bits is mapped to the equivalence class of one or more level, wherein has at least an equivalence class to contain level more than one;
In this equivalence class, select this level, to represent this group of all data bits to be sent;
Determine previous selected all level mobile digital and, wherein this selection step is selected this level to be sent like this, makes this mobile digital and approach zero most; And
Send in this digital network corresponding to this byte of selected level, all bytes that are sent out with box lunch produce when being converted into all level in this digital network has this direct current analog signal at zero point.
27. method as claimed in claim 26, wherein determine mobile digital and this step comprise the mobile digital of determining a weighting and, and this selection step comprise selection numerically approach most this weighting mobile digital and this level.
28. method as claimed in claim 26, wherein all level are separated unevenly.
29. method as claimed in claim 28, all level that wherein separate unevenly all are selected from this A rule or mu rule specification.
Applications Claiming Priority (6)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US2668896P | 1996-09-25 | 1996-09-25 | |
| US60/026,688 | 1996-09-25 | ||
| US73043496A | 1996-10-15 | 1996-10-15 | |
| US08/730,434 | 1996-10-15 | ||
| US08/747,840 | 1996-11-13 | ||
| US08/747,840 US5818879A (en) | 1996-10-15 | 1996-11-13 | Device, system and method for spectrally shaping transmitted data signals |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1231789A true CN1231789A (en) | 1999-10-13 |
Family
ID=27362825
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 97198256 Pending CN1231789A (en) | 1996-09-25 | 1997-09-19 | Device, system and method for spectrally shaping transmitted data signals |
Country Status (5)
| Country | Link |
|---|---|
| EP (1) | EP0927481A4 (en) |
| CN (1) | CN1231789A (en) |
| AU (1) | AU4649397A (en) |
| CA (1) | CA2267362A1 (en) |
| WO (1) | WO1998013977A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1652486B (en) * | 2000-02-07 | 2011-05-18 | 通用电气资本公司 | Method for preventing and overcoming phase reversal in communication system |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5926505A (en) * | 1996-10-16 | 1999-07-20 | Cirrus Logic, Inc. | Device, system, and method for modem communication utilizing two-step mapping |
| US6480549B1 (en) | 1997-04-08 | 2002-11-12 | Vocal Technologies, Ltd. | Method for determining attenuation in a digital PCM channel |
| US6222890B1 (en) * | 1997-04-08 | 2001-04-24 | Vocal Technologies, Ltd. | Variable spectral shaping method for PCM modems |
| US6643270B1 (en) | 1998-03-03 | 2003-11-04 | Vocal Technologies, Ltd | Method of compensating for systemic impairments in a telecommunications network |
| US6414989B1 (en) * | 1999-09-10 | 2002-07-02 | Conexant Systems, Inc. | Upstream PCM transmission for a modem system |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5040191A (en) * | 1987-02-24 | 1991-08-13 | Codex Corporation | Partial response channel signaling systems |
| KR100422965B1 (en) * | 1994-12-09 | 2004-03-12 | 브렌트 타운젠드 | High speed data transfer system and method |
| US5659579A (en) * | 1995-02-01 | 1997-08-19 | Lucent Technologies Inc. | Multilevel coding for fractional bits |
-
1997
- 1997-09-19 CN CN 97198256 patent/CN1231789A/en active Pending
- 1997-09-19 CA CA002267362A patent/CA2267362A1/en not_active Abandoned
- 1997-09-19 AU AU46493/97A patent/AU4649397A/en not_active Abandoned
- 1997-09-19 EP EP97945245A patent/EP0927481A4/en not_active Withdrawn
- 1997-09-19 WO PCT/US1997/016936 patent/WO1998013977A1/en not_active Application Discontinuation
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1652486B (en) * | 2000-02-07 | 2011-05-18 | 通用电气资本公司 | Method for preventing and overcoming phase reversal in communication system |
Also Published As
| Publication number | Publication date |
|---|---|
| EP0927481A4 (en) | 2003-04-23 |
| WO1998013977A1 (en) | 1998-04-02 |
| AU4649397A (en) | 1998-04-17 |
| CA2267362A1 (en) | 1998-04-02 |
| EP0927481A1 (en) | 1999-07-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP3310664B2 (en) | Equalization method and equalization system for data communication system | |
| KR100418848B1 (en) | High speed decoder and decoding method | |
| US6084915A (en) | Signaling method having mixed-base shell map indices | |
| US6081555A (en) | Methods and apparatus for implementing shell mapping techniques in the context of a PCM-based modem communications system | |
| US6463103B1 (en) | Frame-based sign inversion method and system for spectral shaping for pulse-coded-modulation modems | |
| US6034991A (en) | Method and apparatus for implementing enhanced multiple modulus conversion techniques in a signal point mapping context | |
| US6115395A (en) | Method of detecting network impairments for high speed data communication over conventional subscriber lines | |
| US5818879A (en) | Device, system and method for spectrally shaping transmitted data signals | |
| WO1999034565A1 (en) | Device and method for precoding data signals for pcm transmission | |
| DE69917304T2 (en) | METHOD AND DEVICE FOR IDENTIFYING THE CODING TYPE OF A CODE IN A TELEPHONE DISTRIBUTION SYSTEM | |
| US6034993A (en) | Method and apparatus for adaptively equalizing a signal received from a remote transmitter | |
| JP3442793B2 (en) | DC component suppression for PCM modem | |
| Gersho et al. | Multidimensional signal constellations for voiceband data transmission | |
| AU607252B2 (en) | Digital communication system using partial response and bipolar coding techniques | |
| CN1231789A (en) | Device, system and method for spectrally shaping transmitted data signals | |
| JP3527508B2 (en) | Apparatus and method for utilizing zero padding constellation switching using frame mapping | |
| SE450679B (en) | PLANT FOR TRANSFER OF N-LEVEL DATA SIGNALS | |
| US7280592B1 (en) | Hybrid digital/analog device | |
| EP0863642B1 (en) | Shaping code for PCM modems | |
| US5995548A (en) | Signaling method using multiple modulus shell mapping | |
| US6901107B1 (en) | Systems, methods, and computer program products for generating a digital impairment learning signal having low energy content at direct current and Nyquist frequencies | |
| US20030194000A1 (en) | Distinguishing between final coding of received signals in a PCM modem | |
| EP1031217B1 (en) | Methods and apparatus for optimizing shell mapping techniques using an approximated power cost function | |
| WO1998039883A1 (en) | Signalling method using multiple modulus conversion and shell mapping | |
| CA1179794A (en) | Apparatus and method of reducing the bit rate of pcm speech |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: WD Ref document number: 1023012 Country of ref document: HK |