U.S. patent application number 10/702652 was filed with the patent office on 2004-06-03 for data transmissions device and data transmissions method capable of minimizing delay of fast data.
This patent application is currently assigned to NEC CORPORATION. Invention is credited to Okamura, Yusaku.
Application Number | 20040105454 10/702652 |
Document ID | / |
Family ID | 17731692 |
Filed Date | 2004-06-03 |
United States Patent
Application |
20040105454 |
Kind Code |
A1 |
Okamura, Yusaku |
June 3, 2004 |
Data transmissions device and data transmissions method capable of
minimizing delay of fast data
Abstract
A data transmission device for executing data transmission in
environments where the noise level changes periodically minimizes
delay of fast data. A multiplexing/synchronization control section
receives data (in which the ratio between fast data and interleave
data is fixed) which are supplied from an upper layer device at a
prespecified data rate, and separates the data into fast data and
interleave data. The fast data is supplied to a tone ordering
section without changing its data rate. Meanwhile, an interleaving
section executes interleaving to the interleave data, and a rate
converter controls the data rate of the interleave data and
supplies the interleave data to the tone ordering section. The tone
ordering section assigns the fast data and the interleave data to
each carrier of a symbol (frame) to be transmitted, based on
preobtained bit distribution with respect to low noise periods or
high noise periods, depending on whether the symbol is transmitted
in the low noise periods or in periods including the high noise
periods. In the data assignment, the amount of fast data included
in each symbol is set constant regardless of whether the symbol is
transmitted in the low noise periods or in periods including the
high noise periods, thereby the amount of fast data transmitted in
a fixed time period is made constant, and thus delay of fast data
is minimized.
Inventors: |
Okamura, Yusaku; (Tokyo,
JP) |
Correspondence
Address: |
YOUNG & THOMPSON
745 SOUTH 23RD STREET 2ND FLOOR
ARLINGTON
VA
22202
|
Assignee: |
NEC CORPORATION
TOKYO
JP
|
Family ID: |
17731692 |
Appl. No.: |
10/702652 |
Filed: |
November 7, 2003 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
10702652 |
Nov 7, 2003 |
|
|
|
09414161 |
Oct 8, 1999 |
|
|
|
6674768 |
|
|
|
|
Current U.S.
Class: |
370/413 |
Current CPC
Class: |
H04L 1/0071 20130101;
H04J 3/1682 20130101; H04L 5/006 20130101; H04L 5/023 20130101;
H04L 27/2608 20130101; H04L 2001/0098 20130101; H04L 1/0002
20130101; H04L 5/0046 20130101; H04L 5/0007 20130101; H04L 5/143
20130101; H04L 1/007 20130101; H04L 1/0018 20130101 |
Class at
Publication: |
370/413 |
International
Class: |
H04L 012/28; H04L
012/56 |
Foreign Application Data
Date |
Code |
Application Number |
Oct 9, 1998 |
JP |
10-288549 |
Claims
1. A device comprising: inputting means for inputting a signal that
represents a change of the noise level; first outputting means for
outputting first data; second outputting means for outputting
second data; wherein at least one of the amount of said first data
and of said second data changes on the basis of said signal.
2. The device as claimed in claim 1, wherein: the amount of said
first data is constant non-zero amount.
3. The device as claimed in claim 1, wherein: said change of the
noise level is that high noise period and low noise period change
periodically.
4. The device as claimed in claim 3, wherein: the amount of said
second data in low noise period is more than that of said first
data in high noise period.
5. The device as claimed in claim 1, wherein: said first data are
fast data.
6. The device as claimed in claim 5, wherein: said second data are
interleave data.
7. A device comprising: inputting circuit which inputs a signal
that represents a change of the noise level; first outputting
circuit for outputting first data; second outputting circuit for
outputting second data; wherein at least one of the amount of said
first data and of said second data changes on the basis of said
signal.
8. The device as claimed in claim 7, wherein: the amount of said
first data is constant non-zero amount.
9. The device as claimed in claim 7, wherein: said change of the
noise level is that high noise period and low noise period change
periodically.
10. The device as claimed in claim 9, wherein: the amount of said
second data in low noise period is more than that of said first
data in high noise period.
11. The device as claimed in claim 7, wherein: said first data are
fast data.
12. The device as claimed in claim 11, wherein: said second data
are interleave data.
13. A method for transmitting data, comprises the steps of: a step
for inputting a signal that represents a change of the noise level;
a step for controlling at least one of the amount of first data and
of second data; a step for outputting said first and second data on
the basis of said signal so that said at least one of the amount of
first data and of second data can change.
14. The method as claimed in claim 13, wherein: the amount of said
first data is constant non-zero amount.
15. The method as claimed in claim 13, wherein: said change of the
noise level is that high noise period and low noise period change
periodically.
16. The method as claimed in claim 15, wherein: the amount of said
second data in low noise period is more than that of said second
data in high noise period.
17. The method as claimed in claim 13, wherein: said first data are
fast data.
18. The method as claimed in claim 17, wherein: said second data
are interleave data.
Description
BACKGROUND OF THE INVENTION
[0001] The present invention relates to a data transmission device
and a data transmission method, and in particular, to a data
transmission device and a data transmission method which are
applied to XDSL (X Digital Subscriber Line (X: A, S, V, etc.))
which realizes high speed data transmission of the order of Mbit/s
by use of metal cables such as telephone lines.
[0002] Description of the Prior Art
[0003] These days, techniques concerning XDSL (X Digital Subscriber
Line (X: A (Asymmetric), S (Symmetric), V (Very high speed), etc.))
for realizing high speed data transmission of the order of Mbit/s
via metal cables such as telephone lines are attracting attention.
Most of all, ADSL (Asymmetric Digital Subscriber Line), whose bit
rate asymmetry between upstream data transmission and downstream
data transmission is suitable for the access to the Internet, is
receiving much attention.
[0004] In an ADSL transceiver unit, digital signals are converted
to analog signals according to a modulation/demodulation method
called DMT (Discrete Multi-Tone), and the analog signals are
transmitted by the ADSL transceiver unit. In the DMT method,
modulation according to QAM (Quadrature Amplitude Modulation) is
executed to 256 carriers, and the modulated carriers are
multiplexed together by means of IDFT (Inverse Discrete Fourier
Transform), and the multiplexed signal is transmitted. On the
receiving side, each modulated carrier is extracted from the
multiplexed signal by means of DFT (Discrete Fourier Transform),
and each QAM-modulated carrier is demodulated, and thereby high
speed data transmission is implemented.
[0005] However, in the case of the ADSL transceiver unit, if a
cable used for ADSL (an ADSL cable) is included in a cable bundle
that also includes an ISDN (Integrated Services Digital Network)
cable, noise is caused in the ADSL cable due to a variety of
effects of the ISDN cable, and thereby the transmission bit rate of
the ADSL cable is necessitated to be deteriorated. Most of all,
crosstalk noise from the ISDN cable has a large effect on the ADSL
cable.
[0006] In the following, the crosstalk noise which occurs to the
ADSL transceiver unit in the case where a TCM (Time Compression
Multiplex)-ISDN line is placed adjacent to an ADSL line will be
explained referring to FIG. 1. FIG. 1 shows crosstalk noise which
occurs to an ADSL terminal ATU-R (ADSL Transceiver Unit at the
Remote end) due to data transmission on the TCM-ISDN line when
downstream data transmission is being executed on the ADSL
line.
[0007] In the TCM-ISDN, upstream data transmission (1.25 msec) and
downstream data transmission (1.25 msec) are alternately executed.
If upstream data transmission on the TCM-ISDN line is executed when
downstream data transmission is being executed on the ADSL line,
high power signals on the TCM-ISDN line before attenuation exert
effects on attenuated signals on the ADSL line, and thereby NEXT
(Near End X(cross)-Talk) occurs to the terminal ATU-R. If
downstream data transmission on the TCM-ISDN line is executed when
downstream data transmission is being executed on the ADSL line,
signals on the TCM-ISDN line exert effects on attenuated signals on
the ADSL line, and thereby FEXT (Far End X(cross)-Talk) occurs to
the terminal ATU-R. Incidentally, the same effects also occur on
the side of the central office ATU-C (ADSL Transceiver Unit at the
Central office).
[0008] FIG. 2 is a schematic diagram showing noise levels of the
NEXT and the FEXT. As shown in FIG. 2, the noise level of NEXT is
higher than that of FEXT, since high power signals on the TCM-ISDN
line before attenuation exert effects on attenuated signals on the
ADSL line in the case of NEXT. There has been proposed a data
transmission method which pays attention to the difference of the
noise level, in which the data rate (data transmission bit rate) in
FEXT periods (i.e. periods in which the lower noise due to the FEXT
occurs) is set larger than the data rate in NEXT periods (i.e.
periods in which the higher noise due to the NEXT occurs) as shown
in the bottom of FIG. 2. Such a method is called "dual bit-map
method".
[0009] Since the noise level changes periodically in the ADSL
transceiver unit as mentioned above, generally, the SNR (Signal to
Noise Ratio) of each carrier through the ADSL line is measured both
in the upstream data transmission and in the downstream data
transmission of the TCM-ISDN line (that is, both in the NEXT
periods and in the FEXT periods), and thereby bit distribution and
gain distribution onto the carriers are determined with respect to
FEXT periods and with respect to NEXT periods, based on the SNR.
FIG. 3 shows an example of such bit distribution. In FIG. 3, the
horizontal axis denotes the frequency of each carrier to be used
for the data transmission, and the frequency difference between
adjacent carriers is 4.3125 kHz, and the number of carriers is 256.
Each carrier is modulated when data is transmitted. The SNR of each
carrier is evaluated in the NEXT periods and in the FEXT periods,
as a result of transmission of a pseudo random signal which will be
described below. Thereafter, actual data transmission is executed
according to the bit distribution and the gain distribution which
have been determined based on the measurement of the SNR.
[0010] In the following, a concrete example of the SNR evaluation
and the bit/gain distribution calculation will be described
referring to FIG. 3 through FIG. 5. FIG. 4 is a block diagram
showing an example of an ADSL data communication system. Referring
to FIG. 4, an ATU-C 21 is provided to the system as the central
office, and an ATU-R 22 is provided to the system as a terminal.
The ATU-C 21 includes a transmission section 23A and a reception
section 24A, and the ATU-R 22 includes a transmission section 23B
and a reception section 24B.
[0011] FIG. 5 is a block diagram showing the composition of the
ATU-C 21 and the ATU-R 22 of the ADSL data communication system of
FIG. 4. The ATU-C 21 on the left-hand side of FIG. 5 includes the
transmission section 23A, the reception section 24A, a pseudo
random signal generation section 25A, an SNR measurement section
26A, a bit/gain distribution calculation section 27A and a bit/gain
distribution table 28A. The transmission section 23A includes an
IDFT (Inverse Discrete Fourier Transform) section 10A, and the
reception section 24A includes a DFT (Discrete Fourier Transform)
section 29A. Similarly, the ATU-R 22 on the right-hand side of FIG.
5 includes the transmission section 23B, the reception section 24B,
a pseudo random signal generation section 25B, an SNR measurement
section 26B, a bit/gain distribution calculation section 27B and a
bit/gain distribution table 28B. The transmission section 23B
includes an IDFT section 10B, and the reception section 24B
includes a DFT section 29B.
[0012] In the ATU-C 21, the pseudo random signal generation section
25A generates a pseudo random signal which includes all the
carriers (256 carriers according to the standard of ANSI (American
National Standards Institution), for example) and sends the pseudo
random signal to the IDFT section 10A of the transmission section
23A so as to be transmitted to the ATU-R 22. The SNR measurement
section 26A calculates the SNR of each carrier of a pseudo random
signal supplied from the ATU-R 22, with respect to the FEXT periods
and the NEXT periods. The bit/gain distribution calculation section
27A calculates and determines the bit distribution and the gain
distribution of the carriers with respect to the FEXT periods and
the NEXT periods, based on the SNR of each carrier in the FEXT
periods and the SNR of each carrier in the NEXT periods which have
been obtained by the SNR measurement section 26A. The bit/gain
distribution table 28A receives bit distribution and gain
distribution of the carriers which have been calculated and
determined by the bit/gain distribution calculation section 27B of
the ATU-R 22, and stores the bit distribution and the gain
distribution.
[0013] In the ATU-R 22, the pseudo random signal generation section
25B generates a pseudo random signal which includes all the
carriers and sends the pseudo random signal to the IDFT section 10B
of the transmission section 23B so as to be transmitted to the
ATU-C 21. The SNR measurement section 26B calculates the SNR of
each carrier of a pseudo random signal supplied from the ATU-C 21,
with respect to the FEXT periods and the NEXT periods. The bit/gain
distribution calculation section 27B calculates and determines the
bit distribution and the gain distribution of the carriers with
respect to the FEXT periods and the NEXT periods, based on the SNR
of each carrier in the FEXT periods and the SNR of each carrier in
the NEXT periods which have been obtained by the SNR measurement
section 26B. The bit/gain distribution table 28B receives bit
distribution and gain distribution of the carriers which have been
calculated and determined by the bit/gain distribution calculation
section 27A of the ATU-C 21, and stores the bit distribution and
the gain distribution.
[0014] In the following, an example of a process executed by the
ADSL data communication system of FIG. 5 for determining the bit
distribution and the gain distribution onto the carriers in
downstream data transmission will be explained referring to FIG. 5.
First, the ATU-C 21 outputs a pseudo random signal from its pseudo
random signal generation section 25A for obtaining the bit
distribution and the gain distribution of the carriers in the
downstream data transmission. The pseudo random signal outputted by
the pseudo random signal generation section 25A is supplied to the
IDFT section 10A of the transmission section 23A and IDFT (Inverse
Discrete Fourier Transform) is executed to the pseudo random signal
by the IDFT section 10A. The pseudo random signal to which the IDFT
has been executed is transmitted to the ATU-R 22 via the ADSL line.
The pseudo random signal (to which the IDFT has been executed and
which has been transferred to the ATU-R 22) is received by the
reception section 24B of the ATU-R 22, and the signal is supplied
to the DFT section 29B. The DFT section 29B executes DFT (Discrete
Fourier Transform) to the pseudo random signal (to which the IDFT
has been executed and which has been transferred to the ATU-R 22)
and thereby obtains a received pseudo random signal which includes
all the carriers (256 carriers, for example). The received pseudo
random signal is supplied to the SNR measurement section 26B. The
SNR measurement section 26B obtains the SNR of each carrier based
on the received pseudo random signal including all the carriers,
with respect to NEXT periods and FEXT periods. The SNR of each
carrier with respect to NEXT periods and the SNR of each carrier
with respect to FEXT periods which have been measured by the SNR
measurement section 26B are supplied to the bit/gain distribution
calculation section 27B, therefore, the bit/gain distribution
calculation section 27B receives 512 SNRs (256.times.2), for
example. In the bit/gain distribution calculation section 27B, the
bit distribution and the gain distribution of the carriers are
calculated with respect to NEXT periods and FEXT periods as shown
in FIG. 3. The bit/gain distribution information of the carriers
with respect to NEXT periods and FEXT periods which has been
obtained by the bit/gain distribution calculation section 27B is
transmitted by the transmission section 23B to the ATU-C 21 via the
ADSL line. The ATU-C 21 which received the bit/gain distribution
information via the ADSL line stores the bit/gain distribution in
the bit/gain distribution table 28A. The above process for
determining the bit/gain distribution is executed in initialization
(when data communication is started, for example), in the fast
retraining period (defined in ITU standard), etc.
[0015] In the following, an ATU (ADSL Transceiver Unit) transmitter
for transmitting data according to the bit/gain distribution
obtained by the above process will be described referring to FIG.
6. FIG. 6 is a block diagram showing the composition of a
conventional ATU transmitter. Incidentally, the ATU transmitter of
FIG. 6 is a device which corresponds to the transmission section
(23A or 23B) of the ATU (ADSL Transceiver Unit) (21 or 22) shown in
FIG. 5.
[0016] The ATU transmitter shown in FIG. 6 includes a
multiplexing/synchronization control section 30, a fast data CRC
(Cyclic Redundancy Check) section 31, scrambling/error correction
sections 32 and 35, an interleave data CRC section 34, an
interleaving section 36, rate converters 33 and 37, a tone ordering
section 38, a constellation encoding/gain scaling section 39, and
an IDFT (Inverse Discrete Fourier Transform) section 40.
[0017] The ATU transmitter is provided with two types of channels:
a fast channel which is suitable for communication of data such as
video data, voice data, data for real-time application software,
etc. to which delay is harmful, and an interleave channel which is
suitable for data communication such as file transfer to which
delay is not so harmful but burst error is very harmful. The
difference between the two types of channels is the presence or
absence of interleaving process, in which the order of transmitted
data is changed and thereby the effects of the burst error is
decreased. In the interleave channel, even if burst error occurred
due to impulse noise etc., the receiving side of the interleave
channel restores the order of the rearranged data into the original
order and thereby the burst error is spread out, and thus the data
can be correctly restored by error correction. However, data have
to be temporarily stored in the interleaving process, and thus
delay is caused in the interleave process.
[0018] In the following, the operation of the ATU transmitter of
FIG. 6 will be explained briefly. Data (in which the ratio between
the amount of fast data and the amount of interleave data is
generally fixed according to initial settings) are supplied from an
upper layer device (such as a computer) to the ATU transmitter at a
prespecified data rate, and are separated by the
multiplexing/synchronization control section 30 into fast data and
interleave data. The fast data are supplied to a fast path (the
upper path in FIG. 6) and the interleave data are supplied to an
interleave path (the lower path in FIG. 6). In the fast path, the
fast data CRC section 31 adds CRC code to the fast data. The
scrambling/error correction section 32 scrambles the fast data
supplied from the fast data CRC section 31, adds Reed-Solomon error
correction code to the scrambled fast data, and supplies the data
to the rate converter 33. Also in the interleave path, the
interleave data CRC section 34 adds CRC code to the interleave
data, and the scrambling/error correction section 35 executes the
scrambling and the addition of error correction code. Thereafter,
the interleaving section 36 in the interleave path changes the
order of the interleave data supplied from the scrambling/error
correction section 35 and thereby executes the interleaving, and
sends the interleaved data to the rate converter 37.
[0019] Each of the rate converters 33 and 37 recognizes change of
the noise level due to the TCM-ISDN line by use of a TTR (TCM-ISDN
Timing Reference) signal which is in sync with the change of data
transmission direction in the TCM-ISDN line. Thereby, the rate
converter (33, 37) increases the amount of data to be supplied to
the tone ordering section 38 so as to raise the data rate in
periods in which the noise due to the TCM-ISDN line is low (i.e. in
the FEXT periods). In periods in which the noise due to the
TCM-ISDN line is high (i.e. in the NEXT periods), the rate
converter (33, 37) decreases the amount of data to be supplied to
the tone ordering section 38 so as to lower the data rate.
[0020] The tone ordering section 38 obtains information concerning
the number of bits to be assigned to each carrier, by use of the
bit distribution information of the carriers which is supplied from
the bit/gain distribution table, and determines fast data and
interleave data to be assigned to each carrier. In the
constellation encoding/gain scaling section 39, the data assigned
by the tone ordering section 38 to each carrier is converted into
constellation position information using a constellation, and
thereby corresponding Fourier transform coefficients are obtained.
The Fourier transform coefficients with respect to each carrier are
supplied to the IDFT section 40. The IDFT section 40 which received
the 512 Fourier transform coefficients with respect to the 256
carriers executes IDFT (Inverse Discrete Fourier Transform) to the
coefficients and thereby obtains a voltage signal V(t) which varies
in time. The IDFT section 40 outputs 512 sampled voltages of the
voltage signal V(t) (i.e. V(t) corresponding to 512 times "t" in a
symbol cycle) successively to the ADSL line.
[0021] FIG. 7 is a bar graph showing the ratio between the fast
data and the interleave data which are assigned to each symbol
(frame), in the case where data transmission is executed according
to the dual bit-map method which has been described above. In FIG.
7, each bar corresponds to a symbol (frame), in which all the 256
carriers are included. The FEXT/NEXT cycle (2.5 msec) of the
TCM-ISDN line corresponds to (10+.alpha.) symbols. As shown in FIG.
7, the amount of data to be assigned to a symbol is increased in
the period in which the noise is low (in the FEXT period on the
left-hand side of the graph), and is decreased in the period in
which the noise is high (in the NEXT period on the right-hand side
of the graph). However in FIG. 7, the ratio between the fast data
and the interleave data assigned to one symbol is fixed, regardless
of whether the noise is high or low.
[0022] In the data transmission according to the dual bit-map
method, in the 2.5 msec FEXT/NEXT cycle (upstream/downstream data
transmission cycle) of the TCM-ISDN line, the number of symbols in
which the amount of transmitted data can be set large (hereafter,
referred to as "FEXT symbols") and the number of symbols in which
the amount of transmitted data has to be set small (hereafter,
referred to as "NEXT symbols") can not be fixed, since the 2.5 msec
FEXT/NEXT cycle of the TCM-ISDN line corresponds to (10+.alpha.
(not an integer)) symbols on the ADSL line.
[0023] The phenomenon will be explained more concretely referring
to FIG. 8. FIG. 8 shows the timing when the data transmission
direction is changed in the TCM-ISDN line, and symbols which are
transmitted by the ADSL transceiver unit in each FEXT/NEXT cycle of
the TCM-ISDN line. Symbols shown on the left-hand side of FIG. 8
are the FEXT symbols which can carry larger amount of data because
of the low noise due to the FEXT from the TCM-ISDN line, and
symbols shown on the right-hand side of FIG. 8 are the NEXT symbols
which can carry smaller amount of data because of the high noise
due to the NEXT from the TCM-ISDN line. As shown in FIG. 8, a
little and various difference occurs between the timing when the
data transmission direction is changed in the TCM-ISDN line and the
timing when the amount of data transmitted by the ADSL line is
changed based on the TTR signal in each FEXT/NEXT cycle. Therefore,
the number of FEXT symbols which can be transmitted by the ATU
transmitter in each FEXT/NEXT cycle of the TCM-ISDN line changes
between 3 and 4 as shown in FIG. 8.
[0024] In the ADSL transceiver unit, the bit rate of data which are
supplied from the upper layer device is fixed. Therefore, if the
number of the FEXT symbols in a FEXT/NEXT cycle changes between 3
and 4, the amount of data which can be transmitted in a fixed cycle
is necessitated to be changed, and thus data are required to be
temporarily stored in the rate converter. The storage in the rate
converter causes additional delay.
[0025] Further, in the conventional dual bit-map method, the ratio
between fast data and interleave data which are assigned to one
symbol is fixed as shown in FIG. 7, regardless of whether the noise
is high or low (that is, regardless of whether the symbol is a FEXT
symbol or a NEXT symbol). Therefore, the amount of fast data which
are transmitted in a fixed time period changes in time, and thereby
delay is necessitated to be caused in the transmission of fast data
due to buffering (temporarily storing data), even though delay is
harmful to the fast data such as video data, voice data, data for
real-time application software, etc.
SUMMARY OF THE INVENTION
[0026] It is therefore the primary object of the present invention
to provide a data transmission device and a data transmission
device, by which the delay of the fast data can be minimized.
[0027] In accordance with a first aspect of the present invention,
there is provided a data transmission device which executes data
transmission in environments where the noise level changes
periodically. By the data transmission device, data transmission is
executed setting the amount of fast data which are assigned to each
frame at a constant amount even if the amount of data which can be
assigned to each frame changed due to the periodic change of the
noise level.
[0028] In accordance with a second aspect of the present invention,
in the first aspect, the amount of interleave data which are
assigned to each frame to be transmitted is controlled, based on
the change of the amount of data which are assigned to each frame
due to the periodic change of the noise level.
[0029] In accordance with a second aspect of the present invention,
in the first aspect, the data transmission device is a device which
executes data transmission using multi-carrier, comprising a
separation means, an interleaving means, an interleave data rate
control means, a tone ordering means, a coefficient conversion
means and a transform means. The separation means receives data
which are supplied from an upper layer device at a fixed data rate
including fast data and interleave data at a prespecified ratio,
and separates the data into the fast data and the interleave data.
The interleaving means executes interleaving to the interleave data
which have been separated by the separation means. The interleave
data rate control means controls the data rate of the interleave
data according to the periodic change of the noise level. The tone
ordering means assigns the fast data to the carriers of the frame
starting from a carrier that can carry the smallest number of bits
so that the amount of the fast data included in the frame will be
the constant amount, and assigns the interleave data to the
carriers of the frame starting from a carrier that can carry the
largest number of bits. The data assignment is executed based on
preobtained bit distribution of the multi-carrier with respect to
low noise periods or with respect to high noise periods, depending
on whether the frame is a frame that is transmitted in the low
noise periods or a frame that is transmitted in periods including
the high noise periods. The coefficient conversion means converts
data which have been assigned by the tone ordering means to each
carrier of the frame into corresponding coefficients, based on the
preobtained bit distribution and preobtained gain distribution of
the multi-carrier with respect to the low noise periods or with
respect to the high noise periods depending on whether the frame is
a frame that is transmitted in the low noise periods or a frame
that is transmitted in periods including the high noise periods.
The transform means transforms the coefficients with respect to the
carriers of the frame into a signal which varies in time.
[0030] In accordance with a fourth aspect of the present invention,
in the third aspect, the transform means is an IDFT (Inverse
Discrete Fourier Transform) means for executing inverse discrete
Fourier transform to Fourier transform coefficients with respect to
the carriers of the frame which have been obtained by the
coefficient conversion means and thereby obtaining the signal which
varies in time.
[0031] In accordance with a fifth aspect of the present invention,
in the third aspect, the data transmission device further comprises
a fast data CRC (Cyclic Redundancy Check) means, a first
scrambling/error correction means, an interleave data CRC means and
a second scrambling/error correction means. The fast data CRC means
executes the cyclic redundancy check and adds CRC code to the fast
data which have been separated by the separation means. The first
scrambling/error correction means scrambles the fast data supplied
from the fast data CRC means and adds error correction code to the
scrambled fast data. The interleave data CRC means executes the
cyclic redundancy check and adds CRC code to the interleave data
which have been separated by the separation means. The second
scrambling/error correction means scrambles the interleave data
supplied from the interleave data CRC means and adds error
correction code to the scrambled interleave data.
[0032] In accordance with a sixth aspect of the present invention,
in the first aspect, the data transmission device is an ATU (ADSL
(Asymmetric Digital Subscriber Line) Transceiver Unit) transmitter
which is used as a data transmitter of an ATU (ADSL Transceiver
Unit).
[0033] In accordance with a seventh aspect of the present
invention, there is provided a data transmission method for
transmitting data in environments where the noise level changes
periodically. In the data transmission method, data transmission is
executed setting the amount of fast data which are assigned to each
frame at a constant amount even if the amount of data which can be
assigned to each frame changed due to the periodic change of the
noise level.
[0034] In accordance with an eighth aspect of the present
invention, in the seventh aspect, the amount of interleave data
which are assigned to each frame to be transmitted is controlled,
based on the change of the amount of data which are assigned to
each frame due to the periodic change of the noise level.
[0035] In accordance with a ninth aspect of the present invention,
in the seventh aspect, the data transmission is executed using
multi-carrier. The data transmission method comprises a separation
step, an interleaving step, an interleave data rate control step, a
tone ordering step, a coefficient conversion step and a transform
step. In the separation step, data, which are supplied from an
upper layer device at a fixed data rate including fast data and
interleave data at a prespecified ratio, are separated into the
fast data and the interleave data. In the interleaving step,
interleaving is executed to the interleave data which have been
separated in the separation step. In the interleave data rate
control step, the data rate of the interleave data is controlled
according to the periodic change of the noise level. In the tone
ordering step, the fast data are assigned to the carriers of the
frame starting from a carrier that can carry the smallest number of
bits so that the amount of the fast data included in the frame will
be the constant amount, and the interleave data are assigned to the
carriers of the frame starting from a carrier that can carry the
largest number of bits. The data assignment is executed based on
preobtained bit distribution of the multi-carrier with respect to
low noise periods or with respect to high noise periods, depending
on whether the frame is a frame that is transmitted in the low
noise periods or a frame that is transmitted in periods including
the high noise periods. In the coefficient conversion step, data
which have been assigned to each carrier of the frame in the tone
ordering step are converted into corresponding coefficients, based
on the preobtained bit distribution and preobtained gain
distribution of the multi-carrier with respect to the low noise
periods or with respect to the high noise periods depending on
whether the frame is a frame that is transmitted in the low noise
periods or a frame that is transmitted in periods including the
high noise periods. In the transform step, the coefficients with
respect to the carriers of the frame are transformed into a signal
which varies in time.
[0036] In accordance with a tenth aspect of the present invention,
in the ninth aspect, the transform step is an IDFT (Inverse
Discrete Fourier Transform) step in which inverse discrete Fourier
transform is executed to Fourier transform coefficients with
respect to the carriers of the frame which have been obtained in
the coefficient conversion step and thereby the signal which varies
in time is obtained.
[0037] In accordance with an eleventh aspect of the present
invention, in the ninth aspect, the data transmission method
further comprises a fast data CRC (Cyclic Redundancy Check) step, a
first scrambling/error correction step, an interleave data CRC step
and a second scrambling/error correction step. In the fast data CRC
step, the cyclic redundancy check and addition of CRC code are
executed to the fast data which have been separated in the
separation step. In the first scrambling/error correction step, the
fast data as the result of the fast data CRC step are scrambled and
error correction code is added to the scrambled fast data. In the
interleave data CRC step, the cyclic redundancy check and addition
of CRC code are executed to the interleave data which have been
separated in the separation step. In the second scrambling/error
correction step, the interleave data as the result of the
interleave data CRC step are scrambled and error correction code is
added to the scrambled interleave data.
[0038] In accordance with a twelfth aspect of the present
invention, in the seventh aspect, the data transmission method is
applied to an ATU (ADSL (Asymmetric Digital Subscriber Line)
Transceiver Unit) transmitter which is used as a data transmitter
of an ATU (ADSL Transceiver Unit).
[0039] In accordance with thirteenth through eighteenth aspects of
the present invention, there are provided computer-readable record
mediums storing programs for instructing a computer or a DSP
(Digital Signal Processor) of a data transmission device to execute
the data transmission methods of the seventh through twelfth
aspects.
BRIEF DESCRIPTION OF THE DRAWINGS
[0040] The objects and features of the present invention will
become more apparent from the consideration of the following
detailed description taken in conjunction with the accompanying
drawings, in which:
[0041] FIG. 1 is a schematic diagram for explaining crosstalk noise
which occurs to an ATU-R (ADSL Transceiver Unit at the Remote end)
due to data transmission on a TCM-ISDN line when downstream data
transmission is being executed on an ADSL line;
[0042] FIG. 2 is a schematic diagram showing noise levels of NEXT
(Near End X(cross)-Talk) and FEXT (Far End X(cross)-Talk) and
explaining the dual bit-map method;
[0043] FIG. 3 is a graph showing an example of bit
distribution;
[0044] FIG. 4 is a block diagram showing an example of an ADSL data
communication system;
[0045] FIG. 5 is a block diagram showing the composition of an
ATU-C and an ATU-R of the ADSL data communication system of FIG.
4;
[0046] FIG. 6 is a block diagram showing the composition of a
conventional ATU transmitter;
[0047] FIG. 7 is a bar graph showing the ratio between fast data
and interleave data which are assigned to each symbol, in the case
where a conventional dual bit-map method is employed;
[0048] FIG. 8 is a timing chart showing the timing when the data
transmission direction is changed in the TCM-ISDN line, and symbols
which are transmitted by the ADSL transceiver unit in each
FEXT/NEXT cycle of the TCM-ISDN line;
[0049] FIG. 9 is a block diagram showing the composition of an ATU
transmitter as a data transmission device according to an
embodiment of the present invention;
[0050] FIG. 10 is a bar graph showing assignment of fast data and
interleave data to each carrier which is executed by a tone
ordering section of the ATU transmitter of FIG. 9;
[0051] FIG. 11 is a schematic diagram showing two examples of
constellations: a 4-bit constellation and a 5-bit
constellation;
[0052] FIG. 12 is a graph showing the relationship between each
position (A, B) in the 4-bit constellation of FIG. 11 and a
trigonometric function (A cos .theta.+B sin .theta.);
[0053] FIG. 13 is a graph for explaining the relationship between
gain and noise; and
[0054] FIG. 14 is a bar graph showing the assignment of fast data
and interleave data to each symbol which is executed by the ATU
transmitter according to the embodiment of the present
invention.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0055] Referring now to the drawings, a description will be given
in detail of preferred embodiments in accordance with the present
invention.
[0056] FIG. 9 is a block diagram showing the composition of an ATU
transmitter as a data transmission device according to an
embodiment of the present invention. Referring to FIG. 9, the ATU
transmitter of the embodiment includes a
multiplexing/synchronization control section 1, a fast data CRC
(Cyclic Redundancy Check) section 2, scrambling/error correction
sections 3 and 5, an interleave data CRC section 4, an interleaving
section 6, a rate converter 7, a tone ordering section 8, a
constellation encoding/gain scaling section 9, and an IDFT (Inverse
Discrete Fourier Transform) section 10. The components of the ATU
transmitter shown in FIG. 9 are implemented by, for example, one or
more DSPs (Digital Signal Processors) and appropriate software.
[0057] The multiplexing/synchronization control section 1 receives
data (in which the ratio between the amount of fast data and the
amount of interleave data is generally fixed according to initial
settings) which are supplied from an upper layer device at a
prespecified data rate, separates the data into fast data and
interleave data, and supplies the fast data and the interleave data
to a fast path (the upper path in FIG. 9) and an interleave path
(the lower path in FIG. 9) respectively.
[0058] The fast path is provided with the fast data CRC section 2
and the scrambling/error correction section 3. In the fast path,
the fast data CRC section 2 adds CRC code to the fast data supplied
from the multiplexing/synchronization control section 1. The
scrambling/error correction section 3 scrambles the fast data
supplied from the fast data CRC section 2, adds Reed-Solomon error
correction code to the scrambled fast data, and supplies the data
to the tone ordering section 8.
[0059] The interleave path is provided with the interleave data CRC
section 4, the scrambling/error correction section 5, the
interleaving section 6 and the rate converter 7. Also in the
interleave path, the interleave data CRC section 4 adds CRC code to
the interleave data, and the scrambling/error correction section 5
executes the scrambling and the addition of error correction code.
Thereafter, the interleaving section 6 in the interleave path
changes the order of the interleave data supplied from the
scrambling/error correction section 5 and thereby executes the
interleaving, and sends the interleaved data to the rate converter
7. Incidentally, delay is caused by the interleaving process.
[0060] The rate converter 7 is provided with memory for temporarily
storing interleave data, since the transmission data rate of the
interleave data has to be changed depending on the level of the
noise due to the TCM-ISDN line. The rate converter 7 changes the
amount of interleave data to be supplied to the tone ordering
section 8 depending on whether the noise due to the TCM-ISDN line
is high or low, by recognizing the change of the noise level using
the TTR (TCM-ISDN Timing Reference) signal which is in sync with
the change of the data transmission direction in the TCM-ISDN
line.
[0061] The tone ordering section 8 obtains the bit distribution
information of each carrier (with respect to FEXT periods and with
respect to NEXT periods) from the bit/gain distribution table 28,
and thereby rearranges the carriers from the viewpoint of the
number of bits which can be carried by the carrier (that is,
obtains a new order of the carriers based on the number of
assignable bits) as shown in FIG. 10. Referring to FIG. 10, a
carrier that can carry the largest number of bits is placed at the
left-hand end of FIG. 10, and a carrier that can carry the smallest
number of bits is placed at the right-hand end of FIG. 10. The tone
ordering section 8 assigns (distributes) the interleave data to the
carriers starting from the carrier that can carry the largest
number of bits, and assigns the fast data starting from the carrier
that can carry the smallest number of bits, as shown in FIG.
10.
[0062] The constellation encoding/gain scaling section 9 obtains
the bit distribution onto the carriers and the gain distribution
onto the carriers from the bit/gain distribution table 28.
Thereafter, the constellation encoding/gain scaling section 9
refers to a constellation that corresponds to the number of bits
which are assigned to each carrier according to the bit
distribution. For example, in the case where 4 bits of data has
been assigned by the tone ordering section 8 to a particular
carrier according to the bit distribution, a constellation for
4-bit data (a 4-bit constellation) is prepared for the particular
carrier and referred to. FIG. 11 is a schematic diagram showing two
examples of constellations, in which a 4-bit constellation and a
5-bit constellation are shown. Subsequently, a gain scaling
coefficient "g" is determined for each constellation corresponding
to a carrier, according to the gain distribution information with
regard to the carrier. Subsequently, with respect to each carrier,
a position (i.e. coordinates (A, B)) in the constellation that
corresponds to the bits (data) which have been assigned by the tone
ordering section 8 to the carrier is found out, and the coordinates
(A, B) is multiplied by the gain scaling coefficient "g". For
example, in the case where 4-bit data "0101" has been assigned by
the tone ordering section 8 to a particular carrier according to
the bit distribution and the gain scaling coefficient "g" of the
particular carrier should be set at 1.5 according to the gain
distribution, the 4-bit constellation shown in FIG. 11 is referred
to and thereby a position (coordinates (A, B)) corresponding to the
4-bit data "0101" (i.e. a number "5" in decimal digits) is found
out (that is, (1, -1) in FIG. 11), and thereafter the coordinates
(1, -1) are multiplied by the gain scaling coefficient 1.5, thereby
coordinates (gA, gB)=(1.5, -1.5) corresponding to the 4-bit data
"0101" are obtained as Fourier transform coefficients. Thereafter,
the Fourier transform coefficients (gA, gB) with respect to each
carrier are supplied to the IDFT section 10. Incidentally, FIG. 12
is a graph showing the relationship between each position (A, B) in
the 4-bit constellation of FIG. 11 and a trigonometric function (A
cos .theta.+B sin .theta.) as a Fourier component (before gain
scaling) of a voltage signal V(t) which will be described
below.
[0063] The IDFT section 10 which received the 512 Fourier transform
coefficients with respect to the 256 carriers executes IDFT
(Inverse Discrete Fourier Transform) to the coefficients and
thereby obtains the aforementioned voltage signal V(t) which varies
in time. The IDFT section 10 outputs 512 sampled voltages of the
voltage signal V(t) (i.e. V(t) corresponding to 512 times "t" in a
symbol cycle) successively to the ADSL line.
[0064] Here, an explanation on the relationship between gain and
noise will be given. FIG. 13 is a graph for explaining the
relationship between gain and noise. As shown in FIG. 13, the
distance between the origin and each point becomes large when gain
is increased. When the number of transmitted bits becomes large as
indicated by the downward arrows in FIG. 13, the probability of
error due to noise is necessitated to be high if the bits are
transmitted without raising the gain since the distance between
points in the constellation is small. Therefore, even when the
noise level is high, the probability of error due to noise can be
reduced by raising the gain and increasing the distance between
points in the constellation. Incidentally, information concerning
the gain (i.e. the gain scaling coefficient "g") with respect to
each carrier is transmitted to an ADSL transceiver unit on the
receiving side.
[0065] The bit distribution and the gain distribution which are
stored in and outputted by the bit/gain distribution table 28 can
be obtained by the ADSL data communication system shown in FIG. 5,
as described before in the "Description of the Prior Art". For
example, the process executed by the ADSL data communication system
of FIG. 5 for determining the bit/gain distribution in downstream
data transmission is as follows. First, the ATU-C 21 shown in FIG.
5 outputs a pseudo random signal from its pseudo random signal
generation section 25A. The pseudo random signal outputted by the
pseudo random signal generation section 25A is supplied to the IDFT
section 10A of the transmission section 23A and IDFT (Inverse
Discrete Fourier Transform) is executed to the pseudo random
signal. The pseudo random signal to which the IDFT has been
executed is transmitted to the ATU-R 22 via the ADSL line. The
pseudo random signal (to which the IDFT has been executed and which
has been transferred to the ATU-R 22) is received by the reception
section 24B of the ATU-R 22, and the signal is supplied to the DFT
section 29B. The DFT section 29B executes DFT (Discrete Fourier
Transform) to the pseudo random signal (to which the IDFT has been
executed and which has been transferred to the ATU-R 22) and
thereby obtains a received pseudo random signal which includes all
the carriers (256 carriers, for example). The SNR measurement
section 26B obtains the SNR of each carrier based on the received
pseudo random signal including all the carriers. The measurement of
SNR is executed both with respect to NEXT periods and with respect
to FEXT periods. The SNR of each carrier with respect to NEXT
periods and the SNR of each carrier with respect to FEXT periods
which have been obtained by the SNR measurement section 26B are
supplied to the bit/gain distribution calculation section 27B, and
the bit/gain distribution calculation section 27B calculates the
bit distribution and the gain distribution of the carriers with
respect to NEXT periods and FEXT periods as shown in FIG. 3. The
bit/gain distribution information of the carriers with respect to
NEXT periods and FEXT periods which has been obtained by the
bit/gain distribution calculation section 27B is transmitted by the
transmission section 23B to the ATU-C 21 via the ADSL line. The
ATU-C 21 which received the bit/gain distribution information via
the ADSL line stores the bit/gain distribution in the bit/gain
distribution table 28A. The above process for determining the
bit/gain distribution is executed in initialization (when data
communication is started, for example), in the fast retraining
period (defined in ITU standard), etc.
[0066] The conventional ATU transmitter of FIG. 6 employing the
conventional dual bit-map method sets the ratio between fast data
and interleave data in one symbol (frame) at a fixed ratio,
regardless of whether the symbol is a FEXT symbol (i.e. a symbol
that is transmitted in the FEXT periods in which noise due to the
adjacent line is low) or a NEXT symbol (i.e. a symbol that is
transmitted in periods including the NEXT periods in which noise
due to the adjacent line is high), as has been explained referring
to FIG. 7. Therefore, when the number of FEXT symbols which can be
transmitted in a FEXT period decreases (4 to 3, for example), the
amount of transmitted fast data decreases, and thereby delay is
caused in the transmission of fast data.
[0067] For example, in the TCM-ISDN, upstream data transmission
(1.25 msec) and downstream data transmission (1.25 msec) are
alternately executed, and thus the noise level due to crosstalk
from the TCM-ISDN line to the ADSL line changes periodically with a
noise cycle (FEXT/NEXT cycle) of 2.5 msec. Therefore, the ATU
transmitter has to execute data transmission in sync with the
change of the noise level. As explained before referring to FIG. 8,
a little and various difference occurs between the timing when the
data transmission direction is changed in the TCM-ISDN line and the
timing when the amount of data transmitted by the ADSL line is
changed based on the TTR signal in each FEXT/NEXT cycle. Therefore,
the number of FEXT symbols which can be transmitted by the ATU
transmitter in each FEXT/NEXT cycle changes between 3 and 4,
thereby the amount of fast data transmitted in a fixed time period
changes, and thereby delay due to buffering (temporarily storing
data) is caused in the transmission of fast data. The delay of fast
data has to be minimized since the delay is harmful to the fast
data such as video data, voice data, data for real-time application
software, etc.
[0068] FIG. 14 is a bar graph showing the assignment of fast data
and interleave data to each symbol (frame) which is executed by the
ATU transmitter according to the embodiment of the present
invention. Referring to FIG. 14, the amount of fast data which are
assigned to each symbol is fixed, regardless of whether the symbol
is a NEXT symbol or a FEXT symbol. By such assignment, the amount
of fast data which are transmitted in a fixed time period can be
made constant, thereby the delay of fast data due to buffering
(temporarily storing data) can be eliminated, and thereby the delay
of the fast data (such as video data, voice data, data for
real-time application software, etc. to which delay is harmful) can
be minimized. Incidentally, by the constant assignment of the fast
data, even if the number of FEXT symbols in each FEXT/NEXT cycle
changed, the amount of fast data which are transmitted in each
FEXT/NEXT cycle can be made constant, differently from the case of
the conventional data transmission device shown in FIG. 7.
[0069] In the following, the operation of the ATU transmitter of
FIG. 9 for implementing the data assignment explained above will be
described in detail. Incidentally, in FIG. 9, a rate converter is
not necessary in the fast path, since transmission data rate of the
fast data is set constant in this embodiment regardless of the
change of the noise level. On the other hand, the transmission data
rate of the interleave data has to be changed depending on the
change of the noise level, and thus the rate converter 7 is
provided to the interleave path in order to control the data rate
of interleave data which are supplied to the tone ordering section
8. The rate converter 7 sets the amount of interleave data to be
supplied to the tone ordering section 8 at a smaller level in the
NEXT periods (in which the noise due to the adjacent line is high),
and sets the amount at a larger level in the FEXT periods (in which
the noise due to the adjacent line is low), and sends the
interleave data to the tone ordering section 8.
[0070] Data (in which the ratio between the amount of fast data and
the amount of interleave data is generally fixed according to
initial settings) are supplied from an upper layer device (such as
a computer) to the ATU transmitter of FIG. 9 at a prespecified data
rate. The data are separated by the multiplexing/synchronization
control section 1 into fast data and interleave data, and the fast
data and the interleave data are supplied to the fast path and the
interleave path respectively. In the fast path, CRC code is added
by the fast data CRC section 2 to the fast data, and the fast data
outputted by the fast data CRC section 2 is scrambled by the
scrambling/error correction section 3. The scrambling/error
correction section 3 adds Reed-Solomon error correction code to the
scrambled fast data, and supplies the data to the tone ordering
section 8 without changing data rate. Also in the interleave path,
CRC code is added by the interleave data CRC section 4 to the
interleave data, and the scrambling and the addition of error
correction code are executed by the scrambling/error correction
section 5. Thereafter, the interleaving section 6 in the interleave
path changes the order of the interleave data supplied from the
scrambling/error correction section 5 and thereby executes the
interleaving, and sends the interleaved data to the rate converter
7.
[0071] In the rate converter 7 of the interleave path, the amount
of interleave data to be supplied to the tone ordering section 8 is
set at a larger level for FEXT symbols which are transmitted in the
FEXT periods (in which the noise due to the adjacent TCM-ISDN line
is low), and the amount is set at a smaller level for NEXT symbols
which are transmitted in periods including the NEXT periods (in
which the noise due to the adjacent TCM-ISDN line is high).
[0072] For example, the rate converter 7 can execute the control on
the amount of interleave data supplied to the tone ordering section
8, depending on whether it is in the FEXT periods or in the NEXT
periods. Incidentally, in such control, a little delay due to
processing by the tone ordering section 8, the constellation
encoding/gain scaling section 9 and the IDFT section 10 is taken
into consideration by the rate converter 7.
[0073] Or, it is also possible to let the rate converter 7
recognize the noise cycles (FEXT/NEXT cycles) (0.about.33) shown in
FIG. 8 by use of the TTR signal supplied thereto, and grasp the
changing timing between the FEXT symbols and NEXT symbols in each
noise cycle, and thereby change the amount of interleave data to be
supplied to the tone ordering section 8 in sync with the changing
timing. Also in this example, a little delay due to processing by
the tone ordering section 8, the constellation encoding/gain
scaling section 9 and the IDFT section 10 is taken into
consideration by the rate converter 7.
[0074] The tone ordering section 8 obtains the bit distribution
information of each carrier (with respect to the FEXT periods and
with respect to the NEXT periods) from the bit/gain distribution
table 28, and rearranges the carriers from the viewpoint of the
number of bits which can be carried by the carrier (that is,
obtains a new order of the carriers based on the number of
assignable bits) as shown in FIG. 10 (in which a carrier that can
carry the largest number of bits is placed at the left-hand end,
and a carrier that can carry the smallest number of bits is placed
at the right-hand end) according to the bit distribution
information. The above process is executed by the tone ordering
section 8 with respect to the FEXT periods and the NEXT periods, in
initialization (when data communication is started, for example),
in the fast retraining period (defined in ITU standard), etc. The
tone ordering section 8 assigns the interleave data to the carriers
starting from the carrier that can carry the largest number of bits
and assigns the fast data starting from the carrier that can carry
the smallest number of bits as shown in FIG. 10, according to the
above rearranged bit distribution with respect to the FEXT periods
(for the FEXT symbols) or according to the above rearranged bit
distribution with respect to the NEXT periods (for the NEXT
symbols). In the assignment of the fast data and the interleave
data to the carriers of a (FEXT or NEXT) symbol, the tone ordering
section 8 keeps the amount of fast data included in the symbol
constant (so as to match the amount of fast data supplied from the
upper layer device to the ATU transmitter in each symbol cycle)
regardless of whether the symbol is a FEXT symbol or a NEXT symbol,
as shown in FIG. 14.
[0075] The constellation encoding/gain scaling section 9, which is
supplied with the bit distribution information and the gain
distribution information (with respect to the FEXT periods and the
NEXT periods) from the bit/gain distribution table 28, figures out
the Fourier transform coefficients (gA, gB) with respect to each
carrier of the symbol (to each carrier of which data have been
assigned by the tone ordering section 8), according to the method
which has been described before. The process for obtaining the
Fourier transform coefficients (gA, gB) with respect to each
carrier of a symbol is executed based on different bit/gain
distribution (the bit/gain distribution with respect to FEXT
periods or the bit/gain distribution with respect to NEXT periods)
depending on whether the symbol is a FEXT symbol or a NEXT symbol.
The Fourier transform coefficients (gA, gB) with respect to each
carrier of the symbol are supplied to the IDFT section 10.
[0076] The IDFT section 10 which received the 512 Fourier transform
coefficients with respect to the 256 carriers of the symbol
executes IDFT (Inverse Discrete Fourier Transform) to the
coefficients, thereby obtains the voltage signal V(t) corresponding
to the symbol, and outputs 512 sampled voltages of the voltage
signal V(t) (i.e. V(t) corresponding to 512 times "t" in a symbol
cycle) successively to the ADSL line.
[0077] As described above, in the ATU transmitter according to the
embodiment of the present invention, the amount of fast data which
are assigned to each symbol to be transmitted is set constant,
regardless of whether the symbol is a FEXT symbol or a NEXT symbol.
Therefore, the amount of fast data which are transmitted in a fixed
time period can be made constant, thereby the delay of fast data
due to buffering (temporarily storing data) can be eliminated, and
thereby the delay of the fast data such as video data, voice data,
data for real-time application software, etc. to which delay is
harmful can be minimized. By the constant assignment of the fast
data, the effect of the change of the number of FEXT symbols in
each FEXT/NEXT cycle (that is, the change of the amount of fast
data which are transmitted in each FEXT/NEXT cycle) is also
eliminated, differently from the case of the conventional data
transmission device shown in FIG. 7.
[0078] As set forth hereinabove, in the data transmission device
and the data transmission method according to the present
invention, data transmission is executed setting the amount of fast
data which are assigned to each transmitted frame (symbol) at a
constant amount even if the amount of data (fast data+interleave
data) which can be assigned to each frame changed due to the
periodic change of the noise level. Therefore, the amount of fast
data which are transmitted in a fixed time period can be made
constant, and thereby delay of the fast data (such as video data,
voice data, data for real-time application software, etc. to which
delay is harmful) can be minimized.
[0079] While the present invention has been described with
reference to the particular illustrative embodiments, it is not to
be restricted by those embodiments but only by the appended claims.
It is to be appreciated that those skilled in the art can change or
modify the embodiments without departing from the scope and spirit
of the present invention. For example, the data transmission device
according to the present invention is not limited to ATU (ADSL
Transceiver Unit) transmitters, but can also be applied to various
types of transmitters of XTUs (XDSL Transceiver Units) which are
used for data communication according to XDSL (X Digital Subscriber
Line (X: A (Asymmetric), S (Symmetric), V (Very high speed),
etc.)). The present invention can be applied to various types of
data transmission devices which are employed for data communication
in environments where the noise level (by any causes including
causes other than TCM-ISDN lines) changes periodically. The present
invention can also be applied to data transmission devices which
employs modulation/demodulation methods other than DMT (Discrete
Multi-Tone). Further, the present invention is not limited to wire
data transmission devices, but can also be applied to data
transmission devices which execute data transmission by radio
waves. Moreover, while the above explanation has been given on the
assumption that "fast data" means video data, voice data, data for
real-time application software, etc., the classification of data
into the "fast data" and the "interleave data" can be done in
various ways depending on design requirements etc.
* * * * *