JP2010160496A - Signal processing device and signal processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a signal processing device in which a noise suppressor demonstrates a sufficient function, and which transmits and receives high quality voice, even when use setting is changed. <P>SOLUTION: A voice coding device 120 is provided with: an A system coding section 121; a B system coding section 122; and a C system coding section 123, as three circuits for coding a voice data by algorithms which are different from one another. The noise suppressor 110 is provided with: an X system noise suppressing section 111; a Y system noise suppressing section 112; and a Z system noise suppressing section 113, as three circuits for suppressing a background noise by algorithms which are different from one another. A suppression system switching control section 114 makes the optimum noise suppressing section (any one of 111, 112 or 113) operate, according to the coding section (any one of 121, 122 or 123) which functions in the voice coding device 120, on the basis of information from the coding system switching control section 124. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a noise suppressor (noise canceller) which is used in various digital communication system wireless communication devices such as a digital cellular phone system and reduces noise superimposed on transmitted / received speech.

  As a basic mobile communication service, there is a telephone service using voice communication. The mobile phone system originally started with an analog system, but nowadays digital systems such as the Japanese PDC (Personal Digital Cellar) system have become mainstream.

  In the digital system, an A / D converter is required to convert an analog voice signal into a digital signal. However, only A / D conversion requires a coding rate of about 100 kbps and uses less radio resources. Considering this, it is necessary to compress this to 1/10 to 1/20. In response to this requirement, a high-efficiency encoding method for speech called speech compression is generally used and embodied as a speech codec.

  Currently, voice communication codecs with a coding rate of about 3.5 kbps to 32 kbps are used in mobile communications, but low-rate codecs can reduce the coding rate by making the most of the characteristics of voice signals. Therefore, even if sufficient sound quality is obtained with respect to sound, reproducibility and sound quality tend to be poor with respect to “sound” other than sound.

  Mobile phones that are often used outdoors often use low-rate audio codecs as applications and are used in environments with high background noise.

  If background noise is input to a low-rate speech codec focused on “speech” in this way, the sound will be altered, and the clarity and sound quality of speech in the background noise environment will deteriorate.

  As a countermeasure against this, a technique called a noise suppressor (or noise canceller) has been attracting attention in recent years for the purpose of suppressing background noise taken in from a microphone and inputting only speech into the speech codec.

  As a specific example, there is a description of a noise canceller in the “half-rate speech codec” chapter of “Digital Car Telephone System Standard RCR STD-27” of the Radio Wave Industry Association (ARIB).

  Recently, new audio codecs have been developed due to technological innovations. New audio codecs have been introduced into the system, switching between 2 modes (switching between 2 audio codecs), switching between 3 modes (switching between 3 audio codecs) There is a flow of multi-mode.

  On the other hand, there is only one audio codec, such as EVRC (Enhanced Variable Rate Codec) and AMR (Adaptive Multi Rate), which are known in the TIA IS-127 standard, but it supports multiple different coding rates. In consideration of the direction of multirate and the convenience of the user, a hands-free function that allows a user to talk without holding a terminal has also been taken in.

  However, in conventional communication devices that have been multimode or multirate, the noise suppressor does not perform sufficiently depending on the selected mode or rate due to the compatibility of the voice codec and the noise suppressor. There was a problem that a high quality transmitted voice or received voice could not be obtained.

  Also, with conventional communication equipment with hands-free function, the voice input path to noise suppressors such as microphones and analog amplifiers can be changed, characteristics can be varied, or echo control can be performed by hands-free / non-hands-free. Due to changes in the usage environment such as via a new device such as an echo canceller, there is a problem that the noise suppressor does not function sufficiently and high-quality transmitted or received speech cannot be obtained.

  In conventional communication equipment that has been added multi-mode, multi-rate, or hands-free function, the noise suppressor has sufficient functions depending on the usage setting, such as switching modes and rates, or switching between hands-free / non-hands-free There was a problem that the transmitted voice or the received voice deteriorated.

  The present invention has been made to solve the above problems, and even when the use setting is changed, such as switching between modes and rates, or switching between hands-free / non-hands-free, the noise suppressor exhibits a sufficient function. An object of the present invention is to provide a signal processing device, a signal processing method, and a recording medium capable of transmitting and receiving high-quality sound.

  To achieve the above object, the present invention according to claim 1 encodes a noise signal that suppresses a noise component contained in a voice signal and a voice signal in which the noise component is suppressed by the noise suppression means. A signal processing apparatus including an encoding unit that performs processing to generate compressed encoded audio data, wherein the encoding unit can perform a plurality of different encoding processes, and the noise suppression unit The noise suppression characteristics according to the encoding process performed by the encoding means are provided.

  In the signal processing device having the above configuration, when selectively performing a plurality of different encoding processes, noise corresponding to the encoding process performed in the subsequent stage is performed when the noise component included in the audio signal is suppressed in the previous stage. The component is suppressed.

  Therefore, according to the signal processing device having the above configuration, noise components are suppressed according to the encoding process. Therefore, even if the content of the encoding process is varied, sufficient noise component suppression is performed, resulting in high quality. Sound can be transmitted.

  In order to achieve the above object, the present invention according to claim 8 includes decoding means for decoding compressed encoded audio data into an audio signal, and noise included in the audio signal decoded by the decoding means. A signal processing apparatus including noise suppression means for suppressing components, wherein the decoding means can perform a plurality of different decoding processes according to the decoded speech data, and the noise suppression means is implemented by the decoding means. Noise suppression characteristics according to the decoding process are provided.

  In the signal processing device having the above configuration, when a plurality of different decoding processes are selectively performed and the noise component included in the audio signal decoded in the subsequent stage is suppressed, the noise component corresponding to the decoding process to be performed is reduced. I try to do suppression.

  Therefore, according to the signal processing device having the above configuration, noise components are suppressed according to the decoding process. Therefore, even if the content of the decoding process is varied, sufficient noise component suppression is performed, and high-quality audio is suppressed. Can be received.

  Furthermore, in order to achieve the above object, the present invention according to claim 15 is a signal processing device used in a device that can selectively use a hands-free function. The apparatus is configured to include noise suppression means for suppressing noise components according to whether or not voice input uses a function.

In the signal processing apparatus having the above-described configuration, noise components are suppressed according to whether or not the input voice signal is a voice input using a hands-free function.
Therefore, according to the signal processing apparatus having the above configuration, even if the voice input path is changed depending on whether the voice input uses the hands-free function, sufficient noise components are suppressed, and high-quality voice is input. be able to.

  In order to achieve the above object, the present invention according to claim 20 provides a noise suppression step for suppressing a noise component included in a speech signal, and a speech signal in which the noise component is suppressed in the noise suppression step. A signal processing method comprising: an encoding step of performing encoding processing to generate compressed encoded speech data, wherein the encoding step selectively performs a plurality of different encoding processing, and noise suppression In the process, noise components are suppressed according to the encoding process selectively performed in the encoding process on the speech signal.

  In the signal processing method having the above-described configuration, when a plurality of different encoding processes are selectively performed, noise corresponding to the encoding process performed in the subsequent stage is performed when the noise component included in the audio signal is suppressed in the previous stage. The component is suppressed.

  Therefore, according to the signal processing method having the above configuration, noise components are suppressed according to the encoding process. Therefore, even if the content of the encoding process is varied, sufficient noise component suppression is performed, resulting in high quality. Sound can be transmitted.

  In order to achieve the above object, the present invention according to claim 21 includes a decoding step of decoding compressed encoded audio data into an audio signal, and noise included in the audio signal decoded by the decoding step. A noise suppression step of suppressing a component, wherein the decoding step selectively performs a plurality of different decoding processes on the encoded speech data according to the encoded speech data, and noise suppression In the process, the noise component is suppressed for the audio signal according to the decoding process selectively performed in the decoding process.

  In the signal processing method having the above configuration, when a plurality of different decoding processes are selectively performed and the noise component included in the audio signal decoded in the subsequent stage is suppressed, the noise component corresponding to the decoding process to be performed is reduced. I try to do suppression.

  Therefore, according to the signal processing method having the above-described configuration, noise components are suppressed according to the decoding process. Therefore, even if the contents of the decoding process are varied, sufficient noise component suppression is performed, and high-quality audio is suppressed. Can be received.

  Furthermore, in order to achieve the above object, the present invention according to claim 22 is a signal processing method used for a device capable of selectively using a hands-free function. A noise suppression process for suppressing noise components according to whether or not the voice input uses the free function is provided.

In the signal processing method having the above configuration, noise components are suppressed according to whether or not the input voice signal is a voice input using a hands-free function.
Therefore, according to the signal processing method having the above-described configuration, even if the voice input path is changed depending on whether the voice input is performed using the hands-free function, sufficient noise components are suppressed and high-quality voice is input. be able to.

  In order to achieve the above object, the present invention according to claim 23 provides a noise suppression procedure for suppressing a noise component included in a speech signal, and a speech signal in which the noise component is suppressed by the noise suppression procedure. A computer-readable recording medium storing a program for causing a computer to execute an encoding procedure for generating encoded audio data compressed by performing an encoding process, wherein the encoding procedure includes a plurality of different codes. The noise suppression procedure is a procedure for suppressing noise components according to the encoding process selectively performed in the encoding procedure for the speech signal.

  According to the recording medium having the above configuration, when the computer selectively performs a plurality of different encoding processes, the encoding process performed in the subsequent stage when the noise component included in the audio signal is suppressed in the previous stage. The noise component is suppressed according to.

  Therefore, according to the recording medium having the above configuration, in the computer, the noise component is suppressed according to the encoding process. Therefore, even if the content of the encoding process is varied, sufficient noise component suppression is performed. High quality voice can be transmitted.

  Furthermore, in order to achieve the above object, the present invention according to claim 24 provides a decoding procedure for decoding compressed encoded audio data into an audio signal, and a noise included in the audio signal decoded by the decoding procedure. A computer-readable recording medium storing a program for causing a computer to execute noise suppression means for suppressing a component, wherein a decoding procedure selectively encodes a plurality of different decoding processes according to encoded audio data. The noise suppression procedure is a procedure for suppressing a noise component according to a decoding process selectively performed in the decoding procedure for the audio signal.

  According to the recording medium having the above configuration, when the computer selectively performs a plurality of different decoding processes and suppresses the noise component included in the audio signal decoded in the subsequent stage, the computer performs the decoding process. The noise component is suppressed.

  Therefore, according to the recording medium having the above configuration, the computer suppresses the noise component according to the decoding process. Therefore, even if the content of the decoding process is varied, the noise component is sufficiently suppressed, and the high quality is achieved. Voice can be received.

  Furthermore, in order to achieve the above-mentioned object, the present invention according to claim 25 uses a hands-free function for a procedure for selectively enabling the hands-free function and an input audio signal. A program for causing a computer to execute a noise suppression procedure for suppressing noise components depending on whether or not a voice is input is recorded.

  According to the recording medium having the above configuration, the computer suppresses noise components according to whether or not the input audio signal is an audio input using a hands-free function.

  Therefore, according to the recording medium having the above-described configuration, even if the voice input path is varied depending on whether or not the voice input is performed using the hands-free function, the computer can sufficiently suppress noise components and generate high-quality voice. Can be entered.

  In order to achieve the above object, the present invention according to claim 26 provides a noise suppression means for suppressing a noise component included in a speech signal, and a speech signal in which the noise component is suppressed by the noise suppression means. A signal processing apparatus including an encoding unit that performs encoding processing to generate compressed encoded audio data, wherein the encoding unit selectively performs encoding processing by a plurality of different encoding methods The noise suppression unit includes a plurality of noise suppression units having noise component suppression characteristics different from each other, and a plurality of noises are selected according to encoding processing selectively performed by the encoding unit on the audio signal. The noise component is suppressed by selectively using one noise suppressor from among the suppressors. The number of different encoding methods in the encoding means is P, and the noise suppressor of the noise suppression means The Q if (P, Q are both positive integers) was, P and Q, characterized in that in the P ≧ Q> 1 relationship.

  In the signal processing device having the above-described configuration, when selectively performing encoding processing using a plurality of different encoding schemes, encoding processing performed in the subsequent stage when suppressing noise components included in the audio signal in the preceding stage A noise suppressor that suppresses noise components in accordance with is selectively used from among a plurality of noise suppressors, and the number P of the plurality of encoding methods and the number Q of the plurality of noise suppressors are The relationship of P ≧ Q> 1 was established.

  Therefore, according to the signal processing device configured as described above, even when the relationship between the number P of coding schemes and the number Q of noise suppression units is P ≧ Q> 1, suppression of noise components according to the coding processing is suppressed. Therefore, even if the content of the encoding process is varied, sufficient noise component suppression is performed, and high-quality speech can be transmitted.

  Furthermore, in order to achieve the above object, the present invention according to claim 27 provides noise suppression means for suppressing a noise component included in a speech signal, and a speech signal in which the noise component is suppressed by the noise suppression means. A signal processing apparatus including an encoding unit that performs encoding processing to generate compressed encoded audio data, wherein the encoding unit selectively performs encoding processing at a plurality of different encoding rates. The noise suppression unit includes a plurality of noise suppression units having noise component suppression characteristics different from each other, and a plurality of noises are selected according to encoding processing selectively performed by the encoding unit on the audio signal. The noise component is suppressed by selectively using one noise suppressor from among the suppressors. The number of different encoding rates in the encoding means is R, and the noise suppression of the noise suppression means. The number of parts Q if (R, Q are both positive integers) was, R and Q, characterized in that a relation of R ≧ Q> 1.

  In the signal processing device having the above configuration, when selectively performing encoding processing at a plurality of different encoding rates, encoding processing performed in the subsequent stage when suppressing noise components included in the audio signal in the preceding stage The noise suppressor that suppresses the noise component according to the signal is selectively used from the plurality of noise suppressors, and the number R of the plurality of different encoding rates and the number Q of the plurality of noise suppressors are The relation of R ≧ Q> 1 was established.

  Therefore, according to the signal processing device having the above configuration, even when the relationship between the number R of coding rates and the number Q of noise suppression units is R ≧ Q> 1, the suppression of noise components according to the coding processing is suppressed. Therefore, even if the content of the encoding process is varied, sufficient noise component suppression is performed, and high-quality speech can be transmitted.

  Furthermore, in order to achieve the above object, the present invention according to claim 28 includes a noise suppression unit that suppresses a noise component included in the audio signal, and an audio signal in which the noise component is suppressed by the noise suppression unit. A signal processing apparatus comprising: encoding means for performing encoding processing on the signal to generate compressed encoded audio data, wherein the encoding means selectively performs encoding processing by a plurality of different encoding methods. The noise suppression unit performs noise component suppression on the speech signal, and is selectively performed by the noise suppression unit capable of varying the noise component suppression characteristic according to the parameter setting and the encoding unit. Parameter setting means for setting a parameter corresponding to the encoded processing in the noise suppression unit, P being the number of different encoding methods in the encoding means, and noise suppression of the noise suppression means If the number of parameters to be set to the S (P, S are both positive integers) to, P and S, which lies in the P ≧ S> 1 relationship.

  In the signal processing device having the above-described configuration, when selectively performing encoding processing using a plurality of different encoding schemes, encoding processing performed in the subsequent stage when suppressing noise components included in the audio signal in the preceding stage A parameter is selectively used from among a plurality of parameters for the noise suppressor so that the noise suppressor suppresses a noise component with characteristics according to the number P, the number P of the plurality of encoding methods, and the above The relation with the number S of the plurality of parameter sets is set such that P ≧ S> 1.

  Therefore, according to the signal processing apparatus having the above configuration, even when the relationship between the number P of encoding schemes and the number S of parameter sets is P ≧ S> 1, noise components are suppressed according to the encoding processing. Therefore, even if the content of the encoding process is varied, sufficient noise components are suppressed and high-quality speech can be transmitted.

  Furthermore, in order to achieve the above object, the present invention according to claim 29 includes a noise suppression unit that suppresses a noise component included in the audio signal, and an audio signal in which the noise component is suppressed by the noise suppression unit. A signal processing device comprising: encoding means for generating compressed encoded speech data, wherein the encoding means selectively performs encoding processing at a plurality of different encoding rates The noise suppression unit suppresses the noise component of the speech signal, and is selectively performed by the noise suppression unit capable of changing the noise component suppression characteristic according to the parameter setting and the encoding unit. Parameter setting means for setting a parameter according to the encoding processing in the noise suppression unit, R being the number of different encoding rates in the encoding means, and noise suppression of the noise suppression means. If the number of parameter sets to be set in section was S (R, S are both positive integers), R and S, characterized in that a relation of R ≧ S> 1.

  In the signal processing device having the above configuration, when selectively performing encoding processing at a plurality of different encoding rates, encoding processing performed in the subsequent stage when suppressing noise components included in the audio signal in the preceding stage The noise suppressor selectively uses a parameter from among a plurality of parameters so that the noise suppressor suppresses the noise component with a characteristic according to the above. The relationship with the number S of the plurality of parameter sets is set to R ≧ S> 1.

  Therefore, according to the signal processing device having the above configuration, even when the relationship between the number R of the coding rate equations and the number S of parameter sets is R ≧ S> 1, the suppression of noise components according to the coding processing is suppressed. Therefore, even if the content of the encoding process is varied, sufficient noise component suppression is performed, and high-quality speech can be transmitted.

  As described above, according to the present invention, when a plurality of different encoding processes are selectively performed, the noise component included in the speech signal is suppressed in accordance with the encoding process performed in the subsequent stage. The noise component is suppressed.

  Therefore, according to the present invention, the noise component is suppressed according to the encoding process. Therefore, even if the content of the encoding process is varied, the noise component is sufficiently suppressed, and high-quality speech is transmitted. It is possible to provide a signal processing apparatus, a signal processing method, and a recording medium that can be used.

  Further, in the present invention, when a plurality of different decoding processes are selectively performed and the noise component included in the audio signal decoded in the subsequent stage is suppressed, the noise component is suppressed according to the decoding process to be performed. Like to do.

  Therefore, according to the present invention, the noise component is suppressed according to the decoding process. Therefore, even if the content of the decoding process is varied, the noise component is sufficiently suppressed and high quality speech is received. Can be provided, a signal processing method, and a recording medium.

Furthermore, in the present invention, noise components are suppressed according to whether or not the input voice signal is a voice input using a hands-free function.
A signal processing apparatus, a signal processing method, and a signal processing apparatus capable of sufficiently suppressing a noise component and inputting a high-quality voice even if the voice input path is varied depending on whether or not the voice input uses the hands-free function, and A recording medium can be provided.

1 is a circuit block diagram showing the configuration of a first embodiment of a signal processing apparatus according to the present invention; The flowchart for demonstrating operation | movement of the signal processing apparatus concerning 1st Embodiment shown in FIG. The circuit block diagram which shows the structure of 2nd Embodiment of the signal processing apparatus concerning this invention. The flowchart for demonstrating operation | movement of the signal processing apparatus concerning 2nd Embodiment shown in FIG. The circuit block diagram which shows the structure of 3rd Embodiment of the signal processing apparatus concerning this invention. The flowchart for demonstrating operation | movement of the signal processing apparatus concerning 3rd Embodiment shown in FIG. The circuit block diagram which shows the structure of 4th Embodiment of the signal processing apparatus concerning this invention. The flowchart for demonstrating operation | movement of the signal processing apparatus concerning 4th Embodiment shown in FIG. The figure which shows schematic structure by the side of encoding the input audio | voice with which the signal processing apparatus concerning this invention is applied. The circuit block diagram which shows the structure of the modification of this invention taking 3rd Embodiment as an example. The figure which shows an example of the correspondence of an encoding process and a noise suppression process in case the number of types of an encoding process and the number of types of a noise suppression process do not correspond. The figure which shows schematic structure by the side which decodes the output audio | voice to which the signal processing apparatus concerning this invention is applied. FIG. 8 is a circuit block diagram showing a configuration of a modification of the speech encoder of the signal processing apparatus shown in FIGS. 1 to 7. FIG. 8 is a circuit block diagram showing a configuration of a modified example of the noise suppressor of the signal processing device shown in FIGS. 1 to 7; The figure which shows the example of the parameter set to the noise suppressor shown in FIG. The figure which shows the example of the parameter set to the noise suppressor shown in FIG. The figure which shows the example of the parameter set to the noise suppressor shown in FIG. FIG. 8 is a circuit block diagram showing a configuration of a modified example of the noise suppressor of the signal processing device shown in FIGS. 1 to 7; The circuit block diagram which shows the structure of 5th Embodiment of the signal processing apparatus concerning this invention. The flowchart for demonstrating operation | movement of the signal processing apparatus concerning 5th Embodiment shown in FIG. The circuit block diagram which shows the structure of 6th Embodiment of the signal processing apparatus concerning this invention. The flowchart for demonstrating operation | movement of the signal processing apparatus concerning 6th Embodiment shown in FIG.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 shows the configuration of a signal processing apparatus according to the first embodiment of the present invention.

  101 is a microphone that converts a user's transmitted voice into an electric signal, and 102 is an A / D converter that converts an analog voice signal obtained by the microphone 101 into voice data of a digital signal. 110 is a digital signal. A noise suppressor that suppresses background noise included in the audio data by signal processing, 103 is audio data in which background noise is suppressed by the noise suppressor 110, and 120 is an audio encoder that compresses and encodes 103 digital data. , 104 are encoded data compressed by the speech encoder 120.

  The speech encoder 120 includes an A system encoding unit 121, a B system encoding unit 122, and a C system encoding unit 123 as three circuits for encoding speech data using different algorithms. A switching control unit 124 is provided.

  Here, for example, the A method encoding unit 121 is a method in which the encoding rate is low but the encoding sound quality with respect to the background noise is not so good, and the C method encoding unit 123 has a high encoding rate with respect to background noise. The B method encoding unit 122 has a sound quality intermediate between the A method encoding unit 121 and the C method encoding unit 123.

  The coding method switching control unit 124 functions as one of the A method coding unit 121, the B method coding unit 122, and the C method coding unit 123 according to the coding method selection command 105 from the outside. In addition, the switching control is performed, and information indicating a scheme functioning by the switching control is output to the noise suppressor 110 as the encoding scheme selection information 106.

  The noise suppressor 110 includes an X-type noise suppressor 111, a Y-type noise suppressor 112, and a Z-type noise suppressor 113 as three circuits for suppressing background noise by different algorithms, and a suppress method switching control unit 114. Is provided.

  The suppression method switching control unit 114 is configured so that any one of the X method noise suppression unit 111, the Y method noise suppression unit 112, and the Z method noise suppression unit 113 functions according to the encoding method selection information 106. Is controlled to be switched.

  That is, in the switching control by the suppression system switching control unit 114, the optimum noise suppression unit (111, 112, 113) is matched with the encoding unit (121, 122, 123) functioning in the speech encoder 120. If the A method encoding unit 121 functions based on the encoding method selection information 106, the X method noise suppression unit 111 is selected, and the B method encoding unit 122 When functioning, the Y method noise suppression unit 112 is selected, and when the C method encoding unit 123 functions, the Z method noise suppression unit 113 is selected.

  In order to optimize the correspondence between such an encoding unit and a noise suppression unit, for example, the X-type noise suppression unit 111 has a spectral process in a frequency region that is a little complicated but has a high noise suppression performance. A method using the subtraction method (SS method) is used for the Y-type noise suppression unit 112 in the same manner, but a method simplified from the X-type noise suppression unit 111 is used for the Z-type noise suppression unit 112. For the 113, a filter with adaptive filtering in the time domain having a relatively simple configuration is used.

Next, the operation of the signal processing apparatus according to the first embodiment will be described. FIG. 2 is a flowchart for explaining this operation.
If a coding method selection command 105 indicating “use the A method as a coding method” is input to the coding method switching control unit 124 from the command input standby state in step 2a, the step 2b Then, the system to be designated is determined, and since it is the system A, the process proceeds to step 2c.

  In step 2c, the encoding method switching control unit 124 performs switching control so that the digital data 103 is input to the A method encoding unit 121, and the A method encoding unit 121 is input in response to the digital data 103. 103 encoding is started.

  In step 2c, in parallel with this switching control, the encoding scheme switching control section 124 indicates that the A scheme encoding section 121 is used for encoding the digital data 103, and the encoding scheme selection information 106. Is output to the suppression system switching control unit 114, and the process proceeds to step 2d.

  In step 2 d, the output of the A / D converter 102 is output in order for the suppression method switching control unit 114 to perform noise suppression by the X method noise suppression unit 111 optimized for encoding by the A method encoding unit 121. Switching control is performed so that the signal is input to the X-type noise suppression unit 111, and the process proceeds to step 2i.

  By such a switching control operation, the output of the A / D converter 102 is noise-suppressed by the X-type noise suppression unit 111 and input to the A-type encoding unit 121 as digital data 103, where it is encoded. The encoded data 104 is compressed as data.

  In step 2i, when an encoding method selection command 105 indicating that “use the B method as an encoding method” is input to the encoding method switching control unit 124, the process proceeds to step 2b to specify the specified method. Since it is a B system, it transfers to step 2e.

  In step 2e, the encoding system switching control unit 124 controls the switching so that the digital data 103 is input to the B system encoding unit 122 at a timing, and the A system encoding unit 121 responds to the switching. The operation is stopped, and encoding of the digital data 103 to which the B system encoding unit 122 is input is started instead.

  In step 2e, in parallel with this switching control, the encoding method switching control unit 124 indicates that the B method encoding unit 122 is used for encoding the digital data 103, and the encoding method selection information 106. Is output to the suppression system switching control unit 114, and the process proceeds to step 2f.

  In step 2 f, the output of the A / D converter 102 is output so that the suppression method switching control unit 114 performs noise suppression by the Y method noise suppression unit 112 optimized for encoding by the B method encoding unit 122. Switching control is performed so that the signal is input to the Y-type noise suppression unit 112, and the process proceeds to step 2i.

  By such a switching control operation, the output of the A / D converter 102 is noise-suppressed by the Y-system noise suppressor 112, and is input as digital data 103 to the B-system encoder 122, where it is encoded. The encoded data 104 is compressed as data.

  In step 2i, when the digital data 103 is encoded by the A method encoding unit 121 or the B method encoding unit 122 as described above, the message “Use C method as the encoding method” is used. When the encoding method selection command 105 indicating is input to the encoding method switching control unit 124, the process proceeds to step 2b to determine the designated system, and the process proceeds to step 2g because it is the C system.

  In step 2g, the encoding system switching control unit 124 controls the switching so that the digital data 103 is input to the C system encoding unit 123 in a timely manner, and in response thereto, the A system encoding being used. The operation of the unit 121 or the B system encoding unit 122 is stopped, and encoding of the digital data 103 to which the C system encoding unit 123 is input instead is started.

  In step 2g, in parallel with the switching control, information indicating that the encoding method switching control unit 124 uses the C method encoding unit 123 for encoding the digital data 103 is indicated by the encoding method selection information 106. Is output to the suppression system switching control unit 114, and the process proceeds to step 2h.

  In step 2 h, the output of the A / D converter 102 is output so that the suppression method switching control unit 114 performs noise suppression by the Z method noise suppression unit 113 optimized for encoding by the C method encoding unit 123. Switching control is performed so that the signal is input to the Z-type noise suppression unit 113, and the process proceeds to step 2i.

  By such a switching control operation, the output of the A / D converter 102 is noise-suppressed by the Z-system noise suppressor 113, and is input to the C-system encoder 122 as digital data 103, where it is encoded. The encoded data 104 is compressed as data.

  If no command is input in step 2i, the process proceeds to step 2j. In step 2j, it is determined whether or not a communication termination request has been made. If there is a termination request, the process is terminated. If there is no termination request, command input is again performed in step 2i. Monitor.

  As described above, in the signal processing apparatus having the above-described configuration, in order to obtain the encoded data 104 that has been compressed, it is matched to the encoding unit (any of 121, 122, and 123) that functions in the speech encoder 120. The optimum noise suppression unit (any of 111, 112, 113) is made to function.

  Therefore, according to the signal processing device having the above configuration, noise suppression is performed by the noise suppression unit that is optimal for encoding by the speech encoder 120. Therefore, the noise suppression unit exhibits a sufficient function and produces high-quality speech. Can be sent.

  The present invention is not limited to the above embodiment. For example, in the first embodiment, the suppression method switching control unit 114 is a coding unit that functions in the speech encoder 120 based on the coding method selection information 106 from the coding method switching control unit 124. In addition, the optimal noise suppression unit was made to function.

  Instead, for example, the suppression method switching control unit 114 causes the optimum noise suppression unit to function in accordance with the encoding unit functioning in the speech encoder 120 based on the encoding method selection command 105. Can exhibit the same effect.

  In this case, the suppression method switching control unit 114 performs switching control so that the optimum noise suppression unit functions at a timing that takes into account the switching timing of the encoding unit in the speech encoder 120.

Next explained is a signal processing apparatus according to the second embodiment of the invention. FIG. 3 shows the configuration.
Reference numeral 201 denotes a microphone that converts a user's transmission voice into an electric signal, and 202 denotes an A / D converter that converts an analog voice signal obtained by the microphone 201 into voice data of a digital signal. 210 denotes a digital signal. A noise suppressor that suppresses background noise included in the audio data by signal processing, 203 is audio data whose background noise is suppressed by the noise suppressor 210, and 220 is an audio encoder that compresses and encodes the digital data 203. , 204 are encoded data compressed by the speech encoder 220.

  The speech encoder 220 includes, as three circuits that encode speech data using different algorithms, an A method encoding unit 221, a B method encoding unit 222, and a C method encoding unit 223, as well as an encoding method. A switching control unit 224 is provided.

  Here, for example, the A method encoding unit 221 is a method in which the encoding rate is low but the encoding sound quality with respect to the background noise is not so good, and the C method encoding unit 223 has a high encoding rate with respect to the background noise. The B system encoding unit 222 has a sound quality intermediate between the A system encoding unit 221 and the C system encoding unit 223.

  The coding method switching control unit 224 has one of the A method coding unit 221, the B method coding unit 222, and the C method coding unit 223 functioning according to the coding method selection command 205 from the outside. In addition, the switching control is performed, and information indicating a scheme functioning by the switching control is output to the noise suppressor 210 as the encoding scheme selection information 206.

The noise suppressor 210 includes a noise suppression unit 215, a parameter table 216, and a parameter switching control unit 217.
The noise suppression unit 215 suppresses background noise included in the audio data output from the A / D converter 202, and its suppression characteristic is controlled by parameters input from the parameter table 216.

  The parameter table 216 is a table for storing parameters for setting the characteristics of the background noise suppression processing performed by the noise suppression unit 215. The A system encoding unit 221, the B system encoding unit 222, and the C system encoding unit 223 are stored in the parameter table 216. The three parameter sets that have optimum suppression characteristics for each encoding method are stored, and the parameter sets are input to the noise suppression unit 215 under the control of the parameter switching control unit 217.

  In this embodiment, a parameter set having five parameters as elements is assumed, and parameter sets (three in this embodiment) are prepared for each encoding method.

  The parameter switching control unit 217 controls the parameter table 216, and based on the encoding method selection information 206, the optimum parameter for the encoding unit (any one of 221, 222, 223) that functions in the speech encoder 220 The set is selectively set in the noise suppression unit 215.

  In order to optimize the correspondence between the parameter setting (suppression characteristic setting) of the encoding unit and the noise suppression unit, for example, as a parameter set corresponding to the A method encoding unit 221, a noise suppression amount is set as follows. The parameter set is set to be strong and can obtain a characteristic that suppresses noise as much as possible even if the audio part is slightly distorted, and the parameter set corresponding to the C system encoding unit 223 has a light noise suppression amount. For this reason, the parameter set is set so as to obtain a characteristic of transmitting noise that can be heard naturally.

  In addition, the parameter set corresponding to the B system encoding unit 222 is a parameter set that can obtain a characteristic intermediate between the characteristics for the A system encoding unit 221 and the characteristics for the C system encoding unit 223.

Next, the operation of the signal processing apparatus according to the second embodiment will be described. FIG. 4 is a flowchart for explaining this operation.
When a coding method selection command 205 indicating that “use the A method as the coding method” is input to the coding method switching control unit 224 from the command input standby state in step 4a, the step 4b. Then, the designated system is determined, and since it is the A system, the process proceeds to Step 4c.

  In step 4c, the encoding system switching control unit 224 performs switching control so that the digital data 203 is input to the A system encoding unit 221, and the A system encoding unit 221 responds to this by inputting the digital data. 203 encoding is started.

  In step 4c, in parallel with this switching control, information indicating that the encoding method switching control unit 224 uses the A method encoding unit 221 for encoding the digital data 203 is displayed as encoding method selection information 206. Is output to the parameter switching control unit 217, and the process proceeds to step 4d.

  In step 4d, the parameter switching control unit 217 sets the parameter set corresponding to the A method encoding unit 221 so that the noise suppression characteristic of the noise suppression unit 215 becomes an optimum characteristic for encoding by the A method encoding unit 221. Is input to the noise suppression unit 215 from the parameter table 216, and the process proceeds to step 4i.

  With such parameter setting (suppression characteristic setting) control operation, noise is suppressed in the output of the A / D converter 202 with a suppression characteristic suitable for encoding by the A-system encoding unit 221, and A The data is input to the system encoding unit 221, is encoded here, and is output as encoded data 204 after data compression.

  In step 4i, when an encoding method selection command 205 indicating that “use the B method as the encoding method” is input to the encoding method switching control unit 224, the process proceeds to step 4b to specify the specified method. Since it is a B system, it transfers to step 4e.

  In step 4e, the encoding method switching control unit 224 performs switching control so that the digital data 203 is input to the B method encoding unit 222 in a timely manner, and the A method encoding unit 221 responds accordingly. The operation is stopped, and the encoding of the digital data 203 input by the B system encoding unit 222 is started instead.

  Further, in step 4e, in parallel with this switching control, information indicating that the encoding method switching control unit 224 uses the B method encoding unit 222 for encoding the digital data 203, and encoding method selection information 206 Is output to the parameter switching control unit 217, and the process proceeds to step 4f.

  In step 4f, the parameter switching control unit 217 sets the parameter set corresponding to the B method encoding unit 222 so that the noise suppression characteristic of the noise suppression unit 215 becomes the optimum characteristic for encoding by the B method encoding unit 222. Is input to the noise suppression unit 215 from the parameter table 216, and the process proceeds to step 4i.

  By such parameter setting (suppression characteristic setting) control operation, noise is suppressed in the output of the A / D converter 202 with a suppression characteristic suitable for encoding by the B-system encoding unit 222, and B The data is input to the system encoding unit 222, encoded here, and output as encoded data 204 after data compression.

  Further, in step 4i, when the digital data 203 is encoded by the A method encoding unit 221 or the B method encoding unit 222 as described above, the message “Use C method as the encoding method” is used. Is input to the encoding system switching control unit 224, the process proceeds to step 4b to determine the designated system. Since the system is the C system, the process proceeds to step 4g.

  In step 4g, the coding system switching control unit 224 controls switching so that the digital data 203 is input to the C system coding unit 223 at the timing, and in response to this, the A system coding in use. The operation of the unit 221 or the B system encoding unit 222 is stopped, and the encoding of the digital data 203 to which the C system encoding unit 223 is input instead is started.

  In step 4g, in parallel with the switching control, information indicating that the coding method switching control unit 224 uses the C method coding unit 223 for coding the digital data 203 is displayed as coding method selection information 206. Is output to the parameter switching control unit 217, and the process proceeds to step 4h.

  In step 4h, the parameter switching control unit 217 sets the parameter set corresponding to the C method encoding unit 223 so that the noise suppression characteristic of the noise suppression unit 215 becomes the optimum characteristic for encoding by the C method encoding unit 223. Is input to the noise suppression unit 215 from the parameter table 216, and the process proceeds to step 4i.

  By such parameter setting (suppression characteristic setting) control operation, noise is suppressed in the output of the A / D converter 202 with a suppression characteristic suitable for encoding by the C-system encoding unit 223, and the digital data 203 is converted to C The data is input to the system encoding unit 223, is encoded here, and is output as encoded data 204 that has been subjected to data compression.

  If no command is input in step 4i, the process proceeds to step 4j. In step 4j, it is determined whether or not a communication termination request has been made. If there is a termination request, the process is terminated. On the other hand, if there is no termination request, command input is again performed in step 4i. Monitor.

  As described above, in the signal processing apparatus having the above configuration, in order to obtain encoded data 204 that has been compressed, it is matched with the encoding unit (any one of 221, 222, and 223) that functions in the speech encoder 220. The parameters of the noise suppression unit 215 are varied to set the noise suppression characteristic of the noise suppression unit 215 to an optimum characteristic for the encoding process.

  Therefore, according to the signal processing apparatus configured as described above, optimal noise suppression is performed in accordance with the encoding of the audio encoder 220, so that the noise suppressor exhibits a sufficient function and transmits high-quality audio. can do.

  The present invention is not limited to the above embodiment. For example, in the second embodiment, the parameter switching control unit 217 matches the coding unit functioning in the speech encoder 220 based on the coding method selection information 206 from the coding method switching control unit 224. The noise suppression characteristic of the noise suppression unit 215 is changed to an optimum characteristic.

  Instead of this, for example, the parameter switching control unit 217 adjusts the noise suppression characteristic of the noise suppression unit 215 to the optimum characteristic in accordance with the encoding unit functioning in the speech encoder 220 based on the encoding method selection command 205. Even if it is made variable, the same effect can be exhibited.

  In this case, the parameter switching control unit 217 performs control to set a parameter set that obtains an optimum noise suppression characteristic at a timing that takes into account the switching timing of the encoding unit in the speech encoder 220.

Next explained is a signal processing apparatus according to the third embodiment of the invention. FIG. 5 shows the configuration.
Reference numeral 301 denotes a microphone that converts a user's transmitted voice into an electric signal and takes it in. 302 denotes an A / D converter that converts an analog voice signal obtained by the microphone 301 into voice data of a digital signal. 310 denotes a digital signal. A noise suppressor that suppresses background noise included in the audio data by signal processing, 303 is audio data in which background noise is suppressed by the noise suppressor 310, and 320 is an audio encoder that compresses and encodes 303 digital data. , 304 are encoded data compressed by the speech encoder 320.

  The speech encoder 320 includes an A rate encoder 321, a B rate encoder 322, and a C rate encoder 323 as three circuits for encoding speech data at different encoding rates. A conversion rate switching control unit 324 is provided.

  Here, for example, the A rate encoding unit 321 has the lowest encoding rate among the above three circuits that perform encoding, and the C rate encoding unit 323 includes the above three circuits that perform encoding. Of these, the coding rate is the highest, and the B rate coding unit 322 has a coding rate intermediate between the A rate coding unit 321 and the C rate coding unit 323.

  The coding rate switching control unit 324 functions as one of the A rate coding unit 321, the B rate coding unit 322, and the C rate coding unit 323 according to the coding rate selection command 305 from the outside. In addition, the switching control is performed, and information indicating the rate functioning by the switching control is output to the noise suppressor 310 as the coding rate selection information 306.

  The noise suppressor 310 includes an X-type noise suppressor 311, a Y-type noise suppressor 312, and a Z-type noise suppressor 313 as three circuits for suppressing background noise by different algorithms, and a suppress method switching control unit 314. Is provided.

  The suppression method switching control unit 314 is configured so that any one of the X method noise suppression unit 311, the Y method noise suppression unit 312, and the Z method noise suppression unit 313 functions according to the coding rate selection information 306. Is controlled to be switched.

  That is, in the switching control by the suppression system switching control unit 314, the optimum noise suppression unit (311, 312, 313) is matched with the encoding unit (any one of 321, 322, 323) functioning in the speech encoder 320. When the A rate encoding unit 321 functions based on the encoding rate selection information 306, the X-type noise suppression unit 311 is selected, and the B rate encoding unit 322 When functioning, the Y method noise suppression unit 312 is selected, and when the C rate encoding unit 323 functions, the Z method noise suppression unit 313 is selected.

  In order to optimize the correspondence between such an encoding unit and a noise suppression unit, for example, the X-type noise suppression unit 311 has a slightly complicated process but a spectral in a frequency region where noise suppression performance is high. The one using the subtraction method (SS method) is the SS method similarly to the Y method noise suppression unit 312, but the Z method noise suppression unit is a process simplified from the X method noise suppression unit 311. For the reference numeral 313, an adaptive filtering in the time domain having a relatively simple configuration is used.

Next, the operation of the signal processing apparatus according to the third embodiment will be described. FIG. 6 is a flowchart for explaining this operation.
When a coding rate selection command 305 indicating that “use the A rate as the coding rate” is input to the coding rate switching control unit 324 from the command input standby state in step 6a, the step 6b. Then, the designated rate is determined, and since it is the A rate, the process proceeds to step 6c.

  In step 6c, the coding rate switching control unit 324 performs switching control so that the digital data 303 is input to the A rate encoding unit 321, and the A rate encoding unit 321 is input in response to the switching. The encoding of 303 is started.

  In step 6c, in parallel with this switching control, the coding rate switching control unit 324 indicates that the A rate coding unit 321 is used for coding the digital data 303, and coding rate selection information 306. Is output to the suppression system switching control unit 314, and the process proceeds to step 6d.

  In step 6 d, the output of the A / D converter 302 is output so that the suppression method switching control unit 314 performs noise suppression by the X method noise suppression unit 311 optimized for encoding by the A rate encoding unit 321. Switching control is performed so that the signal is input to the X-type noise suppression unit 311, and the process proceeds to Step 6 i.

  By such a switching control operation, the output of the A / D converter 302 is noise-suppressed by the X-type noise suppression unit 311 and input to the A-rate encoding unit 321 as digital data 303, where it is encoded. It is output as encoded data 304 that has been compressed.

  In step 6i, when an encoding rate selection command 305 indicating that “use the B rate as the encoding rate” is input to the encoding rate switching control unit 324, the process proceeds to step 6b to specify the specified rate. Since it is a B rate, it transfers to step 6e.

  In step 6e, the coding rate switching control unit 324 switches the timing so that the digital data 303 is input to the B rate coding unit 322 in response to timing, and the A rate coding unit 321 responds accordingly. The operation is stopped, and encoding of the digital data 303 to which the B rate encoding unit 322 is input instead starts.

  In step 6e, in parallel with this switching control, the coding rate switching control unit 324 indicates that the B rate coding unit 322 is used for coding the digital data 303, and coding rate selection information 306. Is output to the suppression method switching control unit 314, and the process proceeds to step 6f.

  In step 6 f, the output of the A / D converter 302 is output so that the suppression method switching control unit 314 performs noise suppression by the Y method noise suppression unit 312 optimized for encoding by the B rate encoding unit 322. The switching control is performed so that the signal is input to the Y-type noise suppression unit 312 and the process proceeds to Step 6i.

  By such a switching control operation, the output of the A / D converter 302 is noise-suppressed by the Y-type noise suppression unit 312 and input to the B-rate encoding unit 322 as digital data 303, where it is encoded. It is output as encoded data 304 that has been compressed.

  In step 6i, when the digital data 303 is encoded by the A rate encoding unit 321 or the B rate encoding unit 322 as described above, the message “Use C rate as the encoding rate” is used. When the coding rate selection command 305 indicating is input to the coding rate switching control unit 324, the process proceeds to step 6b, where the designated system is determined. Since the system is the C system, the process proceeds to step 6g.

  In step 6g, the coding rate switching control unit 324 controls the switching so that the digital data 303 is input to the C rate coding unit 323 in a timely manner, and in response to this, the A rate coding in use. The operation of the unit 321 or the B rate encoding unit 322 is stopped, and encoding of the digital data 303 to which the C rate encoding unit 323 is input instead is started.

  In step 6g, in parallel with this switching control, the coding rate switching control unit 324 indicates that the C rate coding unit 323 is used for coding the digital data 303, and coding rate selection information 306. Is output to the suppression method switching control unit 314, and the process proceeds to step 6h.

  In step 6h, the output of the A / D converter 302 is output so that the suppression method switching control unit 314 performs noise suppression by the Z method noise suppression unit 313 optimized for encoding by the C rate encoding unit 323. Switching control is performed so that the signal is input to the Z-type noise suppression unit 313, and the process proceeds to Step 6i.

  By such a switching control operation, the output of the A / D converter 302 is noise-suppressed by the Z-type noise suppression unit 313, input to the C-rate encoding unit 322 as digital data 303, and encoded there. It is output as encoded data 304 that has been compressed.

  If no command is input in step 6i, the process proceeds to step 6j. In step 6j, it is determined whether or not an end request for communication has been made. If there is an end request, the process ends. On the other hand, if there is no end request, command input is again made in step 6i. Monitor.

  As described above, in the signal processing apparatus having the above-described configuration, in order to obtain encoded data 304 that is data-compressed, it is matched with the encoding unit (any one of 321, 322, and 323) that functions in the speech encoder 320. The optimum noise suppression unit (any of 311, 312, 313) is made to function.

  Therefore, according to the signal processing apparatus configured as described above, noise suppression is performed by the noise suppression unit that is optimal for encoding by the speech encoder 320. Therefore, the noise suppression unit exhibits a sufficient function and can produce high-quality speech. Can be sent.

  The present invention is not limited to the above embodiment. For example, in the third embodiment, the suppression scheme switching control unit 314 is a coding unit that functions in the speech coder 320 based on the coding rate selection information 306 from the coding rate switching control unit 324. In addition, the optimal noise suppression unit was made to function.

  Instead, for example, the suppression method switching control unit 314 causes the optimum noise suppression unit to function in accordance with the encoding unit functioning in the speech encoder 320 based on the encoding rate selection command 305. Can exhibit the same effect.

  In this case, the suppression method switching control unit 314 performs switching control so that the optimum noise suppression unit functions at a timing that takes into account the switching timing of the encoding unit in the speech encoder 320.

Next explained is a signal processing apparatus according to the fourth embodiment of the invention. FIG. 7 shows the configuration.
401 is a microphone that converts a user's transmitted voice into an electric signal, and 402 is an A / D converter that converts an analog voice signal obtained by the microphone 401 into voice data of a digital signal. 410 is a digital signal. A noise suppressor that suppresses background noise included in the audio data by signal processing, 403 is audio data in which background noise is suppressed by the noise suppressor 410, and 420 is an audio encoder that compresses and encodes digital data 403. , 404 are encoded data compressed by the speech encoder 420.

  The audio encoder 420 includes an A rate encoding unit 421, a B rate encoding unit 422, and a C rate encoding unit 423 as three circuits for encoding audio data at different encoding rates. A conversion rate switching control unit 424 is provided.

  Here, for example, the A rate encoding unit 421 has the lowest encoding rate among the above three circuits that perform encoding, and the C rate encoding unit 423 includes the above three circuits that perform encoding. Among them, the encoding rate is the highest, and the B rate encoding unit 422 is set to an intermediate encoding rate between the A rate encoding unit 421 and the C rate encoding unit 423.

  In the coding rate switching control unit 424, any one of the A rate coding unit 421, the B rate coding unit 422, and the C rate coding unit 423 functions according to the coding rate selection command 405 from the outside. In addition, the switching control is performed, and information indicating the rate functioning by the switching control is output to the noise suppressor 410 as the coding rate selection information 406.

The noise suppressor 410 includes a noise suppression unit 415, a parameter table 416, and a parameter switching control unit 417.
The noise suppression unit 415 suppresses background noise included in the audio data output from the A / D converter 402, and its suppression characteristic is controlled by parameters input from the parameter table 416.

  The parameter table 416 is a table for storing parameters for setting the characteristics of the background noise suppression processing performed by the noise suppression unit 415. The A rate encoding unit 421, the B rate encoding unit 422, and the C rate encoding unit 423 are stored in the parameter table 416. The three parameter sets having optimum suppression characteristics for each coding rate are stored, and the parameter sets are input to the noise suppression unit 415 under the control of the parameter switching control unit 417.

  The parameter switching control unit 417 controls the parameter table 416, and based on the coding rate selection information 406, the optimum parameter for the coding unit (any one of 421, 422, and 423) that functions in the speech encoder 420 The set is selectively set in the noise suppression unit 415.

  In order to optimize the correspondence between the parameter setting (suppression characteristic setting) of the encoding unit and the noise suppression unit, for example, as a parameter set corresponding to the A rate encoding unit 421, the noise suppression amount is The parameter set is set so as to obtain a characteristic that suppresses noise as much as possible even if the sound part is slightly distorted. The parameter set corresponding to the C-rate encoding unit 423 has a light noise suppression amount. For this reason, the parameter set is set so as to obtain a characteristic of transmitting noise that can be heard naturally.

  In addition, the parameter set corresponding to the B rate encoding unit 422 is a parameter set that provides an intermediate characteristic between the characteristics for the A rate encoding unit 421 and the characteristics for the C rate encoding unit 423.

Next, the operation of the signal processing apparatus according to the fourth embodiment will be described. FIG. 8 is a flowchart for explaining this operation.
When a coding rate selection command 405 indicating “use A rate as the coding rate” is input to the coding rate switching control unit 424 from the command input standby state in step 8a, the step 8b Then, the designated system is determined, and since it is the A system, the process proceeds to Step 8c.

  In step 8c, the encoding rate switching control unit 424 performs switching control so that the digital data 403 is input to the A rate encoding unit 421, and in response to this, the A rate encoding unit 421 receives the input digital data. The encoding of 403 is started.

  In step 8c, in parallel with this switching control, the coding rate switching control unit 424 indicates that the A rate coding unit 421 is used for coding the digital data 403, and coding rate selection information 406. Is output to the parameter switching control unit 417, and the process proceeds to step 8d.

  In step 8d, the parameter switching control unit 417 sets the parameter set corresponding to the A rate encoding unit 421 so that the noise suppression characteristic of the noise suppression unit 415 becomes an optimum characteristic for encoding by the A rate encoding unit 421. Is input to the noise suppression unit 415 from the parameter table 416, and the process proceeds to step 8i.

  By such parameter setting (suppression characteristic setting) control operation, noise is suppressed in the output of the A / D converter 402 with a suppression characteristic suitable for encoding by the A rate encoding unit 421, and the digital data 403 is A The data is input to the rate encoding unit 421, encoded here, and output as encoded data 404 with data compression.

  In step 8i, when a coding rate selection command 405 indicating that “use B rate as a coding rate” is input to the coding rate switching control unit 424, the process proceeds to step 8b to specify the designated method. Since it is C system, it transfers to step 8e.

  In step 8e, the coding rate switching control unit 424 controls the switching so that the digital data 403 is input to the B rate coding unit 422 in a timed manner, and the A rate coding unit 421 responds accordingly. The operation stops, and encoding of the digital data 403 to which the B rate encoding unit 422 is input instead starts.

  In step 8e, in parallel with the switching control, the coding rate switching control unit 424 indicates that the B rate coding unit 422 is used for coding the digital data 403, and coding rate selection information 406. Is output to the parameter switching control unit 417, and the process proceeds to step 8f.

  In step 8f, the parameter switching control unit 417 sets the parameter set corresponding to the B rate encoding unit 422 so that the noise suppression characteristic of the noise suppression unit 415 becomes an optimum characteristic for encoding by the B rate encoding unit 422. Is input to the noise suppression unit 415 from the parameter table 416, and the process proceeds to step 8i.

  By such parameter setting (suppression characteristic setting) control operation, noise is suppressed in the output of the A / D converter 402 with suppression characteristics suitable for encoding by the B rate encoding unit 422, and B The data is input to the rate encoding unit 422, encoded here, and output as encoded data 404 with data compression.

  Further, in step 8i, when the digital data 403 is encoded by the A rate encoding unit 421 or the B rate encoding unit 422 as described above, “Use C rate as the encoding rate”. When the coding rate selection command 405 indicating is input to the coding rate switching control unit 424, the process proceeds to step 8b, the designated system is determined, and since it is the C system, the process proceeds to step 8g.

  In step 8g, the coding rate switching control unit 424 performs switching control so that the digital data 403 is input to the C rate coding unit 423 in a timely manner, and in response to this, the A rate coding in use. The operation of the unit 421 or the B rate encoding unit 422 is stopped, and encoding of the digital data 403 to which the C rate encoding unit 423 is input instead is started.

  In step 8g, in parallel with this switching control, the coding rate switching control unit 424 indicates that the C rate coding unit 423 is used for coding the digital data 403, and coding rate selection information 406. Is output to the parameter switching control unit 417, and the process proceeds to step 8h.

  In step 8h, the parameter switching control unit 417 sets the parameter set corresponding to the C rate encoding unit 423 so that the noise suppression characteristics of the noise suppression unit 415 are optimal characteristics for encoding by the C rate encoding unit 423. Is input to the noise suppression unit 415 from the parameter table 416, and the process proceeds to step 8i.

  By such parameter setting (suppression characteristic setting) control operation, noise is suppressed in the output of the A / D converter 402 with a suppression characteristic suitable for encoding by the C rate encoding unit 423, and the digital data 403 is C The data is input to the rate encoding unit 423, encoded here, and output as encoded data 404 that is data-compressed.

  If there is no command input in step 8i, the process proceeds to step 8j. In step 8j, it is determined whether or not a communication termination request has been made. If there is a termination request, the process is terminated. If there is no termination request, command input is again performed in step 8i. Monitor.

  As described above, in the signal processing apparatus having the above-described configuration, in order to obtain encoded data 404 with data compression, it is matched with the encoding unit (either 421, 422, or 423) that functions in the speech encoder 420. The parameters of the noise suppression unit 415 are varied to set the noise suppression characteristic of the noise suppression unit 415 to the optimum characteristic for the encoding process.

  Therefore, according to the signal processing device configured as described above, optimal noise suppression is performed in accordance with the encoding of the audio encoder 420, so that the noise suppressor exhibits a sufficient function and transmits high-quality audio. can do.

  The present invention is not limited to the above embodiment. For example, in the fourth embodiment, the parameter switching control unit 417 matches the coding unit functioning in the speech coder 420 based on the coding rate selection information 406 from the coding rate switching control unit 424. The noise suppression characteristic of the noise suppression unit 415 is changed to an optimum characteristic.

  Instead, for example, the parameter switching control unit 417 adjusts the noise suppression characteristic of the noise suppression unit 415 to the optimum characteristic in accordance with the encoding unit functioning in the speech encoder 420 based on the encoding rate selection command 405. Even if it is made variable, the same effect can be exhibited.

  In this case, the parameter switching control unit 417 performs control to set a parameter set that obtains an optimum noise suppression characteristic at a timing that considers the switching timing of the encoding unit in the speech encoder 420.

  Further, in the above first to fourth embodiments, the case where the transmitted voice is encoded has been described as an example, as summarized in FIG. In this figure, 1 is a microphone, 2 is an A / D converter, the noise suppressor 10 corresponds to the noise suppressors 110, 210, 310, 410, and the speech encoder 20 is the speech code. Corresponds to the generators 120, 220, 320, 420.

  By the way, in the first and second embodiments described above, the number of encoding methods (three, A method, B method, C method) and the number of noise suppression methods (three, X method, Y method, Z method). ) And the number of encoding methods and the number of parameter sets set in the noise suppression unit are the same.

  Also in the third and fourth embodiments, the number of encoding rates (three rates of A rate, B rate, and C rate) and the number of noise suppression methods (three of the X method, Y method, and Z method) and The description has been made assuming that the number of encoding rates is the same as the number of parameter sets set in the noise suppression unit.

  However, in implementing the present invention, the number of encoding schemes and the number of noise suppression schemes, and the number of encoding schemes and the number of parameter sets set in the noise suppression section may not be the same.

  Also, the number of encoding rates, the number of noise suppression schemes, and the number of encoding rates may not be the same as the number of parameter sets set in the noise suppression unit.

  The third embodiment will be described as an example. For example, as shown in FIG. 10, the speech encoder 320 includes eight types of encoding units (A rate, B rate, C rate,..., H rate). There may be only two noise suppression units, the X method and the Y method.

  In this case, for example, as shown in FIG. 11, in the case of an A-rate encoding unit 321 with poor sound quality, an X-type noise suppression unit 311 is used, and encoding units having other encoding rates are used. There is a method in which a Y-type noise suppression unit 312 is used.

  In addition, for example, X-type noise suppression is used for the A rate, B rate, C rate, and D rate, and Y-type noise suppression is used for the E rate, F rate, G rate, and H rate. It goes without saying that various forms of implementation are possible, such as using.

  In short, it is important to associate the coding rate with the noise suppression method used according to the coding rate (or the parameter set set for noise suppression control according to the coding rate) in advance. Thus, the present invention can realize various embodiments.

  In the first embodiment shown in FIG. 1, the number of encoding units having different schemes in the speech encoder 120 is P, and the number of noise suppressing units having different schemes in the noise suppressor 110 is Q (P , Q are both positive integers), P and Q need only have a relationship of P ≧ Q> 1.

  In the second embodiment shown in FIG. 3, the number of encoding units having different schemes in speech encoder 220 is P, and the number of parameters set in noise suppression unit 215 of noise suppressor 210. Is S (P and S are both positive integers), P and S need only have a relationship of P ≧ S> 1.

  Furthermore, in the third embodiment shown in FIG. 5, the number of encoding units having different encoding rates in the speech encoder 320 is R, and the number of noise suppressing units having different methods in the noise suppressor 310 is used. Is Q (R and Q are both positive integers), R and Q need only have a relationship of R ≧ Q> 1.

  In the fourth embodiment shown in FIG. 7, the number of encoding units having different encoding rates in the speech encoder 420 is R, and the number of parameters set in the noise suppression unit 415 of the noise suppressor 410. Is S (R and S are both positive integers), R and S need only have a relationship of R ≧ S> 1.

  Further, the application of the present invention is not limited to the case where such transmitted speech is encoded, and can also be applied to the case where encoded speech data as shown in FIG. 12 is decoded.

  In this figure, 3 is a speaker, 2 is a D / A converter, 40 is a decoder that selectively decodes audio data using a plurality of decoding methods or a plurality of encoding rates, The noise suppressor 30 performs optimum background noise suppression processing in accordance with the decoding processing of the speech decoder 40.

  As described above, even when the present invention is applied to the decoding side, it is possible to configure four modes as in the above-described encoding example. In each of these configurations, the decoding scheme and the encoding rate can be configured. Even when switching, the noise suppressor can function sufficiently and receive high-quality sound.

  Further, in both the third and fourth embodiments, three circuits as shown in the speech encoders 320 and 420 in FIGS. 5 and 7 are selectively used as encoders capable of changing the encoding rate. The configuration used is described as an example.

  However, the present invention is not limited to this, and the coding rate may be varied by varying the parameters of one coding unit 725 as in the speech coder 720 shown in FIG.

  In the configuration of this figure, a parameter set for performing encoding at a plurality of encoding rates is stored in the parameter table 726 in advance, and the encoding rate switching control unit 727 is configured in response to an external request. A parameter set according to the request is output from the parameter table 726 and set in the encoding unit 725.

Even when the speech coder 720 having such a configuration is used, as in the third and fourth embodiments, the noise suppressor exhibits a sufficient function and can transmit high-quality speech.
In addition, any of the encoding units 121 to 123 and 221 to 223 constituting the speech encoders 120 and 220 in FIGS. 1 and 3 may be a speech encoder 720 as shown in FIG. Is possible.

  Although there is a difference between encoding and decoding, it is also possible to apply the configuration shown in FIG. 13 to the configuration on the decoding side shown in FIG. 12, and even in this case, the decoding scheme and encoding rate are also applicable. Even when switching between, the noise suppressor exhibits a sufficient function and can receive high-quality sound.

  Furthermore, in the fourth embodiment, the noise suppressor 410 is configured such that when specific coding rate information is detected, noise suppression is turned off in all frequency bands or some frequency bands (noise suppression is suppressed). It is also possible to adopt a configuration that does not.

FIG. 14 shows a configuration in which such noise suppression is applied to the noise suppressor 410, which will be described in detail below with reference to this figure.
Here, a description will be given by taking as an example a method of dividing a speech signal into M frequency bands and performing noise suppression by a spectral subtraction method (SS method). The value of M is usually about 6 to 32, although it varies depending on the noise suppression method.

  The parameter switching control unit 417 detects the coding rate of coding used in the speech coder 420 from the coding rate selection information 406, and the parameter table 416 corresponding to this rate according to the detected coding rate. Are output from the parameter table 416 to the band-by-band suppression coefficient calculation unit 460 of the noise suppression unit 415.

  At this time, the parameter set input from the parameter table 416 to the band-specific suppression coefficient calculation unit 460 includes L control parameters. At this time, when controlling noise suppression to the same extent in all frequency bands, L = 1 control parameter is output for one coding rate.

  On the other hand, in order to control different noise suppression for each of M frequency bands, L = M control parameters are generated for one coding rate. The value of L is not limited to this.

  Here, for simplification of the following description, there are three types of coding rates, A, B, and C, and L = M. At this time, the control parameters for the coding rate A can be expressed as C (A, 0), C (A, 1),..., C (A, M−1).

  C (A, k) represents a control parameter for controlling the K-th band among the M frequency bands divided with respect to the coding rate A. A summary of this is shown in FIG.

  As described in the fourth embodiment, the noise suppression unit 415 performs noise suppression of the input signal according to the control parameter from the parameter table 416. The noise suppression unit 415 includes an FFT unit 440, a band-specific noise amount estimation unit 450, A band-specific suppression coefficient calculation unit 460, a noise suppression unit 470, and an inverse FFT unit 480 are included.

  The FFT unit 440 converts the input audio signal from the time domain to the frequency domain by FFT (Fast Fourier Transform). In addition, other conversion methods such as DFT and DCT can be used as the conversion method to the frequency domain.

  The band-specific noise amount estimation unit 450 divides the audio signal converted into the frequency domain into predetermined M bands, and estimates the noise amount included in the audio signal for each band. The band-by-band suppression coefficient calculation unit 460 calculates the band-by-band noise suppression coefficient based on the band-by-band noise amount estimated by the band-by-band noise amount estimation unit 450.

  Here, it is assumed that the noise suppression coefficients for each band are D (0), D (1),..., D (M−1). D (k) represents a noise suppression coefficient used for controlling the kth band among the M frequency bands.

  In the present invention, in addition to the noise suppression coefficient obtained only by analyzing the input signal, the noise suppression process is controlled using a control parameter obtained from the coding rate information. One way to achieve this is to set the control parameter so that the control parameter multiplied by the noise suppression coefficient can be used as the new noise suppression coefficient.

  Specifically, for example, by performing the following operations, the noise suppression coefficient D (k) is updated using the control parameter C (k) obtained from the coding rate information. The updated noise suppression coefficient D (k) is output to the noise suppression unit 470.

D (k) ← D (k) × C (k) (k = 0,..., M−1)
Noise suppression section 470 multiplies 1-D (k) for each band by using the updated suppression coefficient obtained by band-specific suppression coefficient calculation section 460 for the frequency domain speech spectrum obtained from the input speech signal. Thus, a speech spectrum with noise suppression is generated. The inverse FFT unit 480 converts the audio spectrum generated by the noise suppression unit 470 into an audio signal in the time domain.

  Here, for example, in order to turn noise suppression OFF (no noise suppression) in all frequency bands at the highest coding rate C, the bit rate C is detected as shown in FIG. All the band-specific control parameters used at times may be set to “0”.

In order to turn off noise suppression only in the frequency band M-1 at the coding rate B (no noise suppression is performed), the band M used when the coding rate B is detected as shown in FIG. The control parameter for -1 may be set to “0”.
Needless to say, according to such a configuration, various other settings are possible.

  In this way, by controlling the noise suppression processing performed by the noise suppression unit 415 using the control parameters generated from the coding rate information, the total balance between noise suppression and variable rate speech coding is taken into consideration. The variable rate speech processing apparatus can be realized.

  It is a well-known fact that known noise suppression processing cannot completely remove only noise contained in an input voice signal. If the noise is completely removed, part of the audio signal will be removed together with the noise. It causes deterioration that is lost.

  On the other hand, for speech coding, a clean speech signal that contains as little noise as possible has better sound quality due to coding because the coding analysis works better and the coding model fits. It is known that degradation is generally small.

  Conversely, if an audio signal that contains a lot of background noise is encoded, especially in low-rate audio encoding, the encoding of the non-audio portion will be significantly degraded. Encoding improves the quality.

  On the other hand, in speech coding with a high coding rate, even if the speech contains a lot of background noise, the performance of the coding itself is high, so the degradation of sound quality due to background noise is small, and the sound quality is relatively natural. Is obtained.

  In this way, a certain amount of noise suppression tends to have a good effect on the total speech quality when speech coding at a low coding rate is used. In the case of using the normalization, normal noise suppression is not necessarily required in an application that requires a high natural sound quality.

  In such a case, noise suppression is performed in all frequency bands by setting all the band-specific control parameters to “0” at the specific coding rate (not limited to one coding rate) described in FIG. Is effective (no noise suppression is performed).

  Therefore, by using the noise suppressor 420 shown in FIG. 14, the noise suppression function can be controlled more flexibly than in the past in accordance with the coding rate, so that variable rate speech can be used in a real environment where a lot of background noise is mixed. The voice quality when using the processing device can be improved.

  Here, the case where the present invention is applied to the noise suppressor 410 configured to vary the encoding rate on the encoding side as shown in FIG. 7 has been described, but noise that performs noise suppression in accordance with the encoding rate on the decoding side is described. It is possible to apply to a suppressor, or to a noise suppressor that performs noise suppression according to an encoding method or a decoding method, and it goes without saying that the same effect can be obtained in this case.

  Furthermore, in the fourth embodiment, the noise suppressor 410 is configured as a noise suppressor 411 as shown in FIG. 18, and the noise suppression is forcibly turned off according to an external request regardless of the encoding rate. A configuration (without noise suppression) is also possible.

  In the figure, the noise suppression unit 415, the parameter table 416, and the parameter switching control unit 417 are the same as those described in the fourth embodiment, and thus the description thereof is omitted here. Here, the newly provided ON / OFF information detector 419 and changeover switch 418 will be described.

  The ON / OFF information detection unit 419 detects / determines information for instructing ON / OFF of the function for suppressing noise from the outside, and controls the changeover switch 418 according to the determination result.

  That is, when an instruction to turn on the noise suppression function is detected, the audio data output from the A / D converter 402 is switched to be input to the noise suppression unit 415, while the noise is suppressed. When the instruction to turn off the function to be detected is detected, the audio data output from the A / D converter 402 is input to the subsequent audio encoder 420 as the digital data 403 without passing through the noise suppression unit 415. Are switched as follows.

  As an example of externally controlling ON / OFF of the function for suppressing noise, there is ON / OFF control to the noise suppressor 411 from the communication network side. One of the reasons why such control is necessary is that noise suppression is performed twice in a communication path where there is a so-called tandem connection in which encoding / decoding is performed twice between the receiving side and the transmitting side. This is to prevent them from becoming tandem.

  The essence of the control performed to prevent the tandem connection is not to turn noise suppression ON (enabled) but to turn noise suppression OFF (disabled). In consideration of this point, when there is ON / OFF control of a function for suppressing noise from the outside such as from the communication network side, a combination of controls that makes use of both intentions is possible.

  Here, the case where the present invention is applied to the noise suppressor 410 configured to vary the encoding rate on the encoding side as shown in FIG. 7 has been described, but noise that performs noise suppression in accordance with the encoding rate on the decoding side is described. It is possible to apply to a suppressor, or to a noise suppressor that performs noise suppression according to an encoding method or a decoding method, and it goes without saying that the same effect can be obtained in this case.

Next explained is a signal processing apparatus according to the fifth embodiment of the invention. FIG. 19 shows the configuration.
The voice input unit 540 converts the user's transmitted voice into an electric signal, captures it, and digitizes it to obtain voice data. The hands-free microphone 541, the hands-free microphone amplifier 542, and the non-hands-free microphone 543, a non-hands-free microphone amplifier 544, a microphone switching control unit 545, and an A / D converter 546.

  The microphone switching control unit 545 switches between the above-described analog system for hands-free and analog system for non-hands-free according to a control command 553 for switching between hands-free and non-hands-free.

  The A / D converter 546 receives an analog audio signal obtained by any one of the above-described analog systems by the switching control of the microphone switching control unit 545, digitizes the analog audio signal, and outputs it as audio data.

  In non-hands-free, since the direction and distance of voice are almost determined, a microphone having microphone sensitivity and directivity corresponding to this is used. On the other hand, when hands-free, it is necessary to raise the microphone sensitivity in advance so that it can pick up transmitted speech from a distance, and the direction of voice arrival is also uncertain compared to non-hands-free, so directivity It is necessary to widen the angle. For this reason, the characteristics of the analog audio signal input to the A / D converter 546 differ depending on hands-free / non-hands-free.

  The echo control unit 530 includes a hands-free echo control unit 531, a non-hands-free echo control unit 532, and an echo switching control unit 533.

  The hands-free echo control unit 531 is suitable when the hands-free microphone 541 and the hands-free microphone amplifier 542 are used, and reduces echoes superimposed on the audio data output from the A / D converter 546.

  On the other hand, the non-hands-free echo control unit 532 is suitable when the non-hands-free microphone 543 and the non-hands-free microphone amplifier 544 are used, and reduces echoes superimposed on the audio data output from the A / D converter 546. To do. However, when echo suppression is not required, audio data is output without being echo-controlled.

  In response to a control command 553 for switching between hands-free and non-hands-free, the echo switching control unit 533 performs A / D on one of the hands-free echo control unit 531 and the non-hands-free echo control unit 532. By inputting the audio data output from the converter 546, control is performed to make one function.

  With this control, the audio data 551 whose echo has been reduced by the hands-free echo control unit 531 or the non-hands-free echo control unit 532 is output to the noise suppressor 510.

  The noise suppressor 510 includes a suppression method switching control unit 514 in addition to the X-type noise suppression unit 511 and the Y-type noise suppression unit 512 as two circuits for suppressing background noise using different algorithms.

  The X-type noise suppression unit 511 is suitable for noise suppression with respect to the audio data 551 generated through the hands-free microphone 541, the hands-free microphone amplifier 542, and the hands-free echo control unit 531 used during the hands-free operation. Is set.

  On the other hand, the Y-type noise suppression unit 512 applies to the audio data 551 generated through the non-hands-free microphone 543, the non-hands-free microphone amplifier 544, and the non-hands-free echo control unit 532, which are used when the hands are free. It is set to be suitable for noise suppression.

  In response to a control command 553 for switching between hands-free and non-hands-free, the suppression method switching control unit 514 sends audio data 551 to either one of the X method noise suppression unit 511 and the Y method noise suppression unit 512. Is input to control one of them.

Next, the operation of the signal processing apparatus of the fifth embodiment will be described. FIG. 20 is a flowchart for explaining this operation.
If a control command 553 indicating “execute hands-free” is input from the command input standby state in step 20a, the process proceeds to step 20b to determine the designated content and hands-free. Since the execution instruction is, the process proceeds to step 20c.

  In step 20 c, the microphone switching control unit 545 starts switching control so that an analog audio signal input through the hands-free microphone 541 and the hands-free microphone amplifier 542 is input to the A / D converter 546.

  In step 20c, in parallel with this switching control, the echo switching control unit 533 sends the audio data output from the A / D converter 546 to the hands-free echo control unit 531 according to the control command 553. Then, the process proceeds to step 20d.

  In step 20d, the suppression method switching control unit 514 inputs the audio data output from the hands-free echo control unit 531 to the X-type noise suppression unit 511 in response to the control command 553, and proceeds to step 20g.

  Therefore, when a hands-free instruction is given, the user's transmitted voice input through the hands-free microphone 541 is transferred to the hands-free echo control unit 531 by the switching control operation. Echo control suitable for use of the microphone 541 and the hands-free microphone amplifier 542 is performed.

  The echo-controlled audio data 551 is subjected to an optimum noise suppression process when the hands-free microphone 541, the hands-free microphone amplifier 542, and the hands-free echo control unit 531 are used by the X-type noise suppression unit 511. Then, it is output as transmission audio data 552 to the subsequent transmission unit.

  After that, when a control command 553 indicating “Stop Hands-Free” is input in Step 20g, the process proceeds to Step 20b, where the designated content is determined, and it is a hands-free stop instruction. The process proceeds to step 20e.

  In step 20e, the microphone switching control unit 545 starts switching control so that the analog audio signal input through the non-hands-free microphone 543 and the non-hands-free microphone amplifier 544 is input to the A / D converter 546. .

  In step 20e, in parallel with this switching control, the echo switching control unit 533 converts the audio data output from the A / D converter 546 in accordance with the control command 553 to the non-hands-free echo control unit 532. To step 20f.

  In step 20f, the suppression method switching control unit 514 inputs the audio data output from the non-hands-free echo control unit 532 to the Y method noise suppression unit 512 in response to the control command 553, and proceeds to step 20g. .

  Therefore, when a hands-free stop instruction is given, the user's transmitted voice input through the non-hands-free microphone 543 is transferred to the non-hands-free echo control unit 532 by such a switching control operation. Echo control suitable for use of the non-hands-free microphone 543 and the non-hands-free microphone amplifier 544 is performed.

  Then, the echo-controlled audio data 551 is subjected to noise suppression that is optimal when the non-hands-free microphone 543, the non-hands-free microphone amplifier 544, and the non-hands-free echo control unit 532 are used by the Y-type noise suppression unit 512. Processing is performed and output as transmission audio data 552 to a subsequent transmission unit.

  If there is no command input in step 20g, the process proceeds to step 20h. In step 20h, it is determined whether or not a request for termination of communication has been made. If there is a termination request, the process is terminated. On the other hand, if there is no termination request, command input is again performed in step 20g. Monitor.

  As described above, in the signal processing apparatus having the above-described configuration, in order to obtain the transmission voice data 552 subjected to echo control and noise suppression, an optimum noise suppression unit (in accordance with a generation path of hands-free / non-hands-free voice data) 511 or 512) is made to function.

  Therefore, according to the signal processing apparatus configured as described above, even when hands-free / non-hands-free switching is performed, noise suppression is performed by the noise suppression unit suitable for the sound data generation path, that is, the characteristics of the sound data. The suppressor can perform a sufficient function and transmit high-quality sound.

Next explained is a signal processing apparatus according to the sixth embodiment of the invention. FIG. 21 shows the configuration.
The voice input unit 640 converts a user's transmitted voice into an electrical signal, captures it, and digitizes it to obtain voice data. The hands-free microphone 641, hands-free microphone amplifier 642, and non-hands-free microphone 643, a non-hands-free microphone amplifier 644, a microphone switching control unit 645, and an A / D converter 646.

  The microphone switching control unit 645 switches between the above-described hands-free analog system and non-hands-free analog system in response to a control command 653 that switches between hands-free and non-hands-free.

  The A / D converter 646 receives an analog voice signal obtained by any one of the above-described analog systems by the switching control of the microphone switching control unit 645, digitizes the analog voice signal, and outputs it as voice data.

  In non-hands-free, since the direction and distance of voice are almost determined, a microphone having microphone sensitivity and directivity corresponding to this is used. On the other hand, when hands-free, it is necessary to raise the microphone sensitivity in advance so that it can pick up transmitted speech from a distance, and the direction of voice arrival is also uncertain compared to non-hands-free, so directivity It is necessary to widen the angle. For this reason, the characteristics of the analog audio signal input to the A / D converter 646 differ depending on the hands-free / non-hands-free.

  The echo control unit 630 includes a hands-free echo control unit 631, a non-hands-free echo control unit 632, and an echo switching control unit 633.

  The hands-free echo control unit 631 is suitable when the hands-free microphone 641 and the hands-free microphone amplifier 642 are used, and reduces echoes superimposed on the audio data output from the A / D converter 646.

  On the other hand, the non-hands-free echo control unit 632 is suitable when the non-hands-free microphone 643 and the non-hands-free microphone amplifier 644 are used, and reduces echoes superimposed on the audio data output from the A / D converter 646. To do. However, when echo suppression is not required, audio data is output without being echo-controlled.

  The echo switching control unit 633 performs A / D on one of the hands-free echo control unit 631 and the non-hands-free echo control unit 632 according to a control command 653 for switching between hands-free and non-hands-free. By inputting the audio data output from the converter 646, control is performed to make one function.

  By this control, the audio data 651 whose echo is reduced by the hands-free echo control unit 631 or the non-hands-free echo control unit 632 is output to the noise suppressor 610.

The noise suppressor 610 includes a noise suppression unit 615, a parameter table 616, and a parameter switching control unit 617.
The noise suppression unit 615 suppresses background noise included in the audio data output from the echo control unit 630, and the suppression characteristic is controlled by parameters input from the parameter table 616.

  The parameter table 616 stores parameters for setting the characteristics of the background noise suppression processing performed by the noise suppression unit 615. The parameter set A is optimal for hands-free and the optimal parameter set B for non-hands-free. The parameter set is input to the noise suppression unit 615 under the control of the parameter switching control unit 617.

  That is, the parameter set A uses the characteristics of the noise suppression unit 615 based on the voice data 651 generated through the hands-free microphone 641, the hands-free microphone amplifier 642, and the hands-free echo control unit 631, which are used during hands-free operation. The characteristic is suitable for noise suppression.

  The parameter set B is generated through the non-hands-free microphone 643, the non-hands-free microphone amplifier 644, and the non-hands-free echo control unit 632, which are used when the hands-free operation is performed. The characteristic is suitable for noise suppression for the voice data 651.

  The parameter switching control unit 617 controls the parameter table 616 to selectively select a parameter set optimal for noise suppression processing of the audio data 651 in accordance with a control command 653 for switching between hands-free and non-hands-free. 615 is set.

Next, the operation of the signal processing apparatus according to the sixth embodiment will be described. FIG. 22 is a flowchart for explaining this operation.
When a control command 653 indicating that “execute hands-free” is input from the command input standby state in step 22a, the process proceeds to step 22b to determine the designated contents and hands-free. Since the execution instruction is, the process proceeds to step 22c.

  In step 22 c, the microphone switching control unit 645 starts switching control so that an analog audio signal input through the hands-free microphone 641 and the hands-free microphone amplifier 642 is input to the A / D converter 646.

  In step 22c, in parallel with this switching control, the echo switching control unit 633 sends the audio data output from the A / D converter 646 to the hands-free echo control unit 631 in response to the control command 653. Then, the process proceeds to step 22d.

  In step 22d, the parameter switching control unit 617 controls the parameter table 616 in accordance with the control command 653, sets the optimum parameter set A in the noise suppression unit 615 during hands-free operation, and proceeds to step 22g.

  Therefore, when a hands-free instruction is given, the user's transmitted voice input through the hands-free microphone 641 is transferred to the hands-free echo control unit 631 by the switching control operation. Echo control suitable for use of the microphone 641 and the hands-free microphone amplifier 642 is performed.

  Then, the echo-controlled audio data 651 is optimally used when the hands-free microphone 641, the hands-free microphone amplifier 642, and the hands-free echo control unit 631 are used by the noise suppression unit 615 in which the parameter set A is set. Noise suppression processing is performed and output as transmission audio data 652 to a subsequent transmission unit.

  After that, when a control command 653 indicating “Stop Hands-Free” is input in Step 22g, the process proceeds to Step 22b, where the designated content is determined, and it is a hands-free stop instruction. Then, the process proceeds to step 22e.

  In step 22e, the microphone switching control unit 645 starts switching control so that the analog audio signal input through the non-hands-free microphone 643 and the non-hands-free microphone amplifier 644 is input to the A / D converter 646. .

  In step 22e, in parallel with this switching control, the echo switching control unit 633 converts the audio data output from the A / D converter 646 into the non-hands-free echo control unit 632 in accordance with the control command 653. To step 22f.

  In step 22f, the parameter switching control unit 617 controls the parameter table 616 in accordance with the control command 653, sets the optimum parameter set B in the noise suppression unit 615 in the non-hands-free state, and proceeds to step 22g. .

  Therefore, when a hands-free stop instruction is given, the user's transmitted voice input through the non-hands-free microphone 643 is transmitted by the non-hands-free echo control unit 632 by such a switching control operation. Echo control suitable for use of the non-hands-free microphone 643 and the non-hands-free microphone amplifier 644 is performed.

  The audio data 651 subjected to echo control is used by the non-hands-free microphone 643, the non-hands-free microphone amplifier 644, and the non-hands-free echo control unit 632 by the noise suppression unit 615 in which the parameter set B is set. In some cases, optimal noise suppression processing is performed and output as transmission audio data 652 to a subsequent transmission unit.

  If there is no command input in step 22g, the process proceeds to step 22h. In step 22h, it is determined whether or not an end request for communication has been made. If there is an end request, the process is terminated. On the other hand, if there is no end request, command input is again made in step 22g. Monitor.

  As described above, in the signal processing apparatus having the above-described configuration, when obtaining the transmission voice data 652 subjected to echo control and noise suppression, an optimum noise suppression characteristic and an optimum noise suppression characteristic are selected according to the generation path of the hands-free / non-hands-free voice data. Thus, the characteristics of the noise suppressor 615 are controlled.

  Therefore, according to the signal processing apparatus configured as described above, even when hands-free / non-hands-free switching is performed, noise suppression is performed by the noise suppression unit suitable for the sound data generation path, that is, the characteristics of the sound data. The suppressor can perform a sufficient function and transmit high-quality sound.

  The present invention is not limited to the above embodiment. For example, in each of the above-described embodiments, the noise suppressor, the voice encoder (decoder), the echo control unit, and the like have been described as separate circuits. However, in each embodiment, these are integrated into one chip. It can also be implemented on one or more DSP chips.

Alternatively, for example, using a high-speed processor and memory, a program that performs functions such as a noise suppressor, a speech coder (decoder), and an echo control unit is stored in the memory. Needless to say, this can be realized by operating the.
In addition, it goes without saying that the present invention can be similarly implemented even if various modifications are made without departing from the gist of the present invention.

10, 30, 110, 310, 510 ... noise suppressor 20, 120, 220, 320, 420, 720 ... speech encoder 40 ... speech decoder 111-113, 311-313, 511, 512 ... noise suppressor 114, 314, 514 ... suppress system switching control unit 121-123, 221-223, 725 ... encoding unit 124, 224 ... encoding system switching control unit 210, 410, 411, 610 ... noise suppressor 215, 415, 615 ... noise suppressor Units 216, 416, 616, 726 ... Parameter tables 217, 417, 617 ... Parameter switching control units 321-323, 421-423 ... Rate coding units 324, 424, 727 ... Coding rate switching control unit 419 ... ON / OFF Information detection unit 440 ... FFT unit 450 ... Band according to band Quantity estimation unit 460 ... Band-by-band suppression coefficient calculation unit 470 ... Noise suppression unit 480 ... Inverse FFT unit 530, 630 ... Echo control unit 531, 631 ... Hands-free echo control unit 532, 632 ... Non-hands-free echo control unit 533 , 633 ... Echo switching control section 540, 640 ... Audio input section 541, 641 ... Hands-free microphone 542, 642 ... Hands-free microphone amplifier 543, 643 ... Non-hands-free microphone 544, 644 ... Non-hands-free microphone amplifier 545, 645 ... Microphone switching control unit.

Claims (29)

Noise suppression means for suppressing a noise component included in the audio signal, and encoding means for generating compressed encoded audio data by performing encoding processing on the audio signal in which the noise component is suppressed by the noise suppression means A signal processing device comprising:
The encoding means can perform a plurality of different encoding processes,
The signal processing apparatus, wherein the noise suppression unit has a noise suppression characteristic according to an encoding process performed by the encoding unit.   The signal processing apparatus according to claim 1, wherein the plurality of different encoding processes in the encoding unit are encoding processes using different encoding methods.   The signal processing apparatus according to claim 1, wherein the plurality of different encoding processes in the encoding unit are processes that perform encoding at different encoding rates.   The noise suppression unit does not suppress a noise component in at least a part of a frequency band in accordance with an encoding process selectively performed by the encoding unit. Item 4. The signal processing device according to any one of Items 3 to 4.   The signal processing apparatus according to claim 1, wherein the noise suppression unit does not suppress noise components in response to a request.   The noise suppression unit includes a plurality of noise suppression units having noise component suppression characteristics different from each other, and the noise suppression unit performs the noise processing according to an encoding process selectively performed by the encoding unit on a speech signal. 6. The signal processing apparatus according to claim 1, wherein a noise component is suppressed by selectively using a suppress unit. The noise suppression means includes
A noise suppressor that suppresses noise components for audio signals, and that can vary the noise component suppression characteristics according to parameter settings;
6. The apparatus according to claim 1, further comprising: a parameter setting unit that sets, in the noise suppression unit, a parameter corresponding to an encoding process selectively performed by the encoding unit. The signal processing apparatus as described. A signal processing apparatus comprising: decoding means for decoding compressed encoded audio data into an audio signal; and noise suppression means for suppressing a noise component contained in the audio signal decoded by the decoding means,
The decoding means can perform a plurality of different decoding processes according to the decoded audio data,
The signal processing apparatus, wherein the noise suppression means has a noise suppression characteristic according to a decoding process performed by the decoding means.   The signal processing apparatus according to claim 8, wherein the plurality of different decoding processes in the decoding unit are processes that perform decoding by different decoding methods.   The signal processing apparatus according to claim 8, wherein the plurality of different decoding processes in the encoding unit are processes for decoding encoded audio data having different encoding rates.   11. The noise suppression unit does not suppress noise components in at least some frequency bands in accordance with a decoding process selectively performed by the decoding unit. The signal processing device according to any one of the above.   The signal processing apparatus according to claim 8, wherein the noise suppression unit does not suppress a noise component in response to a request.   The noise suppression unit includes a plurality of noise suppression units having noise component suppression characteristics different from each other, and according to a decoding process selectively performed by the decoding unit on a decoded speech signal, 13. The signal processing apparatus according to claim 8, wherein the noise component is suppressed by selectively using a noise suppression unit. The noise suppression means includes
A noise suppressor that performs noise component suppression on the decoded speech signal, and a noise suppressor that can vary the noise component suppression characteristics in accordance with parameter settings;
The parameter setting means which sets the parameter according to the decoding process selectively implemented in the said decoding means in the said noise suppression part is provided. Signal processing device. In a signal processing device used for equipment that can selectively use the hands-free function,
A signal processing apparatus comprising noise suppression means for suppressing noise components in accordance with whether or not the input voice signal is voice input using the hands-free function.   The signal processing according to claim 15, wherein the noise suppression unit does not suppress a noise component in at least a part of a frequency band depending on whether or not the voice input uses the hands-free function. apparatus.   The signal processing apparatus according to claim 15 or 16, wherein the noise suppression unit does not suppress a noise component in response to a request.   The noise suppression means includes a plurality of noise suppression units having suppression characteristics of different noise components, and the noise is controlled depending on whether or not the input audio signal is an audio input using the hands-free function. The signal processing apparatus according to claim 15, wherein a noise component is suppressed by selectively using a suppress unit. The noise suppression means includes
A noise suppressor that performs noise component suppression on the input audio signal, and that can vary the noise component suppression characteristics in accordance with parameter settings; and
18. The signal according to claim 15, further comprising: a parameter setting unit that sets, in the noise suppression unit, a parameter according to whether or not voice input using the hands-free function is performed. Processing equipment. A noise suppression step for suppressing a noise component included in the audio signal, and an encoding step for generating compressed encoded audio data by performing encoding processing on the audio signal in which the noise component is suppressed in the noise suppression step A signal processing method comprising:
The encoding step selectively performs a plurality of different encoding processes,
The signal processing method characterized in that the noise suppression step suppresses noise components corresponding to the encoding process selectively performed in the encoding step on the speech signal. A signal processing method comprising: a decoding step of decoding compressed encoded audio data into an audio signal; and a noise suppression step of suppressing a noise component included in the audio signal decoded by the decoding step,
The decoding step selectively performs a plurality of different decoding processes on the encoded audio data according to the encoded audio data,
The signal processing method characterized in that the noise suppression step suppresses a noise component corresponding to a decoding process selectively performed in the decoding step on the audio signal. In a signal processing method used for a device that can selectively use a hands-free function,
A signal processing method comprising: a noise suppression step of suppressing a noise component according to whether or not the input voice signal is a voice input using the hands-free function. A noise suppression procedure for suppressing a noise component included in a speech signal, and a coding procedure for generating encoded speech data compressed by performing a coding process on the speech signal in which the noise component is suppressed by the noise suppression procedure. A computer-readable recording medium storing a program for causing a computer to execute functions,
The encoding procedure is a procedure for selectively performing a plurality of different encoding processes,
The recording medium according to claim 1, wherein the noise suppression procedure is a procedure for suppressing a noise component according to an encoding process selectively performed in the encoding procedure on an audio signal. A computer-readable recording of a program for causing a computer to execute a decoding procedure for decoding compressed encoded voice data into a voice signal and a noise suppression unit for suppressing a noise component included in the voice signal decoded by the decoding procedure A possible recording medium,
The decoding procedure is a procedure for selectively performing a plurality of different decoding processes on the encoded audio data according to the encoded audio data,
The recording medium according to claim 1, wherein the noise suppression procedure is a procedure for suppressing a noise component corresponding to a decoding process selectively performed in the decoding procedure with respect to an audio signal. A procedure for selectively enabling the hands-free function;
A computer-readable recording medium having recorded thereon a program for causing a computer to execute a noise suppression procedure for suppressing a noise component in accordance with whether or not the voice input using the hands-free function is performed on the input audio signal . Noise suppression means for suppressing a noise component included in the audio signal, and encoding means for generating compressed encoded audio data by performing encoding processing on the audio signal in which the noise component is suppressed by the noise suppression means A signal processing device comprising:
The encoding means selectively performs encoding processing by a plurality of different encoding methods,
The noise suppression unit includes a plurality of noise suppression units having noise component suppression characteristics different from each other, and the plurality of noise suppression units according to an encoding process selectively performed by the encoding unit on a speech signal. The noise component is suppressed by selectively using one noise suppressor from among the noise suppressors,
When the number of different encoding methods in the encoding means is P and the number of noise suppression units in the noise suppression means is Q (P and Q are both positive integers), P and Q are P ≧ Q A signal processing apparatus having a relationship of> 1. Noise suppression means for suppressing a noise component included in the audio signal, and encoding means for generating compressed encoded audio data by performing encoding processing on the audio signal in which the noise component is suppressed by the noise suppression means A signal processing device comprising:
The encoding means selectively performs encoding processing at a plurality of different encoding rates,
The noise suppression unit includes a plurality of noise suppression units having noise component suppression characteristics different from each other, and the plurality of noise suppression units according to an encoding process selectively performed by the encoding unit on a speech signal. The noise component is suppressed by selectively using one noise suppressor from among the noise suppressors,
When the number of the plurality of different encoding rates in the encoding unit is R and the number of noise suppression units of the noise suppression unit is Q (R and Q are both positive integers), R and Q are R ≧ Q. A signal processing apparatus having a relationship of Q> 1. Noise suppression means for suppressing a noise component included in the audio signal, and encoding means for generating compressed encoded audio data by performing encoding processing on the audio signal in which the noise component is suppressed by the noise suppression means A signal processing device comprising:
The encoding means selectively performs encoding processing by a plurality of different encoding methods,
The noise suppression unit performs noise component suppression on a speech signal, and is selectively implemented by a noise suppressor capable of varying noise component suppression characteristics according to parameter settings and the encoding unit. Parameter setting means for setting a parameter corresponding to the encoding processing in the noise suppression unit,
When P is the number of different encoding methods in the encoding means and S is the number of parameter sets to be set in the noise suppression unit of the noise suppression means (P and S are both positive integers), P and S Is a signal processing apparatus characterized by P ≧ S> 1. Noise suppression means for suppressing a noise component included in the audio signal, and encoding means for generating compressed encoded audio data by performing encoding processing on the audio signal in which the noise component is suppressed by the noise suppression means A signal processing device comprising:
The encoding means selectively performs encoding processing at a plurality of different encoding rates;
The noise suppression unit performs noise component suppression on a speech signal, and is selectively implemented by a noise suppressor capable of varying noise component suppression characteristics according to parameter settings and the encoding unit. Parameter setting means for setting a parameter corresponding to the encoding processing in the noise suppression unit,
When the number of the plurality of different encoding rates in the encoding unit is R, and the number of parameter sets to be set in the noise suppression unit of the noise suppression unit is S (R and S are both positive integers), R and S is a signal processing apparatus characterized in that R ≧ S> 1.

Images