JPH088933A - Voice cell coder - Google Patents

Abstract

PURPOSE:To obtain a voice cell coder in which deterioration in quality of a received voice signal due to missing of information on the cell abort is minimized. CONSTITUTION:The coder is a voice cell coder in which a voice signal is processed into a cell, the processed cell is sent from a transmission section to a reception section and the cell is aborted on a transmission line. The coder is provided with plural transmission interpolation means 103-106, subtractor means 107-110, accumulation means 111-114, a comparator means 115, a multiplexer means 117, a demultiplexer means 118, plural reception interpolation means 120-123, a 1st selection means 124, a cell abort detection means 120 and a 2nd selection means 125. Then the transmission interpolation means 103-106 and the reception interpolation means 120-123 have the interpolation methods of the same kind. Tn this case, the 1st selection means 124 selects a reception interpolation means having the same interpolation method as that of a transmission interpolation means selected by the comparator means 115 to provide a reception optimum interpolation signal 37.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to coded transmission of a voice signal, and more particularly to compensation for missing information when a cell is discarded in cell transmission.

[0002]

2. Description of the Related Art Cell discard may occur when a voice signal is converted into cells and transmitted, such as when the voice encoded signal is converted into ATM (Asynchronous Transfer Mode) cells and transmitted.
Conventionally, cell interpolation has been performed to compensate for the discarded cells. As a conventional cell interpolation method, measures have been taken such as interpolating the immediately preceding cell at a location missing due to cell discarding, or inserting silence to perform interpolation.

[0003]

As the audio signal to be encoded, signals of various levels, frequencies and waveforms are input. Therefore, if cell discard is performed at the time of high-level signal input, if silence is inserted and interpolation is performed, the quality degradation of the received voice signal due to cell discard becomes very noticeable, and the level changes linearly. Even when the immediately preceding cell is interpolated during the operation, there is a problem that the quality deterioration of the received voice signal is very noticeable.

[0004] In view of the above, it is an object of the present invention to provide a speech cell coding apparatus which minimizes the quality degradation of a received speech signal due to a loss of information when a cell is discarded.

[0005]

SUMMARY OF THE INVENTION An audio cell encoding apparatus according to the present invention is an audio cell encoding apparatus for converting an audio signal into cells, transmitting the cells from a transmitting section to a receiving section, and discarding cells on a transmission path. The transmitting unit outputs a transmission interpolation signal by interpolating a cell, a plurality of transmission interpolation units having different types of interpolation methods, and a subtraction unit that calculates an interpolation error between the transmission interpolation signal and the transmission audio signal, Accumulating means for accumulating the interpolation error by a capacity constituting one cell to output a cumulative interpolation error, and selecting one of the interpolation methods by comparing the cumulative interpolation error for each interpolation method; A comparison means for selecting the transmission interpolation means having the selected interpolation method and generating a transmission optimum interpolation means selection signal; the transmission sound signal and the transmission optimum interpolation means selection signal of a cell immediately before the transmission sound signal; Multiplex Multiplexing means for generating a transmission cell signal and outputting the transmission cell signal to a transmission path, wherein the receiving section inputs the transmission cell signal as a reception cell signal, and a reception voice signal corresponding to the transmission voice signal and the transmission optimum. Separation means for separating the reception optimum interpolation means selection signal corresponding to the interpolation means selection signal, a plurality of reception interpolation means having different kinds of interpolation methods for interpolating cells and outputting reception interpolation signals, and the reception optimization First selection means for selecting one of the reception interpolation signals by the interpolation means selection signal and outputting it as a reception optimum interpolation signal, cell discard detection means for detecting cell discard, the received voice signal and the reception A second selection means for selecting and outputting one of the optimum interpolation signals, and the plurality of transmission interpolation means and the plurality of reception interpolation means have the same kind of interpolation method .

In the above-described speech cell encoding apparatus of the present invention, the comparing means can select the interpolation method for outputting the minimum value of the cumulative interpolation error for each cell.

Further, in the above-mentioned speech cell encoding apparatus of the present invention, the first selection means may include the reception interpolation means having the same interpolation method as that of the transmission interpolation means selected by the comparison means. The selected optimal interpolation signal can be output.

Further, in the speech cell encoding apparatus according to the present invention, the second selecting means outputs the reception optimum interpolation signal when a cell is discarded on a transmission line, and the cell discarding is performed on the transmission line. If no signal is generated, the received audio signal can be output.

[0009]

A plurality of transmission interpolation means and a plurality of reception interpolation means have the same type of interpolation method, and the first selection means uses the same interpolation method as that of the transmission interpolation means selected by the comparison means. Since the receiving interpolation means is selected and the optimum receiving interpolation signal is output, the receiving unit selects the same type of interpolation method as the optimum one selected by the transmitting unit to optimize the cells discarded on the transmission path. It is possible to perform interpolation by a simple interpolation method, and it is possible to minimize the deterioration of the quality of the received voice signal due to the lack of information when the cell is discarded.

[0010]

Embodiments of the present invention will now be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

First, the structure and operation of the transmitter will be described. The transmission voice signal 10 is input from the voice input terminal A. The transmission voice signal 10 is input to the A / D converter 101, subjected to analog / digital conversion, and output as a transmission encoded voice signal 11. The transmission encoded speech signal 11 is input to a cell conversion circuit 102, transmission interpolation circuits 103 to 106, and subtracters 107 to 110.

[0012] The cell conversion circuit 102 assembles the input transmission coded voice signal 11 into M bytes (M is a natural number) as one cell, and outputs it as a transmission cell coded signal 12. On the other hand, the transmission interpolation circuits 103 to 106 perform interpolation on the input transmission coded speech signal 11 by different methods, respectively, and subtracters 107 to 110 calculate respective interpolation errors. The accumulating circuits 111 to 114 accumulate the respective interpolation errors and input them to the comparison circuit 115 as the cumulative interpolation error.

The transmission interpolation circuit 103 interpolates the input M bytes of the transmission encoded voice signal 11 with a silence signal, and outputs it as a transmission interpolation signal 13 to the subtractor 107. Subtractor 1
07, subtracts the transmission interpolation signal 13 and the transmission encoded voice signal 11 to calculate the interpolation error 17, and outputs it to the accumulating circuit 111. The accumulating circuit 111 accumulates the interpolation error 17 for M bytes forming one cell, and the accumulated result is the cumulative interpolation error 2
It is output as 1 to the comparison circuit 115.

The transmission interpolation circuit 104 performs 0-order hold interpolation on the input M-byte transmission coded voice signal 11 from the last byte signal of the immediately preceding cell, and generates a transmission interpolation signal 14.
To the subtractor 108. The subtractor 108 subtracts the transmission interpolation signal 14 and the transmission encoded voice signal 11 to calculate the interpolation error 18, and outputs the interpolation error 18 to the accumulating circuit 112. The accumulation circuit 112 accumulates the interpolation error 18 for M bytes constituting one cell, and outputs the accumulation result to the comparison circuit 115 as the accumulated interpolation error 22.

The transmission interpolator 105 outputs the input M bytes of the transmission encoded voice signal 11 to 1 from the signal level of the last byte of the cell immediately before and the signal level of the first byte of the cell one after. Subsequent interpolation is performed and the transmission interpolation signal 15 is output to the subtractor 109. The subtracter 109 outputs the transmission interpolation signal 1
5 and the transmission coded audio signal 11 to perform subtraction to obtain an interpolation error 1
9 is calculated and output to the accumulation circuit 113. Accumulation circuit 11
3 accumulates the interpolation error 19 for M bytes constituting one cell, and uses the accumulation result as a cumulative interpolation error 23 in the comparison circuit 11.
5 is output.

The transmission interpolation circuit 106 interpolates the input M-byte transmission encoded voice signal 11 with all the bytes of the immediately preceding cell, and outputs the result to the subtracter 110 as a transmission interpolation signal 16. The subtractor 110 subtracts the transmission interpolation signal 16 and the transmission encoded voice signal 11 to calculate the interpolation error 20 and outputs the interpolation error 20 to the accumulating circuit 114. The accumulating circuit 114 accumulates the interpolation error 20 for M bytes forming one cell, and outputs the accumulated result as a cumulative interpolation error 24 to the comparison circuit 115.

The comparison circuit 115 detects the minimum cumulative interpolation error among the input cumulative interpolation errors 21 to 24,
The number of the interpolation circuit that has output the minimum cumulative interpolation error is output to the multiplexing circuit 117 as the transmission optimum interpolation circuit selection signal 25. The cell counter 116 counts continuous cell numbers and outputs the transmission cell number 26 to the multiplexing circuit 117. The multiplexing circuit 117 multiplexes the input transmission cell coded signal 12 with the transmission optimum interpolation circuit selection signal 25 and the transmission cell number 26 and outputs the multiplexed signal as the transmission cell signal 27. The transmission cell signal 27 is output to the transmission line from the transmission line output terminal B of the transmission unit.

Next, the structure and operation of the receiver will be described.
The transmission cell signal 27 output from the transmission unit is input to the transmission line input terminal C as a reception cell signal 28 via the transmission line. The reception cell signal 28 is input to the separation circuit 118 and separated into a reception cell coded signal 29, a reception optimum interpolation circuit selection signal 30 and a reception cell number 31. Here, the reception cell coded signal 29 corresponds to the transmission cell coded signal 12, the reception optimum interpolation circuit selection signal 30 corresponds to the transmission optimum interpolation circuit selection signal 25, and the reception cell number 31 corresponds to the transmission cell number 26. To do. The received cell coded signal 29 is input to the cell decomposition circuit 119 and the reception interpolation circuits 120 to 123. The reception optimum interpolation circuit selection signal 30 is
4 is input. The received cell number 31 is input to the cell discard detection circuit 127.

The cell disassembling circuit 119 performs cell disassembly of the received received cell coded signal 29, outputs a received coded speech signal 36, and inputs it to the selection circuit 125. On the other hand, the reception interpolation circuits 120 to 123 interpolate the input received cell coded signal 29 by different methods and input the result to the selection circuit 124.

The reception interpolation circuit 120 interpolates the input received cell coded signal 29 of one cell with a silence signal, and outputs it as a reception interpolation signal 32 to the selection circuit 124. The reception interpolation circuit 121 performs 0-order hold interpolation on the last byte signal of the immediately preceding cell of the input reception cell coded signal 29 of one cell, and outputs the result to the selection circuit 124 as a reception interpolation signal 33. The reception interpolation circuit 122 performs primary interpolation on the input received cell coded signal 29 of one cell from the signal level of the last byte of the previous cell and the signal level of the first byte of the next cell, Selection circuit 1 as reception interpolation signal 34
To 24. The reception interpolation circuit 123 receives the input 1
The received cell coded signal 29 of the cell is interpolated by all bytes of the immediately preceding cell, and output to the selection circuit 124 as a received interpolation signal 35.

The selection circuit 124 is a circuit for performing interpolation in the same manner as the interpolation circuit that outputs the minimum cumulative interpolation error detected by the transmission unit, based on the input optimum reception interpolation circuit selection signal 30. , And outputs the received interpolation signal output by the selected reception interpolation circuit to the selection circuit 125 as the reception optimum interpolation signal 37.

The cell discard detection circuit 127 receives the received cell number 3
By checking the continuity of 1, the cell discard on the transmission path is detected. When the reception cell numbers 31 are continuous, the cell discard is not performed, and when the reception cell numbers 31 are discontinuous, it is determined that the cell discard is performed, and the determination result is output as the cell discard determination signal 40. The output cell discard determination signal 40 is input to the selection circuit 125.

The selection circuit 125 selects the reception optimum interpolation signal 37 when it is determined that the cell is discarded according to the input cell discard determination signal 40, and selects the reception coded voice signal 36 when the cell is not discarded. Then, it is output as a reception selection coded audio signal 38. The reception selection coded audio signal 38 output from the selection circuit 125 is the D / A converter 12
6 is input. The D / A converter 126 performs digital / analog conversion of the input reception selection coded audio signal 38, and outputs it as a reception audio signal 39 from the audio output terminal D.

In this embodiment, four types of interpolation circuits are used to simplify the description, and four types of interpolation methods, namely, silent interpolation, zero-order hold interpolation, primary interpolation, and interpolation by repeating the previous cell, will be described. However, the present invention is not limited to these numerical values and interpolation methods.

[0025]

As described above, according to the present invention, a plurality of transmission interpolation means and a plurality of reception interpolation means have the same kind of interpolation method, and the first selection means selects the transmission interpolation means selected by the comparison means. By selecting a reception interpolation unit having the same interpolation method as that of the interpolation unit and outputting a reception optimal interpolation signal, the reception unit can execute the same type of interpolation method as the optimal interpolation method selected by the transmission unit. Cells that are selected and discarded on the transmission path can be interpolated by the optimal interpolation method, and the optimal interpolated signal can be obtained for each cell, minimizing the quality degradation of the received voice signal due to missing information when the cells are discarded. It has the effect that it can be suppressed.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

[Explanation of symbols]

A input terminal B transmission line output terminal C transmission line input terminal D output terminal 10 transmission audio signal 11 transmission encoded audio signal 12 transmission cell encoded signal 13, 14, 15, 16 transmission interpolation signal 17, 18, 19, 20 interpolation Error 21, 22, 23, 24 Cumulative interpolation error 25 Transmission optimal interpolation circuit selection signal 26 Transmission cell number 27 Transmission cell signal 28 Reception cell signal 29 Reception cell encoding signal 30 Reception optimal interpolation circuit selection signal 31 Reception cell number 32, 33 , 34, 35 Reception interpolation signal 36 Reception coded speech signal 37 Reception optimal interpolation signal 38 Reception selection coded speech signal 39 Reception speech signal 40 Cell discard determination signal 101 A / D converter 102 Celling circuit 103, 104, 105, 106 Transmission interpolation circuit 107, 108, 109, 110 Subtractor 111, 112, 113, 11 Accumulation circuit 115 comparator circuit 116 cell counter 117 multiplexing circuit 118 separation circuit 119 cell disassembling circuits 120, 121, 122, 123 receives the interpolation circuit 124, 125 selection circuit 126 D / A converter 127 cell discard detecting circuit

Claims (4)

[Claims] 1. An audio cell coding apparatus for converting an audio signal into cells, transmitting the cells from a transmission unit to a reception unit, and discarding cells on a transmission path, wherein the transmission unit interpolates cells and outputs a transmission interpolation signal. A plurality of transmission interpolation means having different types of interpolation methods; a subtraction means for calculating an interpolation error between the transmission interpolation signal and the transmission audio signal; and accumulating the interpolation error by a capacity constituting one cell. Accumulating means for outputting a cumulative interpolation error, comparing the cumulative interpolation error for each interpolation method, selecting one of the interpolation methods, selecting the transmission interpolation means having the selected interpolation method, Comparison means for generating a transmission optimum interpolation means selection signal; multiplexing the transmission sound signal and the transmission optimum interpolation means selection signal of a cell immediately before the transmission sound signal to generate a transmission cell signal; Do A multiplexing unit, wherein the receiving unit inputs the transmission cell signal as a reception cell signal, and selects a reception audio signal corresponding to the transmission audio signal and a reception optimum interpolation unit corresponding to the transmission optimum interpolation unit selection signal. A plurality of reception interpolation means having different types of interpolation methods for interpolating a cell and outputting a reception interpolation signal; and a reception optimum interpolation means selection signal among the reception interpolation signals. First selecting means for selecting one and outputting it as a reception optimum interpolation signal; cell discard detection means for detecting cell discard; and selecting one of the reception voice signal and the reception optimum interpolation signal. A plurality of transmission interpolation means and a plurality of reception interpolation means having the same type of interpolation method. Place 2. The voice cell coding apparatus according to claim 1, wherein the comparison means selects the interpolation method that outputs the minimum value of the cumulative interpolation error for each cell. 3. The first selection means selects the reception interpolation means having the same interpolation method as that of the transmission interpolation means selected by the comparison means, and outputs the reception optimum interpolation signal. The voice cell coding device according to claim 1 or 2. 4. The second selection means outputs the reception optimal interpolation signal when a cell is discarded on a transmission path, and outputs the reception audio signal when no cell is discarded on the transmission path. 4. The speech cell encoding device according to claim 1, wherein