JP4572755B2 - Decoding device, decoding method, and digital audio communication system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce noise generation caused by transmission error without adding an arithmetic circuit such as an interpolation circuit into a decoder regardless of the number of bits of the transmission error when compressing and transmitting audio data in an ADPCM form. <P>SOLUTION: A decoder 6 of an ADPCM scheme comprises switching means 7, 8, 10, 11 for switching and selecting either data of a current sampling term or preceding data by one sampling term as data to be used for decoding and if any error is detected in inputted compression data, preceding data by one sampling term are selected by the switching means 7, 8, 10, 11. <P>COPYRIGHT: (C)2007,JPO&amp;INPIT

Description

  The present invention relates to a decoding apparatus, a decoding method, and a digital audio communication system that employ an ADPCM method.

  ITU-T (International Telecommunication Union) is one of the audio data compression methods. There is an ADPCM (Adaptive Difference PCM) system defined as a 726 recommendation. The ADPCM method is employed for compressing audio data to be transmitted in a digital audio communication system such as a digital wireless microphone system.

  FIG. FIG. 7 is a block diagram illustrating a configuration of an encoder and a decoder that comply with the H.726 recommendation.

  The encoder 21 includes an adder 22, an adaptive quantizer 23, a scale factor updater 24, an adaptive inverse quantizer 25, and a predictor 26. From the PCM sound data X sampled by an AD converter (not shown), the adder 22 subtracts the prediction data P of the predictor 26 based on the sound data one sampling period before.

  The difference data D obtained by the adder 22 is quantized by the adaptive quantizer 23 with the scale factor (quantization width) S from the scale factor updater 24. The scale factor updater 24 updates the scale factor S based on the scale factor S calculated in the current sampling period and the quantized data Q from the adaptive quantizer 23 (the scale factor used in the next sampling period). S is calculated).

  The adaptive inverse quantizer 25 inversely quantizes the upper bits of the quantized data Q of the adaptive quantizer 23 with the scale factor (quantization width) S from the scale factor updater 24. The predictor 26 predicts the audio data of the next sampling period based on the dequantized data D. The prediction data P is subtracted from the audio data X of the next sampling period by the adder 22 as described above.

  From the encoder 21, the quantized data Q of the adaptive quantizer 23 is output as compressed data.

  The decoder 31 includes an adaptive inverse quantizer 32, a scale factor updater 33, a predictor 34, and an adder 35. The scale factor updater 33 updates the scale factor S ′ based on the scale factor S ′ calculated in the current sampling period and the compressed data Q input to the decoder 31 (the scale used in the next sampling period). Factor S ′ is calculated). The scale factor S ′ is data corresponding to the scale factor S on the encoder 21 side.

  In the adaptive inverse quantizer 32, the compressed data Q input to the decoder 31 is inversely quantized with the scale factor S ′ from the scale factor updater 33. Thereby, difference data D ′ corresponding to the difference data D of the adder 22 in the encoder 21 is obtained.

  The predictor 34 predicts the audio data of the next sampling period based on the difference data D ′. Then, the difference data D ′ and the prediction data P ′ of the predictor 34 based on the difference data D ′ one sampling period before are added by the adder 35, whereby the PCM audio data X ′ is decoded. .

  By the way, in the encoder 21 and decoder 31 having the configuration shown in FIG. 5, between the transmission side (side where the encoder 21 is provided) and the reception side (side where the decoder 31 is provided) of speech data. When a data transmission error occurs in the transmission path, audio data different from the original PCM audio data X is decoded by the decoder 31. Therefore, noise is generated when the decoded voice is played back.

  FIG. 6 is a diagram showing a state of decoding in the decoder 31 when such a transmission error occurs. Due to the occurrence of a transmission error in the transmission path, the compressed data Qe having an error is sent to the decoder 31. In the decoder 31, based on the erroneous compressed data Qe, difference data D ′, prediction data P, and scale factor S on the encoder 21 side are obtained as difference data D ′, prediction data P ′, and scale factor S ′. Different difference data De, prediction data Pe, and scale factor Se are obtained. As a result, the PCM audio data X ′ to be decoded becomes data Xe different from the original PCM audio data X. Therefore, noise is generated when the data Xe is reproduced. This noise is significant depending on the erroneous bit position in the compressed data.

  Since the ADPCM method obtains the scale factor and the prediction data based on the data of the previous sampling period, if there is a transmission error in the compressed data of a certain sampling period, the scale factor and the prediction data become normal values. A period of several sampling cycles or more is required before convergence. As a result, noise is generated over a long period.

  Conventionally, there have been the following methods (a) and (b) as methods for reducing the occurrence of noise due to such transmission errors in the ADPCM method.

(A) An error correction code is added to the compressed data on the transmission side, and an error in the compressed data is corrected using the error correction code on the reception side.
(B) An interpolation circuit is added in the decoder to interpolate the average value of data errors (see, for example, Patent Document 1).

JP 62-185414 A (6th page, FIG. 1)

However, in the method (a), when the number of bits of transmission error is large, even if an error correction code is used, the error of the compressed data cannot be corrected, and therefore the generation of noise may not be reduced.
In the method (b), an interpolation circuit including an arithmetic circuit such as an adder or a divider is added to the decoder, which complicates the configuration of the decoder.

  In view of the above, the present invention adds an arithmetic circuit such as an interpolation circuit in the decoder regardless of the number of bits of transmission errors when audio data is compressed and transmitted by the ADPCM method. Thus, it is an object to reduce the generation of noise due to transmission errors.

In order to solve this problem, the present invention provides:
A scale factor updater that updates the scale factor based on the supplied compressed data and the calculated scale factor;
An adaptive inverse quantizer that obtains difference data by inversely quantizing the supplied compressed data with the scale factor from the scale factor updater;
A predictor that predicts speech data of the next sampling period based on the difference data;
An adder for decoding PCM audio data by adding the difference data and the audio data predicted by the predictor;
In an ADPCM decoder having
When no error is detected by the error detection decoder based on a signal indicating an error detection result from an error detection decoder provided in a preceding stage of the decoder, the compressed data from the error detection decoder is When the error is detected by the error detection decoder, the compressed data from the error detection decoder is supplied to the adaptive inverse quantizer and the scale factor updater. A first changeover switch not supplied to the scale factor updater;
A first delay unit that delays the scale factor from the scale factor updater by a time corresponding to one sampling period;
Based on a signal indicating an error detection result from the error detection decoder, if no error is detected by the error detection decoder, a scale factor from the scale factor updater is selected and output, and the error detection decoding A second selector switch that selects and outputs a scale factor one sampling period delayed by the first delay unit when an error is detected by the generator;
A second delayer for delaying the difference data from the adaptive inverse quantizer by a time corresponding to one sampling period;
Based on the signal indicating the error detection result from the error detection decoder, if no error is detected by the error detection decoder, the differential data from the adaptive inverse quantizer is selected and output, and the error detection A third changeover switch for selecting and outputting the difference data one sampling period before delayed by the second delay unit when an error is detected by the decoder;
The scale factor is updated based on the compressed data supplied from the first changeover switch and the scale factor output from the second changeover switch,
The adaptive inverse quantizer obtains difference data by inversely quantizing the compressed data supplied from the first changeover switch with a scale factor from the scale factor updater;
The predictor predicts speech data of the next sampling period based on the difference data output from the third changeover switch,
The adder decodes PCM voice data by adding the difference data output from the third changeover switch and the voice data predicted by the predictor .

The present invention also provides
A scale factor updater that updates the scale factor based on the supplied compressed data and the calculated scale factor;
An adaptive inverse quantizer that obtains difference data by inversely quantizing the supplied compressed data with the scale factor from the scale factor updater;
A predictor that predicts speech data of the next sampling period based on the difference data;
An adder for decoding PCM audio data by adding the difference data and the audio data predicted by the predictor;
In a decoding method in an ADPCM decoder having
Based on a signal indicating an error detection result from an error detection decoder provided in a preceding stage of the decoder, when no error is detected by the error detection decoder,
Supplying compressed data from the error detection decoder to the adaptive inverse quantizer and the scale factor updater;
Causing the scale factor updater to update the scale factor based on the scale factor calculated in the current sampling period;
Based on the difference data from the adaptive inverse quantizer, the predictor predicts speech data of the next sampling period,
Causing the adder to decode the PCM audio data by adding the difference data from the adaptive inverse quantizer and the audio data predicted by the predictor;
If an error is detected by the error detection decoder,
Without supplying compressed data from the error detection decoder to the adaptive inverse quantizer and the scale factor updater;
Causing the scale factor updater to update the scale factor based on the scale factor calculated one sampling period ago;
Causing the predictor to predict speech data of the next sampling period based on the difference data of the previous sampling period from the adaptive inverse quantizer;
The adder causes the PCM speech data to be decoded by adding the difference data one sampling period before from the adaptive inverse quantizer and the speech data predicted by the predictor .

Further, the present invention provides a digital audio communication system in which an ADPCM encoder is provided on the audio data transmission side and an ADPCM decoder is provided on the audio data reception side. An error detection code is added to the generated compressed data and transmitted to the reception side. Based on this error detection code, the reception side transmits the data from the transmission side using an error detection decoder provided in the preceding stage of the decoder. The decoder performs error detection on the compressed data, and this decoder
A scale factor updater that updates the scale factor based on the supplied compressed data and the calculated scale factor;
An adaptive inverse quantizer that obtains difference data by inversely quantizing the supplied compressed data with the scale factor from the scale factor updater;
A predictor that predicts speech data of the next sampling period based on the difference data;
An adder for decoding PCM audio data by adding the difference data and the audio data predicted by the predictor;
When no error is detected by the error detection decoder based on a signal indicating an error detection result from the error detection decoder, the adaptive inverse quantizer and the adaptive inverse quantizer and the compressed data from the error detection decoder are When an error is detected by the error detection decoder supplied to the scale factor updater, the compressed data from the error detection decoder is not supplied to the adaptive inverse quantizer and the scale factor updater. 1 selector switch,
A first delay unit that delays the scale factor from the scale factor updater by a time corresponding to one sampling period;
Based on a signal indicating an error detection result from the error detection decoder, if no error is detected by the error detection decoder, a scale factor from the scale factor updater is selected and output, and the error detection decoding A second selector switch that selects and outputs a scale factor one sampling period delayed by the first delay unit when an error is detected by the generator;
A second delayer for delaying the difference data from the adaptive inverse quantizer by a time corresponding to one sampling period;
Based on the signal indicating the error detection result from the error detection decoder, if no error is detected by the error detection decoder, the differential data from the adaptive inverse quantizer is selected and output, and the error detection A third changeover switch for selecting and outputting the difference data one sampling period before delayed by the second delay unit when an error is detected by the decoder;
The scale factor is updated based on the compressed data supplied from the first changeover switch and the scale factor output from the second changeover switch,
The adaptive inverse quantizer obtains difference data by inversely quantizing the compressed data supplied from the first changeover switch with a scale factor from the scale factor updater;
The predictor predicts speech data of the next sampling period based on the difference data output from the third changeover switch,
The adder decodes PCM voice data by adding the difference data output from the third changeover switch and the voice data predicted by the predictor .

According to the present invention, when an error is detected in the compressed data that is transmitted and input to the decoder, the scale factor updated based on the scale factor before one sampling period, and the difference data before one sampling period, The decoding is performed using the prediction data obtained from the difference data . Thereby, since decoding can be performed based on the compressed data before a transmission error occurs, the generation of noise due to the transmission error can be reduced.

Further, the scale factor calculated based on the scale factor of one sampling period is the scale factor calculated based on the scale factor of the current sampling period when there is no error in the compressed data of the current sampling period. Is not so different. Therefore, after the transmission error is eliminated, the scale factor can be converged to a normal value at an early stage.
In addition, since the predictor obtains the prediction data at every sampling period regardless of whether there is an error in the compressed data, the value of the prediction data does not greatly change.

  Unlike the case of correcting the error of the compressed data, even if the number of bits of the transmission error is large, the generation of noise due to the transmission error can be reduced. Further, it is sufficient to add a simple circuit such as a delay unit and a changeover switch for one sampling period in the decoder, and it is not necessary to add an arithmetic circuit such as an interpolation circuit.

  According to the present invention, when audio data is compressed by the ADPCM method and transmitted, the transmission is performed without adding an arithmetic circuit such as an interpolation circuit in the decoder regardless of the number of transmission error bits. The effect that the generation of noise due to errors can be reduced is obtained.

  Hereinafter, an example in which the present invention is applied to a digital wireless microphone system for broadcasting business will be specifically described with reference to the drawings.

  FIG. 1 is a diagram showing a digital wireless microphone system to which the present invention is applied. This system transmits digital audio data from a wireless microphone transmitter (hereinafter referred to as a transmitter) 1 to a wireless microphone receiver (hereinafter referred to as a receiver) 2 by radio waves.

  The transmitter 1 is used by the interviewer at the interview site. The receiver 2 is attached to, for example, a VTR-integrated video camera 3 used by a cameraman, and audio data is recorded on the video camera 3.

  The transmitter 1 and the receiver 2 are provided with an ADPCM encoder and decoder, respectively. FIG. 2 is a block diagram showing the configuration of the encoding system and decoding system of the transmitter 1 and receiver 2 including the encoder and decoder, and the same reference numerals are given to the parts common to FIG. ing.

  The encoding system of the transmitter 1 is provided with an encoder 21 and an error detection encoder 4. The encoder 21 has the same configuration as that shown in FIG. 5, and includes an adder 22, an adaptive quantizer 23, a scale factor updater 24, an adaptive inverse quantizer 25, and a predictor 26. And have. Predicted data P of the predictor 26 based on the audio data of one sampling period before from the PCM audio data X output from the microphone (not shown) in the transmitter 1 and sampled by the AD converter (not shown). Is subtracted.

  The difference data D obtained by the adder 22 is quantized by the adaptive quantizer 23 with the scale factor (quantization width) S from the scale factor updater 24. The scale factor updater 24 updates the scale factor S based on the scale factor S calculated in the current sampling period and the quantized data Q from the adaptive quantizer 23 (the scale factor used in the next sampling period). S is calculated).

  The adaptive inverse quantizer 25 inversely quantizes the upper bits of the quantized data Q of the adaptive quantizer 23 with the scale factor (quantization width) S from the scale factor updater 24. The predictor 26 predicts the audio data of the next sampling period based on the dequantized data D. The prediction data P is subtracted from the audio data X of the next sampling period by the adder 22 as described above.

  From the encoder 21, the quantized data Q of the adaptive quantizer 23 is sent to the error detection encoder 4 as compressed data. The error detection encoder 4 is a circuit that adds an error detection code such as a Hamming code to input data.

  The compressed data Q to which the error detection code T is added by the error detection encoder 4 is converted into a radio wave signal by a modulation / transmission system (not shown) in the transmitter 1 and transmitted to the receiver 2.

  The decoding system of the receiver 2 is provided with an error detection decoder 5 and a decoder 6. The compressed data Q with the error detection code T received and demodulated by a reception / demodulation system (not shown) in the receiver 2 is sent to the error detection decoder 5. The error detection decoder 5 is a circuit that detects an error in the data based on the error detection code added to the input data.

  The error detection decoder 5 outputs compressed data Q and a 1-bit signal R indicating an error detection result (a signal that is “0” when no error is detected and “1” when an error is detected). And sent to the decoder 6.

  The decoder 6 includes an adaptive inverse quantizer 32, a scale factor updater 33, a predictor 34, an adder 35, delay units 7 and 8, a 1-input 2-output change-over switch 9, and 2-inputs. 1-output selector switches 10 to 12 and a counter 13 are provided. Among them, the adaptive inverse quantizer 32, the scale factor updater 33, the predictor 34, and the adder 35 are the adaptive inverse quantizer 32, the scale factor updater 33, the predictor of the decoder 31 shown in FIG. 34 and the adder 35. The delay units 7 and 8 are circuits that delay input data by a time corresponding to one sampling period of the PCM audio data X.

  The compressed data Q sent from the error detection decoder 5 is input to the changeover switch 9. One output terminal 9a of the changeover switch 9 is connected to the adaptive inverse quantizer 32 and the scale factor updater 33, and the other output terminal 9b of the changeover switch 9 is grounded.

  The changeover switch 9 uses the signal R from the error detection decoder 5 as a control signal and outputs compressed data Q from the output terminal 9a when R = '0', and outputs when R = '1'. The compressed data Q is output from the end 9b. Therefore, the adaptive inverse quantizer 32 and the scale factor updater 33 are not supplied with the compressed data Q in which an error has been detected by the error detection decoder 5, but only with the compressed data Q in which no error has been detected. The

  The scale factor updater 33 updates the scale factor S ′ based on the scale factor S ′ output from the changeover switch 10 and the compressed data Q supplied from the changeover switch 9 (the scale used in the next sampling cycle). Factor S ′ is calculated).

  To the changeover switch 10, the scale factor S 'from the scale factor updater 33 is sent to one input terminal 10a as it is, and this scale factor S' is sent to the other input terminal 10b via the delay unit 7. . The changeover switch 10 selects and outputs the scale factor S ′ input to the input terminal 10a when R = “0”, using the signal R from the error detection decoder 5 as a control signal, and R = “1”. At this time, the scale factor S ′ (scale factor S ′ one sampling period before) input to the input terminal 10b is selected and output.

  In the adaptive inverse quantizer 32, the difference data D ′ is obtained by inversely quantizing the compressed data Q from the changeover switch 9 with the scale factor S ′ from the scale factor updater 33.

  The difference data D ′ obtained by the adaptive inverse quantizer 32 is sent to one input terminal 11a of the changeover switch 11 as it is and also sent to the other input terminal 11b of the changeover switch 11 via the delay unit 8. . The changeover switch 11 selects and outputs the difference data D ′ input to the input terminal 11a when R = “0”, using the signal R from the error detection decoder 5 as a control signal, and R = “1”. At this time, the difference data D ′ (difference data D ′ one sampling period before) input to the input terminal 11b is selected and output.

  The predictor 34 predicts audio data of the next sampling period based on the difference data D ′ output from the changeover switch 11. The prediction data P ′ of the predictor 34 and the difference data D ′ output from the changeover switch 11 are added by the adder 35, whereby the PCM speech data X ′ is decoded.

  The decoded PCM audio data X ′ is sent to one input terminal 12 a of the changeover switch 12. The other input terminal 12b of the changeover switch 12 receives all "0" data having the same number of bits as the PCM audio data X '. When the pulse having the value “1” is not supplied from the counter 13, the changeover switch 12 selects and outputs the PCM sound data X input to the input terminal 12 a, and the pulse having the value “1” is supplied from the counter 13. When selected, all “0” data input to the input terminal 12b is selected and output.

  The counter 13 counts the pulse of the value “1” in the signal R from the error detection decoder 5 and outputs the pulse of the value “1” when the count value becomes N (for example, about 5) or more to switch. Supply to switch 12. Further, the counter 13 resets the count value to zero when the pulse of the value “1” does not exist in the signal R for a period exceeding the sampling period of the PCM audio data X.

  Data selected by the changeover switch 12 is output from the decoder 6, input from an audio output terminal (not shown) of the receiver 2 to, for example, the video camera 3 of FIG. 1, and recorded in the video camera 3.

  Next, the manner in which the audio data is decoded by the decoder 6 in FIG. 2 will be described with respect to the case where no transmission error of the compressed data Q occurs in the transmission path between the transmitter 1 and the receiver 2. This will be described with reference to FIGS. 3 and 4 separately from the case where a transmission error of the compressed data Q occurs.

[When no transmission error occurs]
As shown in FIG. 3, when a transmission error does not occur, the error detection decoder 5 in the receiver 2 detects no error, so that the signal R = '0'. Therefore, the compressed data Q is supplied from the changeover switch 9 (FIG. 2) to the adaptive inverse quantizer 32 and the scale factor updater 33. The changeover switches 10 and 11 (FIG. 2) select and output the input data to the input terminals 10a and 11a, respectively. Since the counter 13 (FIG. 2) is reset, the changeover switch 12 (FIG. 2) selects the PCM audio data X ′ to be input to the input terminal 12a.

  At this time, in the scale factor updater 33 (FIG. 2), based on the scale factor S ′ of the current sampling period, the scale factor S ′ used in the next sampling period (based on the data of the current sampling period t). Is expressed as a scale factor S ′ (t)).

  The adaptive inverse quantizer 32 obtains difference data D ′ (t) obtained by inversely quantizing the compressed data Q with the scale factor S ′ (t).

  In the predictor 34 (FIG. 2), predicted data P ′ (t) of the next sampling period is obtained based on the difference data D ′ (t).

  Then, PCM speech data X ′ obtained by adding the prediction data P (t) ′ and the difference data D (t) ′ is output from the decoder 6.

  In this way, if no transmission error occurs, the decoder 6 updates the scale factor S ′ (t) updated based on the scale factor S ′ of the current sampling period and the scale factor S ′. The PCM audio data X ′ is decoded using the difference data D ′ (t) obtained using (t) and the prediction data P ′ (t) obtained from the difference data D ′ (t). That is, in this case, the decoder 6 performs decoding by the same operation as that of the decoder 31 shown in FIG.

[When a transmission error occurs]
As shown in FIG. 4, when a transmission error occurs, an error is detected by the error detection decoder 5 in the receiver 2, so that the signal R = '1'. Therefore, the compressed data Q is not supplied from the changeover switch 9 (FIG. 2) to the adaptive inverse quantizer 32 and the scale factor updater 33. The changeover switches 10 and 11 (FIG. 2) select and output the input data to the input terminals 10b and 11b, respectively. Further, until the count value of the counter 13 (FIG. 2) reaches N, the changeover switch 12 (FIG. 2) selects the PCM audio data X ′ input to the input terminal 12a.

  At this time, in the scale factor updater 33 (FIG. 2), the scale factor S ′ used in the next sampling cycle (the previous sampling cycle (t−1) is used based on the scale factor S ′ in the previous sampling cycle. ) Is expressed as a scale factor S ′ (t−1) in the sense of being based on data. As a result, the scale factor updater 33 calculates the scale factor S ′ having the same value as that calculated one sampling period before (the value of the scale factor S ′ is not updated).

  In the adaptive inverse quantizer 32, differential data D 'obtained by inversely quantizing the compressed data Q with the scale factor S' (t-1) is obtained.

  The predictor 34 (FIG. 2) obtains predicted data P ′ (t−1) for the next sampling period based on the difference data D ′ (t−1) from the adaptive inverse quantizer 32 one sampling period before. It is done.

  Then, PCM audio data X ′ obtained by adding the prediction data P (t−1) ′ and the difference data D (t−1) ′ is output from the decoder 6.

  In this manner, when a transmission error occurs, the decoder 6 updates the scale factor S ′ (t−1) updated based on the scale factor S ′ one sampling period before and the one sampling period before. The PCM audio data X ′ is decoded using the difference data D ′ (t−1) and the prediction data P ′ (t−1) obtained from the difference data D ′ (t−1).

  As a result, decoding can be performed based on the compressed data before the transmission error occurs, so that noise can be reduced when the audio recorded in the video camera 3 is reproduced.

  Further, the scale factor S ′ (t−1) calculated based on the scale factor S ′ one sampling period before is the scale of the current sampling period when there is no error in the compressed data Q of the current sampling period. It is not so different from the scale factor S calculated based on the factor S ′. Therefore, after the transmission error is eliminated, the scale factor S ′ can be converged to a normal value at an early stage.

  Further, since the predictor 34 obtains the prediction data P ′ at every sampling period regardless of whether or not the compressed data Q has an error, the value of the prediction data P ′ is not greatly changed.

  Unlike the case of correcting the error of the compressed data, even if the number of bits of the transmission error is large, the generation of noise due to the transmission error can be reduced. Further, it is sufficient to add a simple circuit such as a delay unit and a changeover switch for one sampling period in the decoder, and it is not necessary to add an arithmetic circuit such as an interpolation circuit.

  When a transmission error occurs continuously over N (for example, 5 as described above) over a sampling period, a pulse having a value “1” from the counter 13 (FIG. 2) is sent to the changeover switch 12 (FIG. 2). Since it is supplied, the changeover switch 12 (FIG. 2) selects all “0” data to be input to the input terminal 12b.

  As a result, silent audio data is output from the decoder 6 instead of the PCM audio data X ′. Even if decoding is performed using the data before one sampling period, the noise reduction effect is diminished if transmission errors occur continuously over a plurality of sampling periods. Thus, the PCM audio data X ′ itself By stopping the output, the generation of noise can be reliably suppressed.

  In the above example, the configuration of the decoder 6 has been described with reference to the hardware circuit diagram. However, the present invention is not limited to this, and the decoder 6 may be designed as an ASIC (Application Specific LSI) or a programmable logic device (FPGA or the like). Further, a computer program corresponding to the operation of the decoder 6 may be executed by the CPU.

  In the above example, the present invention is applied to a digital wireless microphone system. However, the present invention can be applied to all kinds of wireless or wired digital audio communication systems.

It is a figure which shows the digital wireless microphone system to which this invention is applied. It is a block diagram which shows the structure of the encoding system of a transmitter of FIG. 1, and a receiver, and a decoding system. It is a figure which shows the mode of the decoding in the decoder of FIG. 2 when there is no transmission error. It is a figure which shows the mode of the decoding in the decoder of FIG. 2 when there exists a transmission error. ITU-TG. FIG. 7 is a block diagram illustrating a configuration of an encoder and a decoder that comply with the H.726 recommendation. It is a figure which shows the mode of the decoding in the decoder of FIG. 5 when there exists a transmission error.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Wireless microphone transmitter, 2 Wireless microphone receiver, 3 Video camera, 4 Error detection encoder, 5 Error detection decoder, 6 Decoder, 7, 8 delay device, 9-12 selector switch, 13 counter, 21 code , 22 adder, 23 adaptive quantizer, 24 scale factor updater, 25 adaptive inverse quantizer, 26 predictor, 32 adaptive inverse quantizer, 33 scale factor updater, 34 predictor, 35 adder

Claims (6)

A scale factor updater that updates the scale factor based on the supplied compressed data and the calculated scale factor;
An adaptive inverse quantizer that obtains difference data by inversely quantizing the supplied compressed data with the scale factor from the scale factor updater;
A predictor that predicts speech data of the next sampling period based on the difference data;
An adder for decoding PCM audio data by adding the difference data and the audio data predicted by the predictor;
In an ADPCM decoder having
When no error is detected by the error detection decoder based on a signal indicating an error detection result from an error detection decoder provided in a preceding stage of the decoder, the compressed data from the error detection decoder is When the error is detected by the error detection decoder, the compressed data from the error detection decoder is supplied to the adaptive inverse quantizer and the scale factor updater. A first changeover switch not supplied to the scale factor updater;
A first delay unit that delays the scale factor from the scale factor updater by a time corresponding to one sampling period;
Based on a signal indicating an error detection result from the error detection decoder, if no error is detected by the error detection decoder, a scale factor from the scale factor updater is selected and output, and the error detection decoding A second selector switch that selects and outputs a scale factor one sampling period delayed by the first delay unit when an error is detected by the generator;
A second delayer for delaying the difference data from the adaptive inverse quantizer by a time corresponding to one sampling period;
Based on the signal indicating the error detection result from the error detection decoder, if no error is detected by the error detection decoder, the differential data from the adaptive inverse quantizer is selected and output, and the error detection A third changeover switch for selecting and outputting the difference data one sampling period before delayed by the second delay unit when an error is detected by the decoder;
The scale factor is updated based on the compressed data supplied from the first changeover switch and the scale factor output from the second changeover switch,
The adaptive inverse quantizer obtains difference data by inversely quantizing the compressed data supplied from the first changeover switch with a scale factor from the scale factor updater;
The predictor predicts speech data of the next sampling period based on the difference data output from the third changeover switch,
The adder is a decoder for decoding PCM audio data by adding difference data output from the third changeover switch and audio data predicted by the predictor . The decoder of claim 1, wherein
A determinator for determining whether or not errors are continuously detected by the error detection decoder over a predetermined number of sampling periods based on a signal indicating an error detection result from the error detection decoder; ,
If it is not determined by the determiner that errors have been continuously detected over a predetermined number of sampling periods, PCM audio data from the adder is output to the outside of the decoder, and A fourth switch that does not output the PCM audio data from the adder to the outside of the decoder when it is determined by the determiner that errors have been continuously detected over a predetermined number of sampling periods; And a switch
Decoder. A scale factor updater that updates the scale factor based on the supplied compressed data and the calculated scale factor;
An adaptive inverse quantizer that obtains difference data by inversely quantizing the supplied compressed data with the scale factor from the scale factor updater;
A predictor that predicts speech data of the next sampling period based on the difference data;
An adder for decoding PCM audio data by adding the difference data and the audio data predicted by the predictor;
In a decoding method in an ADPCM decoder having
Based on a signal indicating an error detection result from an error detection decoder provided in a preceding stage of the decoder, when no error is detected by the error detection decoder,
Supplying compressed data from the error detection decoder to the adaptive inverse quantizer and the scale factor updater;
Causing the scale factor updater to update the scale factor based on the scale factor calculated in the current sampling period;
Based on the difference data from the adaptive inverse quantizer, the predictor predicts speech data of the next sampling period,
Causing the adder to decode the PCM audio data by adding the difference data from the adaptive inverse quantizer and the audio data predicted by the predictor;
If an error is detected by the error detection decoder,
Without supplying compressed data from the error detection decoder to the adaptive inverse quantizer and the scale factor updater;
Causing the scale factor updater to update the scale factor based on the scale factor calculated one sampling period ago;
Causing the predictor to predict speech data of the next sampling period based on the difference data of the previous sampling period from the adaptive inverse quantizer;
Causing the adder to decode the PCM audio data by adding the difference data one sampling period before from the adaptive inverse quantizer and the audio data predicted by the predictor;
Decryption method. The decoding method according to claim 3, wherein
Based on a signal indicating an error detection result from the error detection decoder, if no error is continuously detected by the error detection decoder over a predetermined number of sampling periods, the signal from the adder PCM audio data is output to the outside of the decoder, and if errors are detected continuously over a predetermined number of sampling periods by the error detection decoder, the PCM audio data from the adder Is not output to the outside of the decoder
Decryption method. In a digital audio communication system in which an ADPCM encoder is provided on the audio data transmission side and an ADPCM decoder is provided on the audio data reception side,
From the transmission side, an error detection code is added to the compressed data generated by the encoder and transmitted to the reception side,
On the reception side, based on the error detection code, an error detection decoder provided in the preceding stage of the decoder performs error detection of the compressed data transmitted from the transmission side,
The decoder is
A scale factor updater that updates the scale factor based on the supplied compressed data and the calculated scale factor;
An adaptive inverse quantizer that obtains difference data by inversely quantizing the supplied compressed data with the scale factor from the scale factor updater;
A predictor that predicts speech data of the next sampling period based on the difference data;
An adder for decoding PCM audio data by adding the difference data and the audio data predicted by the predictor;
When no error is detected by the error detection decoder based on a signal indicating an error detection result from the error detection decoder, the adaptive inverse quantizer and the adaptive inverse quantizer and the compressed data from the error detection decoder are When an error is detected by the error detection decoder supplied to the scale factor updater, the compressed data from the error detection decoder is not supplied to the adaptive inverse quantizer and the scale factor updater. 1 selector switch,
A first delay unit that delays the scale factor from the scale factor updater by a time corresponding to one sampling period;
Based on a signal indicating an error detection result from the error detection decoder, if no error is detected by the error detection decoder, a scale factor from the scale factor updater is selected and output, and the error detection decoding A second selector switch that selects and outputs a scale factor one sampling period delayed by the first delay unit when an error is detected by the generator;
A second delayer for delaying the difference data from the adaptive inverse quantizer by a time corresponding to one sampling period;
Based on a signal indicating an error detection result from the error detection decoder, if no error is detected by the error detection decoder, the differential data from the adaptive inverse quantizer is selected and output, and the error detection A third changeover switch for selecting and outputting the difference data one sampling period before delayed by the second delay unit when an error is detected by the decoder;
The scale factor is updated based on the compressed data supplied from the first changeover switch and the scale factor output from the second changeover switch,
The adaptive inverse quantizer obtains difference data by inversely quantizing the compressed data supplied from the first changeover switch with a scale factor from the scale factor updater;
The predictor predicts speech data of the next sampling period based on the difference data output from the third changeover switch,
The adder decodes PCM voice data by adding the difference data output from the third changeover switch and the voice data predicted by the predictor.
Digital voice communication system. The digital audio communication system according to claim 5, wherein
The decoder is
A determinator for determining whether or not errors are continuously detected by the error detection decoder over a predetermined number of sampling periods based on a signal indicating an error detection result from the error detection decoder; ,
If it is not determined by the determiner that errors have been continuously detected over a predetermined number of sampling periods, PCM audio data from the adder is output to the outside of the decoder, and A fourth switch that does not output the PCM audio data from the adder to the outside of the decoder when it is determined by the determiner that errors have been continuously detected over a predetermined number of sampling periods; And a switch
Digital voice communication system .

Images