JPH0968997A - Audio processing method and apparatus - Google Patents

Abstract

(57) [Abstract] [PROBLEMS] To perform good amplification without lowering clarity when performing voice amplification processing. An input speech signal is divided into at least a first band containing a pitch frequency component and a first formant component, and a second band containing a second formant component, and a minimum audibility for each band. The value and the discomfort threshold value are set so that the input sound having a level within a predetermined range is amplified to an output level set between the set minimum audible value and the discomfort threshold value.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a hearing aid, a telephone,
The present invention relates to a voice processing method and apparatus suitable for being applied to a device that performs various voice signal processes such as a loudspeaker and a communication device, and particularly to a voice processing technique for converting a voice amplitude of a voice signal.

[0002]

2. Description of the Related Art Conventionally, when audio is reproduced by various audio output devices such as hearing aids, the audio signal is amplified so that the output audio can be clearly heard. By performing and outputting such a process, for example, a voice that is easily heard by a hearing-impaired person or an elderly person whose hearing is deteriorated is output.

In the case of this amplification processing, if the dynamic range of the device is not sufficiently wide, the output sound after the amplification processing will be distorted. For example, in the case of a hearing aid used by a hearing-impaired person, since the dynamic range that can be heard is narrow, it is necessary to suppress the dynamic range of the output sound.When a large-amplitude sound is input, processing to suppress the peak, A process of compressing a large amplitude is performed, and the output voice is distorted.

11 and 12 are diagrams showing an example of amplification processing in a conventional hearing aid. A of FIG. 11 shows a waveform when a voice pronounced "mouse" is input, and A of FIG.
[Fig. 4] is an enlarged view showing a vowel portion at a portion pronounced "ne". 11B shows a waveform when the input waveform pronounced as "rat" is amplified by a hearing aid having a narrow dynamic range, and B in FIG. 12 is pronounced as "ne". It is a figure which expands and shows the vowel part of a location. As can be seen from FIGS. 11 and 12,
The input voice above a certain level is suppressed to a constant peak level.

By thus suppressing the peak level to a constant value, for example, in the case of a hearing aid, the input sound can be suppressed within a dynamic range that can be heard by a hearing-impaired person.

[0006]

However, if the peak level is suppressed or the large amplitude is compressed in this way, the clarity of the amplified voice is lowered and the language cannot be discriminated from the output voice. Will end up.

Further, when this type of device is used, for example, in a place where the noise level is high, the noise is greatly amplified, which makes it difficult to hear the voice itself. In such a case, for example, in a device such as a hearing aid, it may be possible to reduce the volume to reduce the noise. However, when the volume is reduced, the volume of the voice itself is also reduced, making it difficult to hear the voice. turn into.

Further, when the output voice itself does not include noise, as in the case of listening to the output voice of the telephone at a place where the noise level is high, the volume is adjusted to a large degree to output a large voice. However, if it is too large, the intelligibility of the voice is lowered as described above, and it is difficult to always adjust to an appropriate level by manual operation.

In view of the above points, the present invention has an object to enable good amplification without lowering the clarity when performing this kind of audio amplification processing.

[0010]

A voice processing apparatus of the present invention amplifies an input voice having a level within a predetermined range to an output level set between a set minimum audible value and a set discomfort threshold. It is equipped with an amplifying means.

According to this audio processing device, the input audio signal is amplified to an appropriate output level set between the minimum audible value and the discomfort threshold value.

Further, according to the speech processing method of the present invention, the input speech signal is divided into at least a first band containing the pitch frequency component and the first formant component and a second band containing the second formant component. , The minimum audible value and the discomfort threshold are set for each band, and the input sound having a level within a predetermined range is amplified to an output level set between the set minimum audible value and the discomfort threshold. Is.

According to this voice processing method, the voice signal is appropriately amplified for each formant component, and is satisfactorily amplified without impairing the clarity of the voice.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below with reference to FIG.
~ It demonstrates with reference to FIG.

In this example, the speech signal processing circuit is for processing speech such as human speech, and is constructed as shown in FIG. In FIG. 1, reference numeral 11 denotes an audio signal input terminal. An audio signal obtained at this input terminal 11 is supplied to a pitch frequency detection circuit 12, a low-pass filter 21, a mid-pass filter 22, and a high-pass filter 23. Supply.

Here, the pitch frequency detection circuit 12 detects the frequency of the peak portion of the pitch component contained in the audio signal. Here, by detecting the frequency of the pitch component, three types are discriminated, that is, the input voice signal is a male voice, a female voice, and a child voice. This pitch frequency detection processing is performed by a method such as a linear prediction analysis method, a PARCOR method, a cepstrum method, an analytic signal method, or the like.

Then, based on the three types of discrimination results based on the detected pitch frequency, each filter 21, 22, 23
It is configured to control the frequency to be passed by. That is,
For example, when the audio signal obtained at the input terminal 11 is digital data, each filter 21, 22 and 23 is configured by a digital filter such as a transversal filter, and the coefficient of this filter is detected by the pitch frequency detection circuit 12. Control based on
At 23, the frequency to be extracted is selected.

Specifically, the low-pass filter 21 passes the pitch component and the first formant component of the audio signal,
The mid-pass filter 22 passes the second formant component of the audio signal, and the high-pass filter 23 outputs the third formant component of the audio signal.
Pass the formant component and consonant band.

The pitch frequency component of the voice signal and each formant component will be described. In the case of a vowel, the pitch frequency component, the first formant component, the second formant component, ... Can be analyzed. That is, for example, when the power spectrum of the voice pronounced as "A" is shown in FIG. 3, the pitch frequency component P, the first formant component F ₁ , the second formant component F ₂ , and the third formant component F _{3 are} arranged from the lowest frequency. There are places where energy is concentrated.

Therefore, in the case of this example, the boundary between the band extracted by the low-pass filter 21 and the band extracted by the mid-pass filter 22 is defined as the first formant component F ₁ and the second formant component F ₂ . To a frequency between them (hereinafter referred to as the first frequency f ₁ ). In addition, the mid-pass filter 2
The boundary between the band extracted by 2 and the band extracted by the high-pass filter 23 is defined by the frequency between the second formant component F ₂ and the third formant component F ₃ (hereinafter referred to as the second frequency f
(Referred to as ₂ ).

Specifically, when the pitch frequency detecting circuit 12 detects a pitch frequency corresponding to a male voice, the first frequency f ₁ is set to around 800 Hz and the second frequency f _{2 is set.} Is set to around 2.4 kHz.
When the pitch frequency detection circuit 12 detects a pitch frequency corresponding to a female voice, the first frequency f ₁ is set to around 1.2 kHz and the second frequency f ₁ is set.
Set ₂ to around 3.2 kHz. When the pitch frequency detection circuit 12 detects a pitch frequency corresponding to a child's voice, the first frequency f ₁ is set to 1.4k.
And the second frequency f ₂ is set to around 3.6 kHz.

Then, the audio signals extracted by the filters 21, 22, and 23 are supplied to different amplitude control units 31 and 3, respectively.
Supply to 2, 33. Each amplitude control unit 31, 32, 33
Have basically the same configuration, and perform amplitude control processing in their respective amplitude control units. Specific amplitude control processing will be described later.

The audio signal processed by the amplitude control unit 31 is supplied to the low pass filter 41, and the amplitude control unit 32 is supplied.
The audio signal processed in (4) is supplied to the mid-pass filter 42, and the audio signal processed in the amplitude control section 33 is supplied to the high-pass filter 43. Here, each filter 41, 4
Reference numerals 2 and 43 are filters whose pass frequencies are controlled based on the detection result of the pitch frequency detection circuit 12. The low pass filter 41 is controlled to have the same pass frequency as the low pass filter 21, and the mid pass filter 42. Is controlled to the same pass frequency as the mid-pass filter 22, and the high-pass filter 4
3 is controlled to the same pass frequency as the high pass filter 23. The filters 41, 42, 43 are used in the amplitude control unit 3
This is a filter for removing noise generated by the processing of 1, 32, and 33.

Then, the outputs of the filters 41, 42, 43 are supplied to the adder 13 for addition processing, and the added audio signal is supplied to the audio signal output terminal 14. Then, the audio signal output from the audio signal output terminal 14 is supplied to an audio processing circuit in the subsequent stage (a reproduction system circuit for reproduction from a speaker or the like, a transmission system circuit for transmission through a predetermined transmission line, etc.).

Next, the structure of the amplitude control units 31, 32 and 33 will be described. FIG. 2 is a diagram showing the structure of the amplitude control unit 31 (the other amplitude control units 32 and 33 have the same circuit structure). ), The signal supplied to the input terminal 51 from the preceding filter is supplied to the amplifier 52, the amplifier 52 performs amplification processing with a predetermined amplification factor, and the amplified signal is supplied to the output terminal 53. Here, the signal obtained at the input terminal 51 is also supplied to the full-wave rectification circuit 54, and full-wave rectification of the signal supplied by the full-wave rectification circuit 54 is performed. And
The signal rectified by the full-wave rectification circuit 54 is supplied to the audio level detection circuit 55. This voice level detection circuit 55
Then, the rectified signal is detected to detect the voice level.

The audio level data detected by the audio level detecting circuit 55 is supplied to the amplification factor control circuit 56 and the noise level detecting circuit 57. Amplification factor control circuit 5
A circuit 6 controls the amplification factor of the amplifier 52, and controls the amplification factor based on the voice level detected by the voice level detection circuit 55 and the noise level detected by the noise level detection circuit 57. In this case, the amplification factor control circuit 56 is preset with the minimum audible level and the discomfort level, and the output level is controlled so as to be set between the minimum audible level and the discomfort level.

The noise level detection circuit 57 detects the noise level from the audio level data output by the audio level detection circuit 55. Here, the audio level detection circuit 55 and the noise level detection circuit 57 are circuits that detect the levels of the signals respectively supplied.
The characteristics to detect are changed. That is, in the level detection processing in the audio level detection circuit 55, the attack time is set to be short and the recovery time is set to be slightly longer than the attack time so that the level that quickly follows the rising edge of the waveform of the input signal is detected. At the same time, the detected level is lowered slightly after the trailing edge of the waveform of the input signal.

Further, in the level detection processing in the noise level detection circuit 57, the attack time is set to be very long and the recovery time is set to be much shorter than the attack time, so that the waveform of the input signal is significantly delayed. The level to be followed is increased, and the detected level is decreased slightly after the trailing edge of the waveform of the input signal.

Therefore, the voice level detection circuit 55 detects a level that substantially follows the input voice waveform, and the noise level detection circuit 57 detects the level of noise constantly included in the input voice waveform. As an example of the attack time and the recovery time, the attack time in the audio level detection circuit 55 is 0.1 msec and the recovery time is 20 msec.
m seconds. The attack time in the noise level detection circuit 57 is 2 seconds and the recovery time is 20 ms.
Note that the attack time here is the time until the detection value rises to the level of 1 / e of the convergence value when the level of the input signal rises by "1", and the recovery time is the level of the input signal. It is the time until the detected value falls to the level of 1 / e of the converged value when it falls by "1".

Here, an example of the level detection processing state in the amplitude control section 31 is shown in FIG. 7. For example, when the waveform of the input audio signal is the waveform shown in FIG. 7A, this waveform is a full wave. The waveform is rectified by the rectifier circuit 54 as shown in FIG. 7B. Then, the level of this rectified signal is detected by the audio level detection circuit 55, and the detection signal shown in C of FIG. 7 is obtained. This detected signal almost follows the level change of the input audio signal. On the other hand, the detection signal output by the voice level detection circuit 55 is the noise level detection circuit 5
7 and the noise level indicated by D in FIG. 7 is detected. The noise level detection signal is a signal for detecting a level that is constantly included in the input voice waveform.

Then, the amplification factor control circuit 56 controls the amplification factor in the amplifier 52 based on the detected voice level and noise level, and the preset minimum audible level and uncomfortable level. 4, the amplification factor is determined by the input / output sound pressure characteristics shown in FIG. First, FIG. 4 shows a control state of the amplification factor when the noise level is not taken into consideration (that is, when the amplification factor is controlled only based on the level detected by the voice level detection circuit 55).

The characteristic a shown in FIG. 4 shows the control state of the amplification factor.
The output level of the signal amplified by the amplifier 52
The preset minimum audible level y₁And unpleasant level
Ry _FourBetween this level y₁And level y_FourAmong the young
Output audio level range y in a narrow range₀To set. Follow
The minimum audible level y₁Amplified to a greater level
Power level x₁However, this output audio level range y
₀Lower limit y of₂Is amplified to a discomfort level y_FourLess than
Input audio level x that should be limited to level₂But this out
Power voice level range y₀Upper limit of y_ThreeInput sound that is amplified to
Voice level range x ₀Is the output audio level range y₀Corresponding to
Like And this range x₀Too small than
The force sound is the minimum audible level y₁So that
Too much input sound is output voice level y₀Upper limit of y_ThreeControl
Control to limit is performed.

The minimum audible level y ₁ and the discomfort level y
Reference numeral ₄ includes a minimum audible level that can be heard by a person who uses this device, a discomfort level at which a person feels an uncomfortable level higher than that level, a band including a pitch component and a first formant component, and a second formant component. There is one in which measurement is performed for each of three bands including a band and a band including a third formant component and a consonant component, and the measured value is set in the amplitude control unit of the corresponding band. Therefore, the minimum audible level and the discomfort level set for each of the amplitude control units 31, 32 and 33 are slightly different.

By thus controlling the amplification factor, the audio signal obtained at the input terminal 11 has a band containing the pitch component and the first formant component, a band containing the second formant component, and a third formant component. After being amplified between the minimum audible level and the discomfort level set for each of the three bands including the component and the consonant component,
The combined signal of the signals in each band is output from the output terminal 14,
An audio signal that has been appropriately amplified in consideration of the minimum audible level and the unpleasant level is obtained for each band. Therefore, for example, by applying the voice processing of this example to the hearing aid, the voice is output as a voice that is easily heard by a hearing-impaired person. Further, the same effect can be obtained when applied to a voice processing device such as a loudspeaker or a telephone which requires similar processing.

FIG. 8, FIG. 9 and FIG. 10 are diagrams showing an example of a voice signal waveform on which the amplitude control processing of this example is performed. In FIG. 8A, a voice pronounced as "mouse" is input. 9A is an enlarged view of the vowel part of the part pronounced "ne", and FIG. 9A shows the rising part of the waveform of the part pronounced "ne". It is a figure which expands and shows. Note that the waveform shown in A of FIG.
9 is the same as the input waveform shown in FIG.
It is the same as the input waveform shown in A of FIG.

8B, 9B and 10B, the amplification processing of this example is applied to the respective input waveforms (waveforms shown in A of FIG. 8, A of FIG. 9 and A of FIG. 10). The output waveform in the case of applying is shown. Here, the output waveform of B of FIG. 8 and B of FIG. 9 which is a partially enlarged view thereof are shown in FIG.
As can be seen by comparing with B of FIG. 1 and B of FIG. 12, the input voice component is appropriately amplified for each band, the formant structure is maintained, and a waveform in which unnecessary components are suppressed is obtained. It can be seen that the speech waveform has a significantly improved clarity.

Incidentally, FIG. 10C is a diagram showing the output waveform by the conventional amplification processing which is processed without band division, for reference. The output waveform of this conventional processing is C in FIG.
As can be seen by comparing the output waveform of FIG. 10 with the output waveform of the processing of this example, in the case of this example, a component that does not need to be recognized as voice is hardly amplified and needs to be recognized as voice. It can be seen that only the component with is well amplified without causing distortion in the waveform.

Next, considering the result detected by the noise level detection circuit 57, the control state when controlling the amplification factor is
As shown in FIG.

The characteristic a shown in FIG. 5 is the characteristic a of the amplification factor described in FIG. 4 (that is, the characteristic when the noise level is not considered).
The noise level detected by the noise level detection circuit 57 changes in the range x _n . here,
When the noise level is a level x ₁₁ which is sufficiently smaller than the minimum audible level, the amplification factor is controlled with the same characteristic a as when the noise level is not considered.

And sufficiently lower than the minimum audible level
Level x₁₁To a relatively large level x₁₂Noise level
Bell change x_aThere was. At this time, amplification processing
Input audio level range x₀Lower limit of x₁₃Noise level
Bell x₁₂The upper limit of x₁₃Level of
Input voice of the output voice level range y₀Lower limit y of₂To
The characteristic b to be amplified is changed. Input voice level
Range x₀Upper limit x ₁₄The amplification characteristic on the side is not changed.

In this way, the lower limit of the input voice level range is changed to a larger one in accordance with the increase of the noise level, whereby an appropriate amplification factor is set according to the noise level included in the input voice signal, Only speech that can be heard at a level higher than the noise level is amplified above the minimum audible level. Therefore, only the voice louder than the noise is amplified, and the voice can be heard without being disturbed by the noise.

Although the processing in the case where the input audio signal contains noise is shown here, such processing is suitable for an audio processing apparatus such as a hearing aid. On the other hand, when noise is not included in the input voice signal as in the case of processing the voice output by the telephone and there is noise at the place where the telephone is installed, the noise level detection circuit 57 outputs the noise from the input voice signal. Instead of detecting the noise level, the noise level at the installation location of the device may be detected, and the output sound level range may be changed according to the detected noise level.

FIG. 6 shows the amplification control state in this case. The characteristic a shown in FIG. 6 is the same as the characteristic a of the amplification factor described in FIG. 4 (that is, the characteristic when the noise level is not considered). Is. Here, when the noise level is the level y ₁₂ that is the same as or lower than the lower limit of the original output level range, the amplification factor is controlled with the same characteristic a as that when the noise level is not considered.

[0044] Then, occurring changes y _a noise level when the noise level y ₁₃ detected is higher than the lower limit of the original output level range, the lower limit value y ₁₄ of the output sound level y ₀ ' The characteristic c is changed to a level slightly higher than the noise level y ₁₃ . In this case, the input voice level range x ₀ is not changed.

In this way, by controlling the amplification factor according to the noise level around the place where the voice is output,
The voice is output at a level higher than the ambient noise level, and the voice can be heard without being disturbed by the noise.

When there is noise in the input voice and there is noise different from the noise included in the input voice around the place where the voice is output, it is based on the noise level detected from the input voice signal. The control of the input voice level range shown in FIG. 5 and the control of the output voice level range shown in FIG. 6 based on the ambient noise level level may be simultaneously performed.

As shown in the characteristic of FIG. 4, the amplification factor may be set without taking noise into consideration.

Further, the noise level detection process may be performed by another process based on the setting of the attack time and the recovery time shown in the above-mentioned embodiment. For example, the noise level may be detected from the long-term power of the input signal.

Further, in the above-mentioned embodiment, it is divided into three bands, that is, the band containing the pitch component and the first formant component, the band containing the second formant component, and the band containing the third formant component and the consonant. Although the amplification process is performed, the process may be performed in another band division. For example, it may be divided into two bands, that is, a band including the pitch component and the first formant component, and a band including the second formant component and a higher-order formant component, for processing. By doing so, the circuit configuration can be simplified. Alternatively, the amplification process may be performed without dividing the band at all. Alternatively, the processing may be divided into more than three bands for processing.

Further, the specific frequency for dividing the band is not limited to that in the above embodiment. For example, the first frequency f, which is the frequency at the boundary of the band described above.
₁ is a frequency appropriately set between 800 Hz and 1.4 kHz, and the second frequency f ₂ is 2.4 kHz.
The frequency may be set appropriately between z and 3.6 kHz. In this case, although the frequency of each band is changed based on the detection of the pitch frequency in the above-described embodiment, the frequency range of each band may be a fixed frequency.

Further, in the case of the above-mentioned embodiment, the filter 2 for dividing the audio signal of each band into bands after the amplification processing is performed.
Filters 41, 42, 43 having the same characteristics as 1, 22, 23
Since the signals in each band are combined after passing through, distortion components such as harmonics generated by the processing in each band are removed, and a good audio signal without distortion is output. Although effective, the filters 41, 42, 43 connected to the subsequent stages of the amplitude control units 31, 32, 33 may be omitted in order to simplify the circuit configuration.

[0052]

According to the voice processing apparatus of the present invention, the input voice signal is amplified to an appropriate output level set between the minimum audible value and the discomfort threshold value, and the voice clarity is excellently maintained without being deteriorated. An amplified audio signal is obtained.

In this case, the input audio signal is divided into a first frequency band and a second frequency band with a first frequency set between 800 Hz and 1.4 kHz as a boundary, and each band is divided into respective bands. By setting the minimum audible value and the uncomfortable threshold value and performing the amplification process, the pitch component and the formant component of the voice signal are properly amplified, and a more favorably amplified voice signal is obtained.

This second frequency band is further divided into 2.
The second frequency, which is set between 4 kHz and 3.6 kHz, is divided into the third frequency band and the fourth frequency band, and the minimum audible value and the discomfort threshold are set for each band. By performing the amplification process by performing the amplification process, each formant component is more appropriately amplified, and a better amplified audio signal is obtained.

When a pitch frequency detecting means is provided and a voice signal corresponding to a male voice is detected, the first frequency is set to around 800 Hz and the second frequency is set to around 2.4 kHz so that a female voice is set. When the voice signal corresponding to is detected, the first frequency is set to around 1.2 kHz, the second frequency is set to around 3.2 kHz, and when the voice signal corresponding to the voice of the child is detected, the first frequency is set to Around 1.4 kHz,
By setting the second frequency to around 3.6 kHz, an appropriate amplification process according to the type of the input audio signal can be performed.

Further, by supplying the amplified signals of the respective bands to the filters for extracting the respective frequency bands, the distortion caused by the amplification processing in the respective bands is removed and a good audio signal is obtained. Will be obtained.

Furthermore, a noise level detecting means for detecting the noise level included in the input voice signal is provided, and the lower limit value of the input voice level to be amplified is changed in the larger direction according to the increase of the detected noise level. By doing so, even when the noise level included in the voice signal is high, the clarity of the voice is secured and the voice becomes easy to hear.

Furthermore, a noise level detecting means for detecting an external noise level is provided, and the lower limit value of the amplified output sound level is changed in a larger direction according to the increase of the detected noise level. By doing so, the clarity of the voice is secured even when the surrounding noise level is high,
Voice becomes easier to hear.

According to the voice processing method of the present invention, the voice signal is properly amplified for each formant component, and is satisfactorily amplified without impairing the clarity of the voice.

Further, in this case, the noise level included in the input voice signal is detected, and the range of the level of the input voice to be amplified is changed according to the detected noise level. Even if the included noise level is high, the intelligibility of the voice is secured and the voice becomes easy to hear.

Further, by detecting the external noise level and changing the range of the output level of the amplified voice according to the detected noise level, the voice level is increased even if the ambient noise level is high. The intelligibility is secured, and the voice becomes easy to hear.

[Brief description of drawings]

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

FIG. 2 is a configuration diagram illustrating an amplitude control unit according to an embodiment.

FIG. 3 is a characteristic diagram showing a formant component of an audio signal.

FIG. 4 is a characteristic diagram showing an amplitude control state according to an embodiment.

FIG. 5 is a characteristic diagram showing an amplitude control state according to an embodiment when a noise level included in an input audio signal is considered.

FIG. 6 is a characteristic diagram showing an amplitude control state according to an embodiment when a noise level around a reproduced place is considered.

FIG. 7 is a waveform diagram shown for explaining amplitude control processing according to an embodiment.

FIG. 8 is a waveform diagram showing an amplitude control state according to an embodiment.

9 is a waveform diagram showing an enlarged waveform of the amplitude control state shown in FIG.

10 is a waveform diagram showing the waveform of the amplitude control state shown in FIG. 9 in a further enlarged manner.

FIG. 11 is a waveform diagram showing an example of a conventional amplitude control state.

FIG. 12 is a waveform diagram showing an enlarged waveform of the amplitude control state of FIG. 11.

[Explanation of symbols]

 11 voice signal input terminal 12 pitch frequency detection circuit 13 adder 14 voice signal output terminal 21,41 low pass filter 22,42 mid pass filter 23,43 high pass filter 31,32,33 amplitude control section 52 amplifier 54 Full wave rectification circuit 55 Audio level detection circuit 56 Amplification factor control circuit 57 Noise level detection circuit

Claims (10)

[Claims] 1. A voice processing device for converting a voice amplitude of an input voice signal, wherein a minimum audible value setting means, a discomfort threshold setting means, and an input voice having a level within a predetermined range are set by the minimum audible value setting means. The audio processing apparatus comprising: an amplification unit that amplifies to an output level set between the minimum audible value set and the discomfort threshold set by the discomfort threshold setting unit. 2. The input voice signal is converted from 800 Hz to 1.4.
With the first frequency set up to kHz as the boundary, the first
The frequency band and the second frequency band are divided, and the amplification means is prepared for each band, and the amplification process is performed by setting the minimum audible value and the discomfort threshold value for each amplification means. The voice processing device according to claim 1. 3. The second frequency band is further 2.4 k
It is divided into a third frequency band and a fourth frequency band with a second frequency set between Hz and 3.6 kHz as a boundary, and the amplification means is prepared for each band, The audio processing apparatus according to claim 2, wherein the amplification process is performed by setting the minimum audible value and the discomfort threshold value for each amplification unit. 4. A pitch frequency detecting means for an input voice signal is provided, and when the pitch frequency detecting means detects a voice signal corresponding to a male voice, the first frequency is set to about 800 Hz,
When the second frequency is set to about 2.4 kHz and the pitch frequency detecting means detects a voice signal corresponding to a female voice, the first frequency is set to about 1.2 kHz and the second frequency is set to When the voice signal corresponding to a child's voice is detected by the pitch frequency detecting means by setting the frequency to about 3.2 kHz, the first frequency is set to about 1.4 kHz and the second frequency is set to about 3.6 kHz. The audio processing device according to claim 3, wherein 5. The audio processing device according to claim 3, wherein the signals in the respective bands amplified by different amplifying means are supplied to filters for extracting respective frequency bands. 6. An input voice comprising a noise level detecting means for detecting a noise level included in the input voice signal, wherein the amplifying means performs amplification processing in response to an increase in the noise level detected by the noise level detecting means. The lower limit value of the level is changed in a larger direction.
The voice processing device described. 7. A noise level detecting means for detecting an external noise level is provided, and a lower limit value of the output sound level amplified by the amplifying means is set in accordance with an increase in the noise level detected by the noise level detecting means. , The change is made in a larger direction.
The voice processing device described. 8. A voice processing method for converting a voice amplitude of an input voice signal, wherein the input voice signal includes at least a first band including a pitch frequency component and a first formant component, and a second formant component. The second audible band is divided into the second audible band and the minimum audible value and the uncomfortable threshold value are set for each band. A voice processing method designed to amplify to the output level. 9. The voice processing according to claim 8, wherein the noise level included in the input voice signal is detected, and the range of the level of the input voice to be amplified is changed according to the detected noise level. Method. 10. The voice processing method according to claim 8, wherein an external noise level is detected, and the range of the output level of the amplified voice is changed according to the detected noise level.