JP2011075694A - Sound processing device and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a sound processing device capable of improving articulation of a sound signal by improving a signal to noise ratio with natural audibility maintained. <P>SOLUTION: The invention is related to a device for improving articulation of the sound signal by improving the signal to noise ratio by suppressing a noise component and by emphasizing a sound component. A noise suppression lower limit value by frequency band determined on the basis of human audibility characteristics, and a sound emphasis upper limit by frequency band determined on the basis of the human audibility characteristics, are held, and a sound component band or a noise component band is discriminated by frequency band. A gain of the sound signal to be processed is controlled so as not to become smaller than the noise suppression lower limit value of the frequency band, for the frequency band determined to be the noise component, and the gain of the sound signal to be processed is controlled so as not to become larger than the sound emphasis upper limit value of the frequency band, for the frequency band determined to be the sound component. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a sound processing apparatus and program, and can be applied to, for example, improving the clarity of captured speech or pronunciation output speech in a telephone terminal.

  The voice with noise superimposed has problems such as intelligibility and difficulty in listening to the content of the utterance. In order to solve this, conventionally, there are techniques called a noise canceller and a noise suppressor. In addition, there is a speech enhancement technique as disclosed in Patent Document 1. The former improves the speech-to-noise ratio (S / N ratio) by eliminating and suppressing the noise part from the speech with superimposed noise, thereby improving the clarity. The latter enhances the speech-to-noise ratio and enhances the intelligibility by emphasizing the speech part.

JP-A-6-208395

  However, in the former method, no particular consideration is given to the frequency characteristics of the noise after being erased and suppressed, and there is a case where noise is generated although the level is lowered, and the auditory quality is deteriorated. In the latter method, the frequency characteristic of the original voice may be excessively deformed to increase the sense of distortion.

  The present invention has been made in view of the above-described problems, and a sound processing apparatus and program that improve a signal-to-noise ratio while maintaining a natural audibility and improve the clarity of a sound signal (audio signal or sound signal). It is what I tried to provide.

  According to a first aspect of the present invention, there is provided a sound processing apparatus for improving a signal-to-noise ratio by suppressing a noise component to improve a clarity of a sound signal. (1) For each frequency band determined based on human auditory characteristics Suppression lower limit holding means for holding a noise suppression lower limit value, (2) signal / noise determination means for determining whether or not each frequency band is a noise component band, and (3) a frequency determined as a noise component. A gain control means for controlling the gain of the sound signal to be processed is provided for the band so as not to fall below the noise suppression lower limit value of the frequency band.

  According to a second aspect of the present invention, there is provided a sound processing apparatus for improving a signal-to-noise ratio by enhancing a sound component to improve a clarity of a sound signal. (1) For each frequency band determined based on human auditory characteristics Emphasis upper limit holding means for holding a sound emphasis upper limit; (2) signal / noise determination means for determining whether or not each frequency band is a sound component band; and (3) a frequency determined as a sound component. A gain control means for controlling the gain of the sound signal to be processed is provided for the band so as not to exceed the sound enhancement upper limit value of the frequency band.

  A third aspect of the present invention is a sound processing apparatus that improves the signal-to-noise ratio by suppressing noise components and enhancing sound components to improve the intelligibility of sound signals, and is determined based on (1) human auditory characteristics. Suppression lower limit holding means for holding a noise suppression lower limit value for each frequency band, and (2) enhancement upper limit holding means for holding a sound enhancement upper limit value for each frequency band determined based on human auditory characteristics And (3) a signal / noise determining means for determining a sound component band or a noise component band for each frequency band, and (4) for a frequency band determined as a noise component, noise in the frequency band. Control the gain of the sound signal to be processed so that it does not fall below the suppression lower limit, and for the frequency band determined to be a sound component, do not exceed the sound enhancement upper limit for that frequency band. To control the gain of the sound signal Characterized by a control unit.

  A sound processing program according to a fourth aspect of the present invention includes: (1) a suppression lower limit holding unit that holds a noise suppression lower limit for each frequency band determined based on human auditory characteristics; and (2) a frequency A signal / noise determining means for determining whether or not the band is a noise component for each band; and (3) a frequency band determined to be a noise component so as not to fall below a noise suppression lower limit value of the frequency band. And (4) improving the signal-to-noise ratio by suppressing the noise component to improve the clarity of the sound signal.

  A sound processing program according to a fifth aspect of the present invention includes: (1) an emphasis upper limit holding unit that holds a sound emphasis upper limit for each frequency band determined based on human auditory characteristics; and (2) a frequency A signal / noise determining means for determining whether or not the sound component is in each band; and (3) a frequency band determined to be a sound component so as not to exceed a sound enhancement upper limit value of the frequency band. And (4) improving the signal-to-noise ratio by enhancing the sound component to improve the intelligibility of the sound signal.

  A sound processing program according to a sixth aspect of the present invention includes: (1) a suppression lower limit holding unit that holds a noise suppression lower limit value for each frequency band determined based on human auditory characteristics; An emphasis upper limit holding means for holding a sound emphasis upper limit value for each frequency band determined based on the auditory characteristics of the sound signal, and (3) a signal / discriminating a sound component band or a noise component band for each frequency band (4) For the frequency band determined as the noise component, (4) controlling the gain of the sound signal to be processed so as not to fall below the noise suppression lower limit value of the frequency band, The determined frequency band is made to function as gain control means for controlling the gain of the sound signal to be processed so as not to exceed the sound enhancement upper limit value of the frequency band, and (5) noise component suppression and sound Signal by component enhancement Improve noise ratio, characterized in that to improve the clarity of the sound signal.

  According to the present invention, the signal-to-noise ratio can be improved while maintaining a natural audibility, and the clarity of the sound signal can be improved.

It is a block diagram which shows the functional structure of the audio processing apparatus 100 of 1st Embodiment. FIG. 2 is a block diagram illustrating a detailed configuration of a gain change circuit in FIG. 1. It is explanatory drawing which shows the frequency characteristic of the input signal and output signal to the gain change circuit group of FIG. It is a block diagram which shows the functional structure of the audio processing apparatus 100 of 2nd Embodiment. FIG. 5 is a block diagram illustrating a detailed configuration of the gain change circuit of FIG. 4. FIG. 5 is an explanatory diagram of changing a set of a speech enhancement upper limit value and a noise suppression lower limit value in the gain change circuit of FIG. 4.

(A) First Embodiment Hereinafter, a first embodiment in which a sound processing apparatus and a program according to the present invention are applied to sound processing will be described in detail with reference to the drawings. The sound processing device (sound processing device) according to the first embodiment is, for example, a digital signal that is captured by a microphone of a telephone terminal and interposed at a position where a near-end sound signal converted into a digital signal is processed, or applied to a speaker. It is interposed at a position where a far-end audio signal consisting of a signal is processed.

(A-1) Configuration of the First Embodiment FIG. 1 is a block diagram showing a functional configuration of the speech processing apparatus 100 of the first embodiment. For example, a softphone is constructed by installing a softphone application on a computer, and the sound processing program (voice processing program) of the first embodiment is applied as a partial program of the softphone application. However, the speech processing apparatus 100 according to the first embodiment may be constructed. Even in this case, the functional configuration can be represented in FIG.

  In FIG. 1, the speech processing apparatus 100 according to the first embodiment includes a sound / silence detection circuit 101, a frequency band division circuit 102, gain change circuits 103-1 to 103 -M, and a frequency band synthesis circuit 104.

  An audio signal s (n) (where n is a parameter representing time), which is a digital signal on which noise is superimposed, is input to the sound / silence detection circuit 101 and the frequency band division circuit 102.

  The sound / silence detection circuit 101 determines whether the input sound signal s (n) is sound (sound section) or sound (silence section). As the determination method by the sound / silence detection circuit 101, any existing method can be applied. For example, if the signal power threshold value for determination is present and the power of the input audio signal s (n) exceeds this value, the result is sound (V = 1), and if not, the result is silent (V = 0). Output to circuits 103-1 to 103-M.

  The frequency band dividing circuit 102 divides the input audio signal s (n) into M frequency bands, and each of the band divided signals s1 (n) to sM (n) is respectively corresponding to the gain changing circuits 103-1 to 103. Output to -M.

  Each gain changing circuit 103-m (m is 1 to M) controls the gain of the signal component or noise component with respect to the input band-divided signal sm (n), thereby improving the speech intelligibility. In the case of the first embodiment, the gain change circuit 103-m has a gain control limit determined in accordance with human auditory characteristics (loudness curve) such as ISO (International Organization for Standardization) 226.

  The frequency band synthesis circuit 104 synthesizes the output signals x1 (n) to xM (n) from the gain change circuits 103-1 to 103-M, and the synthesized audio signal x (n) is clarified audio signal. Is output as

  FIG. 2 is a block diagram showing a detailed configuration of the gain changing circuit 103-m. The gain changing circuits 103-1 to 103 -M have the same configuration, but the gain changing circuits 103-1 to 103 -M are voice enhancement upper limit values and noise suppression lower limit values stored in a limit value storage unit 203 described later. Is different.

  In FIG. 2, the gain change circuit 103-m includes an SN calculation unit 200, an SN threshold supply unit 201, a voice / noise determination unit 202, a limit value storage unit 203, and a gain change unit 204.

  The SN calculation unit 200 calculates a voice-to-noise ratio (SN ratio) snm of the input band-divided signal sm (n). As a calculation method of the S / N ratio snm, any existing method may be applied, and an example is as follows.

  The power of the band-divided signal sm (n) is calculated and set as the signal power smp. When there is no sound (V = 0), for example, the estimated noise value smnp is updated according to the equation (1) (where 0 <α <1). On the other hand, when there is a sound (V = 1), the noise estimate value smnp is not updated. And SN ratio snm is calculated according to (2) Formula.

smnp = αXsmnp + (1−α) Xsmp (1)
snm = ₁₀ × log ₁₀ (smp / smnp) (2)
The SN threshold supply unit 201 supplies an SN threshold to be compared with the SN ratio snm to the voice / noise determination unit 202. The SN threshold supply unit 201 calculates the average power of the band division signal sm (n) when there is no sound (V = 0), and adds the predetermined offset to the average power value as the SN threshold. You may make it supply to the noise determination part 202. FIG. In addition, a boundary value of an SN ratio that can process a voice signal of many people, and can determine whether the band that the gain changing circuit 103-m is in charge of is a voice band or a noise band, and supplies an SN threshold value. The SN threshold value may be set in advance in the unit 201 as the SN threshold value, and the SN threshold value supply unit 201 may be supplied to the voice / noise determination unit 202 as a preset SN threshold value.

  The voice / noise determination unit 202 compares the SN ratio snm calculated by the SN calculation unit 200 with the SN threshold value supplied from the SN threshold value supply unit 201, so that the band assigned to the gain change circuit 103-m is Speaking from the characteristics of the target speaker (speech), it is determined whether the voice band or noise band is reached. When the SN ratio snm exceeds the SN threshold, the voice / noise determination unit 202 determines that the band handled by the gain change circuit 103-m is a voice band, and otherwise, the gain change circuit. It is determined that the band assigned to 103-m is a noise band.

  In the limit value storage unit 203, the upper limit limit gain (speech enhancement upper limit value) of speech enhancement when the band handled by the gain change circuit 103-m is a speech band, and the gain change circuit 103-m is responsible. The limit gain (noise suppression lower limit value) of the lower limit of noise suppression when the band to be processed is a noise band is stored. Note that the limit value storage unit 203 may store only one of the voice enhancement upper limit value and the noise suppression lower limit value, and the other may be obtained by subtracting or adding a predetermined amount from the stored value. .

  A voice enhancement upper limit curve C11 in FIG. 3 to be described later is a curve obtained by connecting voice enhancement upper limit values in the gain change circuits 103-1 to 103-M, and a noise suppression lower limit curve C12 in FIG. 3 is a gain change circuit 103-. It is the curve which connected the noise suppression lower limit in 1-103-M. These curves C1 and C2 have a curved shape corresponding to human auditory characteristics (loudness curve) such as ISO (International Organization for Standardization) 226.

  The gain changing unit 204 switches processing according to a determination result by the voice / noise determining unit 202 that the band handled by the gain changing circuit 103-m is a voice band or a noise band.

  The gain changing unit 204, when the band handled by the gain changing circuit 103-m is a voice band, the gain of the input band division signal sm (n) is stored in the limit value storage unit 203. It is determined whether or not it is below the emphasis upper limit value. Here, the gain of the band division signal sm (n) is the power of the band division signal sm (n). When the gain of the band division signal sm (n) is below the voice enhancement upper limit value, the gain changing unit 204 changes the input band division signal sm (n) so that the gain matches the voice enhancement upper limit value. When the gain of the band division signal sm (n) is equal to or higher than the voice enhancement upper limit value, the input band division signal sm (n) is passed through without changing the gain.

  When the band assigned to the gain changing circuit 103-m is a noise band, the gain changing unit 204 uses the noise stored in the limit value storage unit 203 as the gain of the input band division signal sm (n). It is determined whether or not the suppression lower limit value is exceeded. When the gain of the band division signal sm (n) exceeds the noise suppression lower limit value, the gain changing unit 204 changes the input band division signal sm (n) so that the gain matches the noise suppression lower limit value. When the gain of the band division signal sm (n) is equal to or less than the noise suppression lower limit value, the input band division signal sm (n) is passed as it is without changing the gain.

(A-2) Operation of the First Embodiment Next, the operation of the speech processing apparatus 100 of the first embodiment will be described with reference to the frequency characteristic diagram shown in FIG.

  An audio signal s (n), which is a digital signal on which noise is superimposed, is input to the sound / silence detection circuit 101 and the frequency band division circuit 102.

  The input audio signal s (n) is divided into M frequency bands by the frequency band dividing circuit 102, and the divided band divided signals s1 (n) to sM (n) are respectively corresponding gain changing circuits 103. -1 to 103-M. The voiced / silent detection circuit 101 determines whether the input voice signal s (n) is voiced or silent, and the determination result is also given to the gain changing circuits 103-1 to 103 -M.

  In the gain changing circuit 103-m (m is 1 to M), the SN calculation unit 200 calculates the SN ratio snm of the input band division signal sm (n), and the SN ratio snm and the SN threshold supply unit 201 Is compared with the SN threshold value supplied by the voice / noise determination unit 202, and it is determined whether the band handled by the gain changing circuit 103-m is a voice band or a noise band.

  When it is determined that the band handled by the gain changing circuit 103-m is the voice band, the gain changing unit 204 stores the gain of the input band division signal sm (n) in the limit value storage unit 203. Whether or not the gain of the band division signal sm (n) is below the voice enhancement upper limit value, the input band division signal sm (n) Is changed (amplified) so as to coincide with the voice enhancement upper limit value, and when the gain of the band division signal sm (n) is equal to or higher than the voice enhancement upper limit value, the input band division signal sm (n) is gain It is output as it is without being changed.

  On the other hand, when the band assigned to the gain changing circuit 103-m is a noise band, the gain changing unit 204 stores the gain of the input band division signal sm (n) in the limit value storage unit 203. If the gain of the band division signal sm (n) exceeds the noise suppression lower limit value, it is determined whether or not the gain of the input band division signal sm (n) is higher than the noise suppression lower limit value. If it is output after being changed (attenuated) to match the noise suppression lower limit value and the gain of the band division signal sm (n) is less than or equal to the noise suppression lower limit value, the input band division signal sm (n) is gain changed. It is output as is without being processed.

  The output signals x1 (n) to xM (n) from the gain changing circuits 103-1 to 103-M are synthesized by the frequency band synthesizing circuit 104, and the synthesized audio signal x (n) is clarified audio signal. Is output from the speech processing apparatus 100.

  3A shows the gains of the band division signals s1 (n) to sM (n) input to the gain change circuit groups 103-1 to 103 -M, and FIG. 3B shows the gain change circuit group. The gains of the band division signals x1 (n) to xM (n) output from 103-1 to 103-M are shown. In FIG. 3, a band indicating a gain with a black vertical bar indicates a band determined as an audio band by the internal audio / noise determination unit 202 (202-1 to 202-M), and hatched. The band indicating the gain by the vertical bar indicates the band determined as the noise band by the internal voice / noise determination unit 202 (202-1 to 202-M).

  As shown in FIG. 3A, the first band is determined to be a noise band, and the gain of the band division signal s1 (n) exceeds the noise suppression lower limit value. As shown, the band division signal s1 (n) is changed (attenuated) so that its gain matches the noise suppression lower limit value, and the changed band division signal x1 (n) is output.

  As shown in FIG. 3A, the fifth band is determined to be a noise band, but the gain of the band division signal s5 (n) is equal to or lower than the noise suppression lower limit value, so that the fifth band is shown in FIG. Thus, the band division signal s5 (n) becomes the output band division signal x5 (n) as it is.

  As shown in FIG. 3A, the second band is determined to be a voice band, and the gain of the band division signal s2 (n) is lower than the voice enhancement upper limit value. As shown, the band division signal s2 (n) is changed (amplified) so that its gain matches the voice enhancement upper limit value, and the changed band division signal x2 (n) is output.

  The third band is determined to be a voice band as shown in FIG. 3A, but the gain of the band division signal s3 (n) is equal to or higher than the voice enhancement upper limit value, and therefore, the third band is shown in FIG. Thus, the band division signal s3 (n) becomes the output band division signal x3 (n) as it is.

(A-3) Effects of the First Embodiment According to the first embodiment, the voice enhancement upper limit value and the noise suppression lower limit value based on the human auditory characteristics are provided, and the voice enhancement upper limit and the noise suppression lower limit are set. Thus, since voice enhancement or noise suppression is performed, the voice-to-noise ratio can be improved while maintaining a natural audibility and voice clarity can be improved.

(B) Second Embodiment Next, a second embodiment in which the sound processing apparatus and the program according to the present invention are applied to sound processing will be described in detail with reference to the drawings.

  FIG. 4 is a block diagram illustrating a functional configuration of the speech processing apparatus 100A according to the second embodiment. The same or corresponding parts as those in FIG. 1 according to the first embodiment are denoted by the same or corresponding reference numerals. Show. FIG. 5 is a block diagram showing a detailed configuration of the gain changing circuit 103A-m in the second embodiment. The same components as those in FIG. 4 according to the first embodiment, corresponding components are denoted by the same reference numerals. It is attached.

  The sound processing apparatus 100A of the second embodiment includes a sound / silence detection circuit 101, a frequency band dividing circuit 102, gain changing circuits 103A-1 to 103A-M, and a frequency band synthesizing circuit 104 similar to those of the first embodiment. In addition, a limit value change command circuit 105 is provided. Further, the internal configuration of the gain changing circuit 103A-m (m is 1 to M) is different from that of the first embodiment. Similarly to the first embodiment, the gain changing circuit 103A-m includes an SN calculation unit 200, an SN threshold supply unit 201, an audio / noise determination unit 202, a limit value storage unit 203A, and a gain change unit 204. The value storage unit 203A is different from that of the first embodiment.

  The limit value storage unit 203A in the second embodiment stores a plurality of sets of speech enhancement upper limit values and noise suppression lower limit values. In the following, for simplification of explanation, it is assumed that two sets of speech enhancement upper limit values and noise suppression lower limit values are stored in the limit value storage unit 203A. Of the two sets, the voice enhancement upper limit value and noise suppression lower limit value of one set take into account the hearing characteristics of the elderly, and the voice enhancement upper limit value and noise suppression lower limit value of the other set are auditory other than the elderly. The characteristics are taken into consideration.

  Note that the limit value storage unit 203A in the second embodiment stores only a certain set of speech enhancement upper limit values and noise suppression lower limit values, and stores other sets of speech enhancement upper limit values and noise suppression lower limit values. It may be obtained by applying a predetermined conversion function to a certain value.

  The limit value change command circuit 105 newly provided in the second embodiment is a voice enhancement used by the gain changing unit 204 (204-1 to 204-M) for the gain changing circuits 103A-1 to 103A-M. A set of an upper limit value and a noise suppression lower limit value is commanded. The gain changing units 204-1 to 204-M change a pair used in conjunction with the change command from the limit value change command circuit 105.

  For example, the limit value change command circuit 105 is configured around a specific key on a push switch or a keyboard, and each time the push switch or a specific key is operated, the switch to the other set is instructed. good. Further, for example, when the limit value change command circuit 105 is configured with a dip switch or two keys corresponding to each group as a center, and the dip switch is operated or a key corresponding to a group not selected is operated. Alternatively, switching to the other set may be instructed.

  For example, if the speech processing apparatus 100A of the second embodiment is incorporated in a soft phone and provided in a system that processes a speech signal from a microphone, the near-end speaker is When the speaker is recognized as an elderly person, the limit value change command circuit 105 may change the voice enhancement upper limit value and the noise suppression lower limit value in consideration of the hearing characteristics of the elderly person. Also, for example, if the speech processing apparatus 100A of the second embodiment is incorporated in a soft phone and provided in a system that processes speech signals to a speaker, When the user is an elderly person, the limit value change command circuit 105 may change the voice enhancement upper limit value and noise suppression lower limit value in consideration of the hearing characteristics of the elderly person.

  The limit value change command circuit 105 may be an application of an existing technique that automatically determines whether or not the voice of the elderly person is input from the input voice signal s (n).

  FIG. 6A shows a selected state of a voice enhancement upper limit value and a noise suppression lower limit value in consideration of auditory characteristics other than the elderly, and in such a state, the voice enhancement upper limit in consideration of the auditory characteristics of the elderly person. When a change command to a set of a value and a noise suppression lower limit value is generated, the voice enhancement upper limit value and the noise suppression lower limit value defined by the voice enhancement upper limit curve C21 and the noise suppression lower limit curve C22 indicated by solid lines in FIG. The group is switched to a state used by the gain changing units 204-1 to 204-M.

  Here, the speech enhancement upper limit curve C21 and the noise suppression lower limit curve C22 in consideration of the elderly's auditory characteristics are determined in consideration of differences in auditory characteristics depending on age based on ISO 7029 and the like. That is, in human hearing, since high frequencies are difficult to perceive with age, in consideration of this, the speech enhancement upper limit curve C21 and the noise suppression lower limit curve C22 take into account the speech enhancement upper limit value and noise suppression in a high frequency region. The lower limit value is set higher than the speech enhancement upper limit curve C11 and the noise suppression lower limit curve C12 related to other groups.

  In the example of FIG. 6, when the speech enhancement upper limit curve C21 and the noise suppression lower limit curve C22 considering the auditory characteristics of the elderly are valid compared to when the speech enhancement upper limit curve C11 and the noise suppression lower limit curve C12 are valid, For example, the attenuation amount of the Mth noise band is reduced, and the (M-3) th audio band is changed from an unamplified one to an amplification target.

(B-3) Effects of Second Embodiment Also according to the second embodiment, a voice enhancement upper limit value and a noise suppression lower limit value based on human auditory characteristics are provided, and a voice enhancement upper limit and a noise suppression lower limit are set. Thus, since speech enhancement and noise suppression are performed, the speech-to-noise ratio can be improved while maintaining a natural audibility, and speech intelligibility can be improved.

  Furthermore, according to the second embodiment, considering the human auditory characteristics that change with age, a plurality of sets of speech enhancement upper limit values and noise suppression lower limit values can be provided for selection and use. Compared with the first embodiment, it is possible to improve the intelligibility of the voice in accordance with the listener.

(C) Other Embodiments In each of the above embodiments, the speech processing device that improves the speech-to-noise ratio (SN ratio) by using the suppression of the noise component and the enhancement of the speech component has been described. Thus, the present invention can be applied to a speech processing apparatus that improves the speech-to-noise ratio, and the present invention can be applied to a speech processing apparatus that improves the speech-to-noise ratio by enhancing speech components.

  In the description of the above embodiment, an example in which the voice processing device is applied to a telephone terminal is given, but the application of the present invention is not limited to this.

  Moreover, although what applied this invention to the audio processing apparatus which processes an audio | voice signal was shown in said each embodiment, this invention is applicable to the audio processing apparatus which processes an acoustic signal (audio signal). The “sound signal” in the claims represents an audio signal or an acoustic signal.

  DESCRIPTION OF SYMBOLS 100, 100A ... Voice processing apparatus, 101 ... Sound / silence detection circuit, 102 ... Frequency band division circuit, 103-1 to 103-M, 103A-1 to 103A-M ... Gain change circuit, 104 ... Frequency band synthesis circuit, 105 ... limit value change command circuit, 200 ... SN calculation unit, 201 ... SN threshold supply unit, 202 ... voice / noise determination unit, 203, 203A ... limit value storage unit, 204 ... gain change unit.

Claims (9)

In a sound processing device that improves the intelligibility of a sound signal by improving the signal-to-noise ratio by suppressing noise components,
A suppression lower limit holding means for holding a noise suppression lower limit for each frequency band determined based on human auditory characteristics;
For each frequency band, a signal / noise determination means for determining whether the band is a noise component band,
And a gain control means for controlling the gain of the sound signal to be processed so that the frequency band determined to be a noise component does not fall below a noise suppression lower limit value of the frequency band. apparatus. The suppression lower limit holding means holds, as noise suppression lower limit values for each frequency band, a plurality of values according to the human auditory age change characteristic, and selects a value to be applied from the plurality of values. Item 2. The sound processing apparatus according to Item 1. In a sound processing device that improves the intelligibility of a sound signal by improving the signal-to-noise ratio by enhancing the sound component,
Emphasis upper limit holding means for holding a sound emphasis upper limit for each frequency band determined based on human auditory characteristics;
For each frequency band, a signal / noise determination means for determining whether or not the sound component band,
And a gain control means for controlling the gain of the sound signal to be processed so that the frequency band determined to be a sound component does not exceed the sound enhancement upper limit value of the frequency band. apparatus. The emphasis upper limit holding unit holds a plurality of values according to the age change characteristics of human hearing as a sound emphasis upper limit value of each frequency band, and selects a value to be applied from the plurality of values. Item 4. The sound processing device according to item 3. In a sound processing apparatus for improving the intelligibility of a sound signal by improving the signal-to-noise ratio by suppressing the noise component and enhancing the sound component,
A suppression lower limit holding means for holding a noise suppression lower limit for each frequency band determined based on human auditory characteristics;
Emphasis upper limit holding means for holding a sound emphasis upper limit for each frequency band determined based on human auditory characteristics;
For each frequency band, a signal / noise determining means for determining a sound component band or a noise component band;
For the frequency band determined to be a noise component, the gain of the sound signal to be processed is controlled so that it does not fall below the noise suppression lower limit value of that frequency band, and for the frequency band determined to be a sound component Has a gain control means for controlling the gain of the sound signal to be processed so as not to exceed the sound enhancement upper limit value of the frequency band. The suppression lower limit holding means holds a plurality of values according to the age change characteristics of human hearing as a noise suppression lower limit value of each frequency band, and selects a value to be applied from a plurality of values,
The emphasis upper limit holding unit holds a plurality of values according to the age change characteristics of human hearing as a sound emphasis upper limit value of each frequency band, and selects a value to be applied from the plurality of values. Item 6. The sound processing device according to Item 5. Computer
A suppression lower limit holding means for holding a noise suppression lower limit for each frequency band determined based on human auditory characteristics;
For each frequency band, a signal / noise determination means for determining whether the band is a noise component band,
For a frequency band determined as a noise component, it functions as a gain control means that controls the gain of the sound signal to be processed so that it does not fall below the noise suppression lower limit value of that frequency band, thereby suppressing the noise component. A sound processing program for improving the intelligibility of a sound signal by improving the signal-to-noise ratio. Computer
Emphasis upper limit holding means for holding a sound emphasis upper limit for each frequency band determined based on human auditory characteristics;
For each frequency band, a signal / noise determination means for determining whether or not the sound component band,
For a frequency band determined to be a sound component, it functions as a gain control means for controlling the gain of the sound signal to be processed so that it does not exceed the sound enhancement upper limit value for that frequency band. A sound processing program for improving the intelligibility of a sound signal by improving the signal-to-noise ratio. Computer
A suppression lower limit holding means for holding a noise suppression lower limit for each frequency band determined based on human auditory characteristics;
Emphasis upper limit holding means for holding a sound emphasis upper limit for each frequency band determined based on human auditory characteristics;
For each frequency band, a signal / noise determining means for determining a sound component band or a noise component band;
For the frequency band determined to be a noise component, the gain of the sound signal to be processed is controlled so that it does not fall below the noise suppression lower limit value of that frequency band, and for the frequency band determined to be a sound component Functions as a gain control means that controls the gain of the sound signal to be processed so that it does not exceed the sound enhancement upper limit of that frequency band, and the signal-to-noise ratio is reduced by suppressing the noise component and enhancing the sound component. A sound processing program characterized by improving the clarity of a sound signal.

Images