JPH0758706A - Speech encoder and speech encoding method - Google Patents

Abstract

PURPOSE:To provide a speech encoder which performs speech encoding with low power consumption. CONSTITUTION:This encoder is equipped with a memory access means 4 which inputs and outputs digital code data 18 and 19 or digital speech data 12 and 15 and is interposed for an internal processor 2 to send and receive the digital code data 18 and 19 or digital speech data 12 and 15 by accessing a memory 3 for speech encoding and decoding in the processor 2, a means which stops the processor 2 from operating when detecting no speech being included in input speech data for encoding.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech coder for encoding and decoding speech used in digital wireless communication and the like, and more particularly to a speech coder suitable for low power consumption.

[0002]

2. Description of the Related Art "VOICEACTIVITY DETECTIO is an example of a process for reducing power consumption when there is no sound in digital wireless communication.
N (VAD) AND THE OPERATION OF DISCONTINUOUS TRANSMIS
SION (DTX) IN THE GSM SYSTEM "(S.Hansen: Digital Cel
There is a method to reduce the power consumption by stopping the radio wave transmission of the radio section as described in lular Radio Conference '1988). This is a method of reducing power consumption by minimizing wireless transmission, which consumes the largest amount of power in digital wireless communication. When the power consumption of other circuits is small, the power consumption can be reduced easily and efficiently. It has the advantage of being worn.

[0003]

In the above-mentioned conventional example, only the method of stopping the wireless transmission section is disclosed, and the method of stopping the operation of the speech encoder for encoding speech is not disclosed. However, in the highly-compressed voice coding adopted in recent digital mobile communications, since the processing amount of the voice coding is increasing, the operating speed of the built-in processor is high and the power consumption is large. Therefore, in order to reduce the power consumption of the entire digital mobile communication device, it is important to reduce the power consumption of the speech coder. An object of the present invention is to provide a speech coder that can realize speech coding with low power consumption.

[0004]

In the speech coder of the present invention for achieving the above object, for example, as shown in FIG. 1, input / output of digital code data 18, 19 or digital speech data 12, 15 By interposing with the built-in processor 2, the speech coding / decoding memory 3 included in the processor 2 is accessed and digital code data 18, 19 or digital speech data 12, is sent to the memory 3.
15 and the memory access means 4 for transmitting and receiving, for example, FIG.
Alternatively, as shown in FIG. 6, the operation of the processor is stopped (27) at the time of silence determination (32) in which it is detected that the input voice data for encoding does not include voice.
Means for doing so.

Here, the memory access means 4 sends and receives digital code data or digital voice data to and from the voice coding / decoding memory 3 even if the processor is not operating. The writing or reading of the data from the memory is performed.

Further, in the voice encoding method of the present invention for achieving the above object, for example, as shown in FIGS. 3 and 6, memory access for transmitting / receiving the digital code data or digital voice data to / from a memory is performed. Means is provided, and between the memory access means, the memory, and the processor, the following steps are included to start encoding by the operating processor or shift to stop of the processor. . That is, (1) the memory access means receives (41) the audio data, writes the data in the memory (42), confirms the number of data, and completes the acquisition of the required number of audio data for encoding (43 or 71), (2) Next, the processor determines whether there is sound or no sound (26), and if there is sound, the encoding process is started (4).
6) If there is no sound, stop the operation of the processor (2
7), (3) The memory access means fetches the next audio data.

As for the start of decoding, for example, FIG.
As shown in, the following procedure may be included. That is, (1) the memory access means receives the digital code data for decoding (52), writes the data in the memory (53), (2) determines whether the processor is operating next (54), and operates. If yes, start decoding (5
6) and (3) If the processor is stopped, the processor is activated (55) and decoding is started (56).

Alternatively, the decoding interrupt and the restoration process may include the following procedure as shown in FIG. 5, for example. That is, (1) when the decoding data is received even when the processor is executing the encoding process, the encoding process is interrupted and the decoding process is performed (61), (62), (2). Check the processing status (6
3) Then, depending on the state, the state returns to the encoding processing state or the operation stop state (37), (38).

Further, regarding the transition from the stopped state to the encoded state, for example, as shown in FIG. 7, (1) the memory access means receives voice data, writes the data in the memory, The number of data is confirmed and the acquisition of the required number of audio data for encoding is completed (71), (2) the memory access means activates the processor (21), (3). The sound / silence determination is performed (26), and if it is determined that there is sound, the voice coding state is set (28).

[0010]

In the voice encoder, by providing the memory access means separately from the processor for encoding and decoding the voice, even if the processor is not operating, the voice data for encoding is taken in and decoded. The resulting voice data can be output. Therefore, when the number of voice data required for one-time encoding taken in does not include voice, the operation of the above processor can be stopped.
This makes it possible to reduce the power consumption of the speech encoder.

[0011]

EXAMPLES Examples of the present invention will be described below. First, FIG. 1 shows the configuration of a digital wireless communication apparatus equipped with a speech coder to which the present invention is applied. The apparatus includes a voice encoder 1 for encoding and decoding voice, a microphone 10 for converting voice into an analog input voice signal 11 and taking it in, and an analog input voice signal 11 for digital input voice data 12. An A / D converter 7 for converting, a D / A converter 6 for converting the decoded digital output audio data 15 into an analog output audio signal 16, and a speaker 9 for outputting the output audio signal 16. Transmission data 19 of the encoded result from the voice encoder 1.
Is formed from the antenna 8 as a received transmission signal 20 and the received signal 17 received by the antenna 8 is passed to the speech encoder 1 as decoding data 18. Further, the speech coder 1 is used for data decoding at the time of speech coding.
Encoding processor 2 having a memory 3 for storing data
And the input voice data 12 received from the A / D converter 7.
Alternatively, the decoding data 18 received from the wireless unit 5 is written in the memory 3 of the processor 2, and the transmission data 19 of the encoding result stored in the memory 3 is passed to the wireless unit 5 to output the decoded output voice. The automatic memory access device 4 outputs the data 15 to the D / A converter 6.

In the above digital radio communication device, first, the voice is taken in as a analog input voice signal 11 in the microphone 10 at regular intervals, converted into digital input voice data 12 in the A / D converter 7, and voice encoded. Hand over to vessel 1. The speech coder 1 receives the speech data 12 by the automatic memory access unit 4 and stores it in the memory 3.
Here, in the automatic memory access device 4, the input voice data 1
The number of received data of 2 is confirmed, and when the data acquisition of the number required for one-time voice coding is completed, the coding start signal 13 is issued to the processor 2. At this time, if the processor 2 has stopped operating, the start signal 21 is set to the encoding start signal 1
Issue before 3. Upon receipt of the coding start signal 13, the processor 2 first determines whether there is sound or not in the voice, and if there is sound, starts the coding process, and if there is no sound, stops the operation. When the encoding is completed, the result is stored in the memory 3,
The automatic memory access device 4 passes it to the radio section 5 as transmission data 19 and the radio section 5 transmits it as a transmission signal 20 through the antenna 8. Further, the reception signal 17 received by the antenna 8 is converted into the decoding data 18 by the radio section 5 and then passed to the speech encoder 1. In the speech coder 1 which receives this, the automatic memory access device 4 decodes the decoding data.
The data 18 is stored in the memory 3 of the processor 2 and the decoding start signal 14 is issued. If the processor 2 is not operating, the activation signal 21 is issued before the decoding start signal 14. As a result, the processor 2 decodes the voice and stores the result in the memory 3. After that, the automatic memory access unit 4 outputs the output voice data 15 decoded in the same period as the voice input period to the D / A converter 6. D / A
The converter 6 converts the received output voice data 15 into an analog output voice signal 16 and the output voice signal 16 is output as voice by the speaker 9.

The operation of the processor 2 will be described below with reference to the state transition diagram shown in FIG. First, as an initial state, the processor 2 is in a wait state 25 in which no signal processing is performed.
When the encoding start signal 13 is received from the automatic memory access device 4, the process moves to the voiced / unvoiced determination 26 (3
0). If it is determined that the voice is present, the state transits to the voice encoding state 28 in which the voice encoding process is performed (31), and if it is determined that the voice is not present, the state transits to the operation stop state 27 for stopping the operation of the processor. (32). Further, when the decoding start signal 14 is received in the voice encoding processing state 28, the state shifts to the decoding state 29 for performing the decoding processing (35), and after the decoding processing ends, the processing returns to the voice encoding processing 28 (36). When the decoding start signal 14 is received in the operation stop state as well, the state shifts to the decoding state 29 in which the decoding process is performed (37), and after the decoding process ends, the processor 2 returns to the operation stop state (38). In addition, when the voice encoding process is completed in the voice encoding state 28, w
In the ait state 25 (33), when the activation signal 21 and the encoding start signal 13 are received in the operation stop state 27, the process moves to the voiced / unvoiced determination 26 (34).

FIG. 3 shows the input voice data 1 in the automatic memory access unit 4 in accordance with the operation of the processor 2.
2 shows a procedure of capturing 2. First, the automatic memory access device 4 waits for the input voice data 12 to arrive from the A / D converter 7 (40). At this time, the input voice data 12 is received from the A / D converter 7 (41), and the received input voice data 12 is written in the memory 3 (4).
2). After that, a judgment (43) is made as to whether or not the necessary number of voice data for one encoding has been prepared.
(47) Wait for the next input voice data 12 (40), and if they are complete (48) Confirm that the processor 2 is operating
(44). At this time, if the processor 2 is operating (50),
The encoding start signal 13 is issued to the processor 2 (46), and if the operation is stopped (49), the processor start signal 21 is issued (45) and then the encoding start signal 13 is issued (4).
6).

FIG. 4 shows a transfer procedure of the decoding data 18 of the automatic memory access device 4. First, the decryption data 18 is waited (51), the decryption data 18 is received (52), and the received decryption data 18 is written in the memory 3 (53). Thereafter, it is determined whether the processor 2 is operating (54), and if it is operating (58), the decoding start signal 1
4 is sent (56), and if the processor 2 is stopped (5
7) The processor start signal 21 is sent (55) and the decoding start signal 14 is sent (56).

Next, the processing flow until the state returns to the previous state after shifting to the decoding processing state 29 in the processor 2.
Is shown in FIG. First, the decoding start signal 14 is received from the automatic memory access device 4 (61), and decoding processing is performed (62).
After that, it is confirmed whether the process before the decoding process is the encoding process 28 or the operation stop state 27 (63), and if it is the encoding process state 28 (67), the process returns to the encoding process state 28.
(36) If the operation stop state 27 is reached (66), the operation stop state 27 is restored (38).

Further, FIG. 6 shows a timing chart until the processor 2 shifts to the operation stop state 27 when it is determined to be silent in the silence determination 26. First, wait for data 2
5 to the processor 2 in FIG.
(71) Automatic memory access device 4
Sends a coding start signal 13 from the processor 2, and the processor 2 determines whether there is sound or no sound 26. Here, it is determined that there is no sound, and the operation stop state 27 is set. Further, the automatic memory access device 4 sends the encoding start signal 13 to the processor 2 and then starts taking in the next encoding voice.

FIG. 7 shows a timing chart from the operation stop state 27 to the speech coding state 28 of the processor 2. First, the start signal 21 is sent to the processor 2 from the automatic memory access device 4 which has completed (71) the acquisition of the voice data necessary for encoding, and the processor 2 receiving this signal starts the operation. After that, automatic memory access device 4
Then, the coding start signal 13 is continuously sent to the processor 2, and the processor 2 makes a voiced / non-voiced determination 26. Here, the voiced state is determined to be voiced and a voice encoding state 28 is set.

Next, the timing of the decoding process will be described. FIG. 8 shows the voice coding process state 28 to the decoding process state 2
9 is a timing chart when the state becomes 9, and FIG. 9 is a timing chart when the state 27 in which the operation of the processor 2 is stopped becomes the voice decoding processing state 29. First, in the voice encoding processing state 28, the automatic memory access device 4 receives the decoding data 18 from the wireless section 5 (7
2). The automatic memory access unit 4 receives the received decoding data.
The data 18 is stored in the memory 3 (73) and the decoding start signal 14 is sent to the processor 2. The processor 2, which has received the decoding start signal 14, stops the coding process and performs the interrupt process 74 for performing the decoding process, and then becomes the decoding process state 29 and performs the restoration process 75 after the decoding process is completed and the voice coding process state 28.
Return to. Further, in the operation stop state 27, the wireless unit 5
When the automatic memory access device 4 receives the decoding data 18 from the memory device 3, the automatic memory access device 4 writes the received decoding data 18 in the memory 73, and the processor 2 sends the start signal 21 and the decoding process start signal 14 to the processor 2. To send. Receiving this signal, the processor 2 starts operation and the decoding processing state 2
9 and the operation stop state 27 is set after the decoding process is completed.

[0020]

According to the present invention, in a voice coder used in a digital wireless communication device or the like, it is possible to reduce the power consumption of the voice coder for performing highly compressed voice coding.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a digital wireless communication device to which the present invention has been applied.

FIG. 2 is a state transition diagram of a processor of a speech coder according to the present invention.

FIG. 3 is a flow chart of encoded data transfer of the automatic memory access device.

FIG. 4 is a flow chart of decoding data transfer of the automatic memory access device.

FIG. 5 is a decryption interrupt processing flow chart.

FIG. 6 is a timing flow chart from silence determination to process stop.

FIG. 7 is a timing flow chart from process stop to encoding.

FIG. 8 is a flow chart of a decoding interrupt processing timing at the time of encoding processing.

FIG. 9 is a flowchart of a decoding interrupt processing timing from an operation stop state.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Voice encoder 2 ... Processor 3 ... Processor memory 4 ... Automatic memory access device 5 ... Radio part 6 ... D / A converter 7 ... A / D converter 8 ... Radio wave transmitting / receiving antenna 9 ... Speaker 10 ... Microphone 11 ... Input voice signal 12 ... Input voice data
Data 13 ... voice coding start signal 14 ... voice decoding start signal 15 ... output voice data 16 ... output voice signal 17 ... reception signal 18 ... reception / decoding data 19 ... transmission data 20 ... transmission signal 21 ... Control signal

Front page continuation (72) Yoshiaki Asakawa, Inventor Yoshiaki Asakawa 1-280, Higashi Koikekubo, Kokubunji, Tokyo

Claims (6)

[Claims] 1. A processor having a built-in processor, a voice coding / decoding memory provided in the processor, and the voice is converted into a digital code for each fixed number of digital input voice data by using the memory. In the speech coder having in the processor, a decoding means for decoding the input digital code into the same number of digital voice data as the fixed number and outputting the digital code data. Or, a memory access means for accessing the voice encoding / decoding memory and transmitting / receiving digital code data or digital voice data to / from the memory, interposed between the input / output of the digital voice data and the processor. And a means for stopping the operation of the processor when it is detected that the input voice data for performing the conversion does not include voice. And a speech coder. 2. The speech coder according to claim 1, wherein the memory access means transmits / receives digital code data or digital speech data to / from the speech coding / decoding memory even if the processor is not operating. A speech coder for writing the data to the memory or reading the data from the memory. 3. A speech encoding / decoding memory is provided in the processor, and the processor converts the speech into a digital code for each fixed number of input speech data and outputs the digital speech by using the memory. In a voice encoding method for decoding and outputting the same number of voice data as the fixed number from the generated digital code, memory access means for transmitting / receiving the digital code data or digital voice data to / from the memory is provided. A voice code characterized by including the following procedure between the memory access means, the memory and the processor, to start encoding by the operating processor or to stop the processor. Method. (1) The memory access means receives the audio data, writes the data in the memory, confirms the number of data, and completes the acquisition of the required number of audio data for encoding. (2) Next, the processor outputs the sound. It is determined whether or not there is sound, and if there is sound, the encoding process is started, and if there is no sound, the operation of the processor is stopped. (3) The memory access means fetches the next sound data. 4. The speech coding method according to claim 3, wherein
A speech encoding method characterized by starting decoding including the following steps. (1) The memory access means receives the digital code data for decoding and writes the data in the memory (2) Next, it is judged whether or not the processor is operating, and if it is operating, the decoding is started, (3) If the processor is stopped, start the processor and then start decoding. 5. The speech coding method according to claim 3 or 4, wherein the decoding interruption and restoration processing is performed by including the following procedure. (1) If the decoding data is received even if the processor is executing the encoding process, the encoding process is interrupted and the decoding process is performed. (2) Next, the processing state of the processor before the decoding process is confirmed and Accordingly, the coding processing state or the operation stop state is restored. 6. A speech coding method according to any one of claims 3 to 5, wherein the processor shifts from a stopped state to a coding state including the following steps. Method. (1) The memory access means receives the audio data, writes the data in the memory, confirms the number of data, and completes the acquisition of the audio data of the number necessary for encoding. (2) The memory access means is a processor (3) Next, the processor determines whether there is sound or no sound, and if it is judged that there is sound, the processor enters the voice coding state.