Process for transmitting and receiving a signal

• US 5,227,990 A
• Filed: 01/17/1992
• Issued: 07/13/1993
• Est. Priority Date: 05/17/1989
• Status: Expired due to Term

• Current Assignee: Thomson Consumer Electronics Sales GmbH (Vantiva SA)
• Original Assignee: Telefunken Fernseh Und Rundfunk GmbH (Vantiva SA)
• Inventors: Vaupel, Thomas, Dickopp, Gerhard, Krahe, Detlef
• Primary Examiner(s): NGUYEN, LONG
• Application Number:
  US07/824,848
• Time in Patent Office: 543 Days
• Field of Search: 364/724.01, 364/726, 381/36
• US Class Current: 708/303
• CPC Class Codes: H04B 1/66 for reducing bandwidth of signals; for improving efficiency of transmission H04B1/68 takes precedence

Abstract

A signal is divided by windows into successive blocks that overlap at least by 50% and signal sections contained in the blocks are evaluated by analysis windows. The signal sections contained in the blocks are subjected to a transformation which permits sub-sampling with compensation of the aliasing components such as, for example, the "time domain aliasing cancellation" method. The spectra resulting from the transformation are subsequently coded, transmitted, decoded after transmission and changed back to signal sections by re-transformation. The blocks including the signal sections are evaluated by synthesis windows and are joined together in an overlapping fashion. The window functions of the synthesis windows are determined in dependence on the window functions of the corresponding analysis windows in the overlap region.

Claims

1. A communication method for use in a communications system having a transmitter and receiver, comprising:

dividing a signal using windows into successive blocks that overlap by at least 50%, signal sections contained in the blocks being evaluated with analysis windows;

subjecting the signal sections contained in the blocks to a transformation which permits sub-sampling with compensation of aliasing components, said transformation performing time domain aliasing cancellation;

wherein spectra resulting from the transformation are subsequently transmitted from the transmitter to a receiver and are changed back to blocks having signal sections by retransformation;

wherein the retransformed blocks having the signal sections are evaluated with synthesis windows and are joined together in an overlapping fashion; and

wherein window functions of the synthesis windows are determined in dependence on window functions of corresponding analysis windows in the overlap region, according to the following equations;

##EQU5## where a_n (t) is the analysis window function for block n;

s_n (t) is the synthesis window function for block n;

a_n+1 (t) is the analysis window function for block n+1;

s_n+1 (t) is the synthesis window function for block n+1; and

T_B is the block time.

2. A method according to claim 1, wherein, for symmetrical windows, the window functions of the synthesis windows are determined in dependence on the window functions of the analysis windows according to the following equation:

##EQU6## where a(t) is the analysis window function;

s(t) is the synthesis window function; and

T_B is the block time.

3. A method according to claim 1, wherein the sub-sampling is effected so that, after the transformation, in the m^th block, the real components of the spectral values having an even-numbered frequency index and the imaginary components of the spectral values having an odd-numbered frequency index and, in the (m+1)^th block, the imaginary components of the spectral values having an even-numbered frequency index and the real components of the spectral values having an odd-numbered frequency index, are selected for coding and transmission.

4. A method according to claim 1, wherein the sub-sampling is effected so that, after the transformation, a pseudo magnitude and pseudo phase representation is formed in that the real components of the spectral values having the frequency indices n are each combined with the adjacent imaginary components of the spectral values having the frequency indices n+1 to form magnitude and phase.

5. A method according to claim 1, wherein the sub-sampling is effected so that, after the transformation, two successive blocks are combined and a pseudo magnitude and pseudo phase representation is formed in that, in each case, the real components of the spectral values of the m^th block and the imaginary components of the spectral components of the (m+1)^th block having the same frequency indices are selected and magnitude and phase are calculated therefrom.

6. A method according to claim 1, wherein the transformation is a modified cosine transformation which is performed alternatingly block by block according to the following equation:

##EQU7## where y represents the spectral values;

k represents the indices of the spectral values;

D is a constant according to the following definition;

space="preserve" listing-type="equation">D.sub.0 (for k=0)=√

N;

D.sub.k =√

2/N;

1≦

k≦

N-1;

x represents the time values;

n represents the indices of the time values of the sampling; and

N represents the number of sampling values within a block,

with a modified sine transformation being performed according to the following equation;

##EQU8## where y represents the spectral values;

k represents the indices of the spectral values;

D is a constant according to the following definition;

space="preserve" listing-type="equation">D.sub.k =√

2/N;

1≦

k≦

N-1;

D.sub.N =√

N

x represents the time values;

n represents the indices of the time values of the sampling; and

N represents the number of sampling values within a block,

with an inverse transformation relative to the forward transformation being performed for the re-transformation.

7. A method according to claim 6, wherein from the available spectral values, the even-numbered spectral values are selected and then complex values are formed from the spectral values that have been employed in that, from the spectral values of every two successive overlapping blocks, the values of the first block are multiplied by 1 and the values of the second block are multiplied by j and the values having the same indices in the two blocks are added together.

8. A communication method for use in a communications system having a transmitter and receiver, comprising:

performing analog to digital conversion on an analog signal to form digital samples;

selecting a series of successive sampling values from the digital samples;

forming blocks that overlap in time by fifty percent from the selected sampling values, wherein adjacent blocks partially contain the same sampling values but at different locations, the sampling values in a first half of a respective block corresponding to sampling values in a second half of a preceding block;

evaluating the formed blocks using analysis windows, wherein at borders of the formed blocks, a soft signal start and decay is produced;

transforming the formed blocks from the time to the frequency domain to thereby produce spectral values;

converting the spectral values into a form containing pseudo magnitudes and pseudo phases, and simultaneously sub-sampling, to produce a result in which a doubling of data caused by the fifty percent overlapping of the formed blocks is cancelled out;

transmitting the result of the converting and sub-sampling step from the transmitter to a receiver;

inverse transforming the spectral values in the transmitted result from the frequency domain to the time domain to produce blocks, wherein the produced blocks contain only one-half of the sampling values of the originally formed blocks;

evaluating the produced blocks using synthesis windows to compensate for the evaluation performed by the analysis windows, wherein window functions of the synthesis windows, in a region of overlap with a corresponding analysis window, supplement each other to equal one, and the analysis window mirrored in the middle of the overlap region multiplied by a synthesis window for block n, minus the analysis window mirrored in the middle of the overlap region multiplied by a synthesis window for block n+1 in the overlap region, equals zero;

adding fifty percent overlapping blocks, aliasing components in two blocks to be added occurring in each case with the opposite sign so that the aliasing components are compensated to zero during the addition;

joining the blocks resulting from the addition to form continuous sampling values; and

performing digital to analog conversion on the continuous sampling values to form an analog signal.

9. The method of claim 8, further comprising performing data reduction prior to transmitting.

10. The method according to claim 8, wherein the sub-sampling is performed so that, in the m^th block, the real components of the spectral values having an even-numbered frequency index are employed for transmission as well as the imaginary components of the spectral values having an odd-numbered frequency index, and in the (m+1) block, the real components of the spectral values having an odd-numbered frequency index and the imaginary components of the spectral values having an even-numbered frequency index are employed for transmission, whereby for a block having N values and an overlap of about fifty percent, only N/4 real components and N/4 imaginary components are required for transmission.

11. The method according to claim 8, wherein the transforming step to transform the blocks from the time to the frequency domain to thereby produce spectral values comprises performing Fourier transformation.