JPH04347900A - Telephone device - Google Patents

Abstract

PURPOSE:To obtain a voice signal while suppressing a noise signal. CONSTITUTION:The telephone device is provided with a voice encoding circuit 13 which extracts characteristics of the voice waveform of a voice signal of words spoken by a user from the voice signal and a voice waveform recording circuit 5 wherein the characteristics of the voice waveform extracted by the voice encoding circuit 13 are recorded. A comparison circuit 3 compares the characteristics of the voice waveform extracted by the circuit 13 with characteristics of voice waveform recorded in the voice waveform recording circuit 5. When a signal is sent, the comparison circuit 30 compares the characteristics of the voice waveform outputted by the voice encoding circuit 13 with the characteristics of the voice waveform recorded in the voice recording circuit 5 and the characteristics of the voice waveform having a smaller noise signal between the both as the sent signal according to the comparison result.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to telephone equipment, particularly mobile telephones.

0002

2. Description of the Related Art When talking on the telephone, if there is noise in the surroundings, this noise is picked up by the speaker, which significantly reduces the clarity of the conversation.

[0003] In this case, depending on the usage environment of the telephone,
Although you can try to make the handset pick up less background noise, there are times when this is not possible. For example, in the case of a car phone, engine noise, wind noise, etc. may be the noise.

[0004] Therefore, as a method to reduce the influence of such noise, a microphone is prepared separately from the transmitter to collect ambient noise, and the spectrum of the noise signal is used to provide feedback to the transmitter. Methods of suppressing noise signals included in output signals have been considered.

[0005]

[Problem to be Solved by the Invention] However, even if such a method is carried out, there is actually a difference between the spectrum of noise picked up by a transmitter and the spectrum of noise picked up by a microphone. Therefore, it is difficult to sufficiently suppress the noise signal contained in the output signal of the transmitter and extract only the voice signal. Moreover, it is almost impossible to enable hands-free use of a car phone or the like.

[0006] This invention attempts to solve these problems.

[0007]

[Means for Solving the Problems] Therefore, in the present invention, if the reference numerals of each part correspond to the embodiments described later, a voice code for extracting the characteristics of the voice waveform from the voice signal of the words uttered by the user is provided. an audio waveform recording circuit 5 that records the characteristics of the audio waveform extracted by the audio encoding circuit 13; A comparison circuit 30 is provided to compare the characteristics of the recorded speech waveform.
The characteristics of the voice waveform output from the voice waveform recording circuit 5 are compared with the characteristics of the voice waveform recorded in the voice waveform recording circuit 5, and as a result of this comparison, the characteristics of the voice waveform with the least noise signal of both are selected as the voice waveform that is transmitted. This is a signal that is sent out as a signal.

[0008]

[Operation] The audio waveform characteristics of the input audio signal are compared with the audio waveform characteristics registered in advance to check for coincidence or approximation, and an audio signal with suppressed noise signals is output.

[0009]

[Embodiment] In FIG. 1, 10 is a transmitting circuit, 20 is a receiving circuit, 30 is a system controller, 41 is an operation button such as a dial button, and 42 is a display device. It is used to control display, etc. For this reason, the system controller 30 is composed of a microcomputer, and includes operation buttons 41 and a display device 4.
2 are connected, and necessary control signals or data are accessed between each circuit.

Further, reference numeral 5 denotes a voice waveform characteristic recording circuit, in which various voice waveform characteristics of the user are recorded and held.

[Registration of voice waveform characteristics] This is a process of registering the characteristics of various voice waveforms of the user in the recording circuit 5, and is realized by, for example, executing the routine 60 shown in FIG. 2 by the system controller 30. .

That is, when the mobile phone is in standby mode, for example, if a function button of the operation buttons 41 is pressed and then a registration button is pressed, the routine 60 starts from step 61 and then from step 62.
, it will wait for voice input.

[0013] Therefore, when the user speaks a specified specific word into the transmitter 11, the voice signal (assuming that it does not include a noise signal) is supplied from the transmitter 11 to the A/D converter 12. The signal is encoded and compressed in the audio encoding circuit 13 to extract the characteristics of the audio waveform, and the characteristics of this audio waveform are provisionally registered in the recording circuit 5 in step 63. , then the process proceeds to step 64.

[0014] Then, when the function button of the operation buttons 41 is pressed and then the confirmation button is pressed, this is detected in step 64 and the process proceeds to step 64.
The process proceeds to step 65, and in this step 65,
The system will wait for voice input again.

Then, when the words input in step 62 are repeated into the transmitter 11, the voice signal is similarly processed to extract the characteristics of the voice waveform, and then in step 66, the words input in step 62 are repeated. and step 63
The characteristics of the audio waveform of the word provisionally registered in step 65
The characteristics of the audio waveform of the input word are compared with the characteristics of the audio waveform of the input word, and it is checked whether the characteristics of the audio waveform of the input word can be recognized through this comparison.

When the feature can be recognized, the process proceeds from step 66 to step 71, and in step 71, it is displayed on the display device 42 that the words input in steps 62 and 65 are adopted, and the process proceeds to step 71. Proceed to step 72.

Then, when the function button of the operation buttons 41 is pressed and then the registration button is pressed, this is detected in step 72 and the process proceeds to step 72.
The process proceeds to step 73, and in this step 73,
The characteristics of the speech waveform of the word temporarily registered in step 63 are officially registered (recorded) in the recording circuit 5, and then, in step 74, this routine 60 is ended.

However, in step 66, if the characteristics of the audio waveform of the input word cannot be recognized, the process proceeds from step 66 to step 75, and in this step 75, the display device 42 displays the input word in steps 62 and 65. An indication that the word cannot be adopted is displayed and the process proceeds to step 76. Then, in step 76, the temporary registration in step 63 is canceled, and then, in step 74, this routine 60 is ended.

Further, if the predetermined button operation is not performed in step 64 or 72, the process proceeds from step 64 or 72 to step 74, and this routine 60 is ended.

Thereafter, such processing is repeated, and the voice waveform characteristics necessary for transmitting a call are registered in the recording circuit 5 as a database (codebook) of voice waveform characteristics.

[Call] The voice signal from the transmitter 11 (this voice signal may include a noise signal) is transmitted to the A/D
A/D signal is supplied to the converter 12 and converted into a digital audio signal.
The converted signal is encoded and compressed in the encoding circuit 13, and the characteristics of the audio waveform are extracted.

The characteristics of the extracted audio waveform are then compared with the characteristics of the audio waveform registered in the recording circuit 5, and a digital audio signal with matching or similar characteristics is extracted. That is, data indicating the degree of ambient noise is supplied from the audio encoding circuit 13 to the system controller 30, and when there is little ambient noise, the encoding circuit 13 outputs a signal obtained by encoding and compressing the actual audio signal. When there is a lot of ambient noise, the audio waveform registered in the recording circuit 5 is output. In this way, a digital audio signal with suppressed noise signals is extracted from the encoding circuit 13.

[0023] Then, this digital audio signal is supplied to an encoding circuit 14 for error correction to perform encoding processing for error correction, and the signal subjected to this encoding processing is supplied to a transmitting circuit 15. It is converted into an uplink transmission signal, and this transmission signal is supplied to the transmission/reception antenna 43 and transmitted to the base station.

Further, a downlink transmission signal from the base station is received by the antenna 43 and supplied to the receiving circuit 21, and a digital audio signal is taken out from the receiving circuit 21. Then, this extracted signal is supplied to the error correction circuit 22 to perform error correction, and this error-corrected signal is supplied to the audio decoding circuit 23 to be decoded into the original digital audio signal. , this signal is supplied to the D/A converter 24 and converted into the original analog audio signal by D/A converter 24.
This audio signal is then supplied to the handset 25.

In this way, the transmitting circuit 10 and the receiving circuit 20
In this case, according to the present invention, the characteristics of the waveform of the voice signal to be transmitted are recorded in advance in the recording circuit 5. Since only matching or similar waveforms are transmitted, even if the audio signal from the transmitter 11 contains a noise signal, the transmitted audio signal is an audio signal with the noise signal suppressed. It can be done.

Furthermore, since noise signals can be sufficiently suppressed, it is also possible to provide a hands-free car telephone or the like.

Moreover, the speech encoding circuit 13, that is,
Since a telephone using a voice codec does not require any new additional circuitry, noise in transmitted voice signals can be suppressed without increasing costs.

FIG. 3 shows an example of the audio encoding circuit 13 and the connection relationship between it and the recording circuit 5. This speech encoding circuit 13 performs encoding using an analysis-synthesis encoding method.
In this example, it is the case of RPE-LTP method,
Circuits 81 to 94 are the original speech encoding circuits.

That is, when registering the characteristics of the audio waveform, the digital audio signal from the A/D converter 12 is supplied from the terminal 81 to the LPC analysis circuit 82, the characteristics of the audio waveform are extracted, and this is the audio waveform. The signal is supplied to the characteristic recording circuit 5 and recorded.

When transmitting a call, the output signal of the LPC analysis circuit 82 is supplied to the LPC comparison and selection circuit 102, and the signal having the characteristics of the voice waveform recorded in the recording circuit 5 is supplied to the comparison and selection circuit. 102, and the comparison and selection circuit 102 compares the two to separate the noise signal included in the signal from the terminal 81 from the LPC information.

Then, the separated noise signal is supplied to the subtraction circuit 101, and the digital audio signal from the terminal 81 is supplied to the subtraction circuit 101, and the noise signal mixed in the digital audio signal from the terminal 81 is removed. is canceled by the noise signal from the comparison and selection circuit 102, and the subtraction circuit 101 takes out the audio signal with the noise signal suppressed.

Furthermore, this extracted audio signal is supplied to the inverse filter 83 and is also supplied to the comparison selection circuit 102.
At , one of the LPC information from the analysis circuit 82 and the LPC information from the recording circuit 5 is selected and taken out;
This extracted LPC information is supplied to the inverse filter 83. In this way, the inverse filter 83 predicts the audio signal.

The predicted audio signal is then sequentially supplied to the subtraction circuit 84 → multiplexing filter 85 → drive pulse down sampler 86 → APCM quantization circuit 87 to extract drive pulse vibration information, which is then sent to the multiplexing circuit. 88. Further, long-term prediction filter information is extracted by the long-term prediction filter 91 and the pitch analysis circuit 92, and is supplied to the multiplexing circuit 88. Note that 93 is a local decoding circuit, and 94 is an adding circuit.

Further, drive pulse position information is extracted from the down sampler 86 and supplied to the multiplexing circuit 88, and LPC information from the comparison and selection circuit 102 is supplied to the multiplexing circuit 88.

In this way, the multiplexing circuit 88 outputs the RPE-
A digital audio signal encoded by the LTP method is extracted, and this signal is supplied to the error correction encoding circuit 14 as described above.

In this case, since the noise signal in the audio signal supplied to the inverse filter 83 is suppressed by the signal from the comparison and selection circuit 102, the audio signal output from the multiplexing circuit 88 is also a noise signal. is being suppressed.

In the above description, when a specific button of the operation buttons 41 is operated, the characteristics of the voice waveform are registered. If so, you can also automatically register or update the audio waveform characteristics during the call.

Furthermore, the characteristics of the voice waveform can also be registered by the contracted operator registering the characteristics of the user's voice waveform in the recording circuit 5 or in the IC card used for the mobile phone when the mobile phone is purchased. You may also do so.

[0039]

[Effects of the Invention] According to the present invention, the characteristics of the waveform of the voice signal to be transmitted are compared with the characteristics of the voice waveform registered in advance in the recording circuit 5, and only those waveforms that match or are similar are transmitted. Therefore, even if the audio signal from the transmitter 11 contains a noise signal, the transmitted audio signal can be an audio signal with the noise signal suppressed.

Furthermore, since noise signals can be sufficiently suppressed, it is also possible to provide hands-free car telephones and the like.

[0041] Moreover, the speech encoding circuit 13, that is,
Since a telephone using a voice codec does not require any new additional circuitry, noise in transmitted voice signals can be suppressed without increasing costs.

[Brief explanation of drawings]

FIG. 1 is a system diagram of an example of the present invention.

FIG. 2 is a flowchart illustrating an example of a voice waveform characteristic registration routine.

FIG. 3 is a flowchart showing an example of a speech encoding circuit.

[Explanation of symbols]

5 Audio waveform characteristic recording circuit 10 Transmitter circuit 11 Transmitter 12 A/D converter 13 Audio encoding circuit 14 Error correction encoding circuit 15 Transmission circuit 20 Receiving circuit 21 Receiving circuit 22 Error correction circuit 23 Audio decoding circuit 24 D/A converter 25　Telephone receiver 30 System controller 41 Operation button 42 Display device

Claims (1)

[Claims] 1. A voice encoding circuit that extracts characteristics of a voice waveform from a voice signal of words uttered by a user; and a voice waveform recording circuit that records the characteristics of the voice waveform extracted in this voice encoding circuit. , a comparison circuit that compares the characteristics of the audio waveform extracted by the audio encoding circuit and the characteristics of the audio waveform recorded in the audio waveform recording circuit, and when transmitting a call, the comparing circuit , the characteristics of the audio waveform output from the audio encoding circuit are compared with the characteristics of the audio waveform recorded in the audio waveform recording circuit, and as a result of this comparison, the audio waveform with the least noise signal of the two is determined. A telephone device that transmits the characteristics of a telephone as a signal to be transmitted.