JP2009092994A - Audio teleconference device - Google Patents

Audio teleconference device Download PDF

Info

Publication number
JP2009092994A
JP2009092994A JP2007264422A JP2007264422A JP2009092994A JP 2009092994 A JP2009092994 A JP 2009092994A JP 2007264422 A JP2007264422 A JP 2007264422A JP 2007264422 A JP2007264422 A JP 2007264422A JP 2009092994 A JP2009092994 A JP 2009092994A
Authority
JP
Japan
Prior art keywords
input data
value
means
signal
analog
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
JP2007264422A
Other languages
Japanese (ja)
Inventor
Shinya Urushizaki
慎也 漆崎
Original Assignee
Audio Technica Corp
株式会社オーディオテクニカ
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Audio Technica Corp, 株式会社オーディオテクニカ filed Critical Audio Technica Corp
Priority to JP2007264422A priority Critical patent/JP2009092994A/en
Publication of JP2009092994A publication Critical patent/JP2009092994A/en
Application status is Pending legal-status Critical

Links

Images

Abstract

<P>PROBLEM TO BE SOLVED: To obtain an audio teleconference device capable of avoiding impulsive sound output from a speaker, by distinguishing voice from the impulsive sound (hitting sound) which is generated by, for example, attendant's handclapping and desk hitting. <P>SOLUTION: The audio teleconference device includes: an analog/digital converter 2 for converting an input signal from a microphone 1 to a digital signal; an input data storage means 4 for temporarily storing the digital signal; a first arithmetic means 5 for calculating an absolute value of the stored input data; a second arithmetic means 6 for calculating an average value of the absolute value; and a determination means 7 for determining that it is an impulse sound, if a probability of the absolute value which exceeds the average value is a predetermined value or more. <P>COPYRIGHT: (C)2009,JPO&amp;INPIT

Description

  The present invention relates to an audio conference apparatus used when a conference is attended by a large number of people, and more particularly, ambient noise, for example, an impulse sound (impact generated when an attendee's applause or a desk is struck). The present invention relates to an audio conference apparatus that distinguishes between sound and audio and avoids output of such an impact sound from a speaker.

  When a conference is attended by a large number of people, the speaker's voice is picked up by a microphone and amplified by an amplifier so that the voice of one speaker can reach all the members. A voice system is used. Many microphones are used in conferences where an audio system is used. When many microphones are turned on at the same time and are in a so-called “live” state, the voice captured by these microphones is amplified and flows from the speaker, so the voice other than the voice of the speaker becomes noise. It ’s hard to hear. Therefore, normally, each microphone output is muted, and when an attendee speaks, a nearby microphone picks up the sound, and if the sound pressure exceeds a predetermined level, the microphone is unmuted, and the microphone There is an audio conference apparatus that outputs a converted audio signal from a speaker. That is, it is an automatic audio conference apparatus that recognizes a speech only when audio is input to a microphone and outputs audio from a speaker.

  An impulse sound is generated when an attendee claps during a meeting, hits the table, or drops an object near the microphone. When the impulse sound is picked up by the microphone and the sound pressure is higher than a predetermined level, it may be mistaken for the speech of the attendee, the mute is canceled, and the impulse sound may be output from the speaker. The impulse sound other than the attendee's speech is noise, and when such noise is output from the speaker, the attendee feels uncomfortable and makes it difficult to hear the speech and disturbs the progress of the meeting. Arise.

  In view of this, in an audio conference apparatus, a device has been proposed in which noise such as an impulse sound input to a microphone is discriminated from human voice and noise is not output from a speaker. One of them is that the noise detection circuit detects the long-term average level of ambient noise from the output signal of each microphone, compares the output signal of each microphone during transmission with the output signal of the noise detection circuit, From the comparison result, the mute is canceled only when the output signal of the microphone is higher than the output signal of the noise detection circuit (see, for example, Patent Document 1).

  Further, as another conventional technique, a feature amount characterizing power and speech phonology is extracted from a large number of speech data that have been previously phoneme-labeled by a feature extraction unit, and for each initial phoneme of a word in the speech data The standard pattern and duration are created by the standard pattern creation unit, the silence average power of the silent part in the audio data is calculated by the silence pattern creation unit, and the power and feature amount are extracted from the input signal by the feature extraction unit. The similarity is calculated by the similarity calculation unit from the amount and the standard pattern, and the sum of logarithmic likelihoods for the duration of each start phoneme from the frame in which the power exceeds the silent average power is obtained by the similarity sum calculation unit. There has been proposed a voice detection method including a comprehensive determination unit that determines that a sound is within a predetermined range by comparing the sum of logarithmic likelihoods with a certain threshold (see, for example, Patent Document 2).

  In addition, the microphone output signal level and the microphone noise level are converted into digital quantities, and at least one microphone is selected in accordance with these converted outputs. If the output level of the selected microphone is large compared to the signal level of the currently selected microphone, but does not exceed the threshold value of the microphone to be selected, the microphone is not switched. There has been proposed a conference telephone device that can switch microphones without being affected by the above (for example, see Patent Document 3).

Japanese Patent Laid-Open No. 08-288999 Japanese Patent Laid-Open No. 08-87293 JP 01-24665 A

  The conventional technique as described in Patent Document 1 compares the output signal of a microphone during transmission with the output signal of a noise detection circuit. However, in the noise detection circuit, it is not easy to detect the noise separately from the voice, and in order to detect the noise accurately, for example, it is predicted that the configuration of the detection means as described in Patent Document 2 is used. However, it has a complicated configuration from both hardware and software perspectives, and there is a problem in terms of cost for practical use.

  In order to detect speech, in other words, in order to detect speech from various sound signals mixed with noise, the prior art as described in Patent Document 2 uses a feature extraction unit, standard A pattern creation unit, a silent pattern creation unit, a similarity calculation unit, a similarity sum calculation unit, and the like are required, and the configuration is complicated in terms of both hardware and software.

  The conventional technique as described in Patent Document 3 is relatively simple in terms of configuration, but when a new microphone is to be selected, the output level of the microphone to be selected is currently selected. Even if it is larger than the signal level of the microphone, if the threshold value of the microphone to be selected is not exceeded, the switching of the microphone is not performed. There are difficulties.

  The above prior art focuses on impulse sound noise peculiar to the conference hall, such as applause of the attendees, sound of hitting the table, or falling sound of objects, and there is no idea of processing that matches the characteristics, and from the speaker The probability of preventing the output of impulse sound is low.

  The present invention has been made in view of such a situation, and by devising countermeasures by paying attention to the characteristics of the impulse sound, which is the main noise generated in the conference hall, noise specific to the conference hall is output from the speaker. It is an object of the present invention to provide an audio conference apparatus that can effectively reduce this.

  The present invention relates to an audio conference apparatus, and an analog / digital converter that converts an input signal from a microphone into a digital signal, and an input data storage that temporarily stores the digital signal converted by the analog / digital converter. Means, first computing means for calculating the absolute value of the input data stored in the input data storing means, second computing means for calculating the average value of the absolute values, and the absolute value exceeding the average value The main feature is that it comprises a judging means for judging that the input signal is an impulse sound when the value probability is equal to or higher than a predetermined value.

  The present invention also relates to an audio conference apparatus, and an analog / digital converter that converts an input signal from a microphone into a digital signal, and an input that temporarily stores the digital signal converted by the analog / digital converter. A data storage means, a first calculation means for calculating an effective value of the input data stored in the input data storage means, and a second-order differential calculation using a result of the quadratic curve approximation calculation using the effective value. And a second calculating means for performing determination, and a determination means for determining that the input signal is an impulse sound when a result of the second derivative calculation exceeds a predetermined threshold with a predetermined probability or more.

  The present invention also relates to an audio conference apparatus, and an analog / digital converter that converts an input signal from a microphone into a digital signal, and an input that temporarily stores the digital signal converted by the analog / digital converter. A data storage means; a first calculation means for calculating an effective value of the input data stored in the input data storage means; a second for performing a product-sum operation with the differential filter coefficient using the effective value of the input data Computation means, and determination means for determining that the input signal is an impulse sound when a result of the product-sum operation exceeds a predetermined threshold with a predetermined probability, are provided.

  According to the present invention, the amount of change of the signal input from the microphone is calculated by the calculation means, and when the change of a certain magnitude or more occurs with a predetermined probability, the input signal is determined to be an impulse sound. By controlling the input data mute means based on the determination, even if an impulse sound such as applause is input to the microphone, it is not output from the speaker.

  Embodiments of an audio conference apparatus according to the present invention will be described below with reference to the drawings. FIG. 1 is a functional block diagram schematically showing an embodiment of an audio conference apparatus installed in a conference hall. In FIG. 1, a voice conference apparatus includes a microphone 1 that converts a speaker's voice into an electric signal and outputs the signal, and an analog / digital converter (A / D) 2 that converts the electric signal (input signal) into a digital signal. A digital signal processing device (DSP) 3 including a memory for storing the digital signal and performing signal processing according to the present invention, and a digital / analog converter (D / D) for converting the digital signal processed in the DSP 3 into an analog signal A) 9, a driver 10 that drives the speaker 11 with the converted analog signal, and a speaker 11 that outputs the analog signal.

The DSP 3 temporarily stores input data d i (i = 0, 1, 2,..., N) obtained by converting the input signal into a digital signal, and reads the input data stored in the memory 4 The absolute value calculating unit 5 that calculates the absolute value and holds the calculated absolute value, and the average that reads the absolute value held in the absolute value calculating unit 5 and calculates the average value in a predetermined number of the absolute values A value calculation unit 6, a determination unit 7 for calculating the number that the absolute value exceeds a predetermined multiple of the average value, and determining whether the calculation result is equal to or greater than a predetermined threshold value; Mute means 8 for releasing the mute of the input data read from the memory 4 according to the determination result is provided.

The average value calculator 6 is the oldest absolute value stored among the absolute values | d i | (i = 0, 1, 2,..., N) held in the absolute value calculator 5. The absolute values from (n absolute values before) to k previous values are read out, and the average value is calculated. The determination unit 7 compares the A-fold average value obtained by multiplying the average value by a predetermined value (for example, A-fold) and the absolute value from the data before k−1 to the present, and exceeds the A-fold average value. It is determined whether or not the number is greater than a predetermined threshold. If the number of absolute values exceeding the A-fold average value is larger than the threshold value, the mute in the mute means 8 is held without being released, and if the absolute value number is smaller than the threshold value, it is determined that the input data is voice. Control is performed to cancel the mute in the mute means 8.

The process of the average value calculation unit 6 will be described in more detail. In the absolute value | d i | stored in the absolute value calculation unit 5, for example, when n is 6, the absolute value calculation unit 5 has | d 0 |, | d 1 |, | d 2 | , | D 3 |, | d 4 |, | d 5 |, | d 6 |, a total of seven absolute values are stored. Here, the above-mentioned “stored oldest absolute value” is | d 6 |, and | d 6 | corresponds to “n absolute values before”. In this process, the average value calculation is performed using the absolute values from the n-th absolute value to a predetermined constant “k-th previous”. For example, when the predetermined constant k is 3, “n-th previous For the absolute value from the absolute value of k to k previous values, an average value using the absolute value | d 6 | to the absolute value | d 3 | is calculated. That is, the average value of four absolute values of | d 6 |, | d 5 |, | d 4 |, and | d 3 | is calculated. By comparing A times the average value calculated here with | d 2 |, | d 1 |, | d 0 | (k−1 absolute data to the present value), The number of absolute values exceeding the A-fold average value is obtained.

Next, the flow of determination processing by the DSP 3 will be described using the flowchart of FIG. In FIG. 2, each processing step is denoted as S1, S2,. First, the input signal converted into a digital signal in A / D2 is temporarily stored in the memory 4 (S1). Next, the absolute value calculation unit 5 reads the input data stored in the memory 4 and calculates and holds the absolute value (S2). Next, the average value calculation unit 6 reads the past n−k + 1 absolute values from the absolute values (n + 1) held in the absolute value calculation unit 5 and calculates these average values (S3). ). That is, when the current absolute value is | d 0 | and the oldest absolute value held is | d n |, the average value of absolute values of | d k | is calculated from | d n |.

Next, the determination unit 7 compares a value (A-fold average value) obtained by multiplying the average value by a predetermined constant with an absolute value after the absolute value used for the average value calculation. That is, the A-fold average value is compared with the absolute values from | d k-1 | to | d 0 | to determine the number of absolute values exceeding the A-fold average value. As a result, if the number of absolute values exceeding the A-fold average value is greater than a predetermined threshold value (YES in S5), it is determined that the input data used for the calculation is an impulse sound, and the mute means 8 An instruction is given to hold the mute (S6). As a result of comparing the A-fold average value and the absolute value, if the number is smaller than the threshold value (NO in S5), it is determined that the input data used for the calculation is voice, and the mute means 8 is muted. An instruction is given to cancel (S7).

  A mechanism for determining whether or not the input data is an impulse sound in the determination process (S5) will be described. FIG. 3 shows an example of input data held in the memory 4, where the horizontal axis represents the time axis and the vertical axis represents the signal level. 3A shows an example of an impulse sound, and FIG. 3B shows an example of a sound. As shown in FIG. 3, since the impulse sound has a sudden and large level change at the time of input, when the impulse sound is input to the microphone 1, it is converted into a digital signal as shown in FIG. Suddenly a large signal level suddenly appears, and a sudden signal level change occurs in a short time. On the other hand, the sound pressure changes more slowly than the impulse sound, and the signal level gradually increases, and the signal level does not decrease sharply.

The absolute value waveform of the signal shown in FIG. 3 is shown in FIG. 4A shows the absolute value of the impulse sound, and FIG. 4B shows the absolute value of the sound. FIG. 5A is an enlarged view of the range A in FIG. 4A, and FIG. 5B is an enlarged view of the range B in FIG. 4B. In FIG. 5, the vertical axis is the signal level, and the horizontal axis is the time axis. | D n |,..., | D k |,..., | D 1 |, | d 0 | written on the horizontal axis are sample timings of each input data, and | d 0 | It is. A diamond is plotted at a position indicating each signal level, and a black triangle is plotted as an average value. Further, the average value A times is represented by a straight line.

As shown in FIG. 5, the impulse sound has a large number of absolute values exceeding the average value d mean (A × d mean ) due to its characteristics (four in the case of FIG. 5A). For audio, there is almost no absolute value exceeding A × d mean . Since the number of absolute values exceeding a predetermined level (A × d mean ) is clearly different between impulse sound and sound when compared in the same sample interval, the microphone 1 is measured by measuring the number of absolute values exceeding a predetermined level. It is possible to determine whether the signal input to is an impulse sound or a sound. Here, the threshold value for determining the impulse sound may be the number of absolute values, or the ratio of the number of absolute values exceeding the A-fold average value among the absolute value numbers used for comparison.

  According to the above embodiment, whether the input sound is an impulse sound or a sound is determined by the determination process using the absolute value of the sound input from the microphone and the probability that the absolute value takes a predetermined value or more. Since the determination can be made, even if an impulse sound such as applause is input from the microphone, it is not output from the speaker, and there is no possibility of hindering the progress of the conference.

  Next, another embodiment of the audio conference apparatus according to the present invention will be described with reference to the drawings. FIG. 6 is a block diagram schematically showing another embodiment of the audio conference apparatus installed in the conference hall. In the voice conference apparatus according to the present embodiment, the same reference numerals are given to the same components as those of the voice conference apparatus shown in the first embodiment. Here, different configurations will be described.

  The DSP 3a shown in FIG. 6 calculates the effective value using the memory 4a that temporarily stores the input data converted into the digital signal, and the input data stored in the memory 4a, and holds the calculated effective value. An effective value calculation unit 5a, a quadratic differential calculation unit 6a that performs a quadratic curve approximation calculation using the effective value held in the effective value calculation unit 5a, and performs a secondary differential calculation on the calculated approximate curve; A probability determination unit 7a that determines whether or not a result of the secondary differentiation operation (secondary differential value) exceeds a predetermined threshold with a predetermined probability or more, and is read from the memory 4a based on the determination result of the probability determination unit 7a. Mute means 8a for attenuating the level of input data is provided.

  The probability determination unit 7a determines that the sound is an impulse sound when a plurality of secondary differential values calculated by the secondary differential calculation unit 6a exceed a predetermined threshold with a predetermined probability, and the mute means 8a Control to keep mute. If the secondary differential value does not exceed a predetermined threshold value with a predetermined probability, it is determined that the voice is a voice, and the mute means 8a is controlled to switch off so as to cancel the mute.

Next, details of the processing by the DSP 3a will be described. FIG. 7 is a flowchart showing the flow of processing according to the present invention using an input signal which is a digital signal in the DSP 3a. In FIG. 7, S10, S20,... Indicate operation steps. First, the input data d i (i = 0, 1, 2,..., N) converted into digital signals in the A / D 2 is temporarily stored in the memory 4a (S10). Next, the effective value calculator 5a uses the k pieces (k <n) of the input data d i stored in the memory 4a in order from the past input data d n to obtain n−k + 2 effective values. A value is calculated and held (S20). Next, the secondary differential calculation unit 6a uses the m effective values in order from the past effective values in the effective values held in the effective value calculation unit 5a, and uses n−k−m + 3 quadratic approximate curves. Is calculated (S30). Next, a secondary differential operation is performed on the calculated secondary approximate curve (S40). In this way, n−k−m + 3 secondary differential values are calculated, and it is determined whether or not these secondary differential values exceed a predetermined threshold with a predetermined probability or more (S50, S60). . Specifically, a predetermined threshold value is compared with the above-mentioned secondary differential value, the number of secondary differential values larger than the threshold value is counted (S50), and if this number is equal to or greater than the predetermined number (S60) YES), it is determined that the input data used for the calculation is an impulse sound, and the mute means 8a is instructed to hold the mute (S70). If the number does not exceed the predetermined number (NO in S60), it is determined that the input data used for the calculation is a voice, and the mute means 8a is instructed to cancel mute (S80). .

The processes of the effective value calculation unit 5a and the secondary differential calculation unit 6a will be described in more detail. If the input data d i (i = 0, 1, 2,..., N) stored in the memory 4a is n = 5, for example, “d 0 , d 1 , d 2 , d 3 , D 4 , d 5 ”are stored in total. Effective value computing unit 5a computes the effective value with the k input data from the previous input data d 5 of the d i in this order. Here, k is a predetermined constant. For example, if k is 2, two pieces of past input data d 5 and d 4 are read to calculate an effective value, and then the input data d 4 and d 3 are calculated. Are read out and the effective value is calculated. Thus, the calculation is performed, and finally the effective values of the input data d 1 and d 0 are calculated, and five (n−k + 2) effective values are calculated.

Next, in the calculated five effective values, a quadratic approximate curve is calculated using m effective values in order from the past effective values. Here, m is also a predetermined constant. For example, if m is 3, in the above five effective values “dj x ” (x = 0, 1, 2, 3, 4), first, the past three ( A quadratic approximation curve is calculated from m effective values (d j4 , d j3 , d j2 ), and then quadratic approximation from three (m) effective values (d j3 , d j2 , d j1 ) A curve is calculated, and finally a quadratic approximate curve is calculated from the effective values (d j2 , d j1 , d j0 ). In this way, since three (n−k−m + 3) quadratic approximate curves are obtained, n−k−m + 3 quadratic differential values are calculated.

  As already described, the waveform of the input data converted into a digital signal in A / D2 is as shown in FIG. As shown in FIG. 3A, the impulse sound has a great change in sound pressure, so that the signal level changes drastically even after being converted into a digital signal. The pressure change is gentle compared to the impulse sound.

When the voice is input data, the second derivative value (for example, a 0 ) calculated by second-order differentiation of the second-order approximation curve calculated based on the effective value is almost zero. Conversely, when the impulse sound is input data, the secondary differential value calculated thereby does not become zero, but the absolute value of the secondary differential value increases.

  Accordingly, as described above, if the number of secondary differential values exceeding a predetermined threshold in a plurality of secondary differential values (for example, n−k−m + 3) is a predetermined number or more, it is determined that the sound is an impulse sound. be able to. In addition, in a plurality of secondary differential values, the probability of calculating the number of secondary differential values exceeding the threshold and dividing this number by n−k−m + 3 is equal to or higher than a predetermined probability (for example, 50%). If it exceeds, it can be determined that the sound is an impulse sound.

  According to the above-described embodiment, a quadratic curve approximation calculation is performed using the effective value of the sound input from the microphone, and a value obtained by performing a quadratic differential calculation on the calculated quadratic curve is used. By determining the probability, it is possible to determine whether the input sound is an impulse sound or a voice, so that an impulse sound such as applause is input from the microphone, so that speaking is not permitted and the progress of the conference is hindered. No fear.

  Next, another embodiment of the audio conference apparatus according to the present invention will be described with reference to the drawings. FIG. 8 is a block diagram schematically showing another embodiment of the audio conference apparatus installed in the conference hall. The voice conference apparatus according to the present embodiment has the same configuration as the voice conference apparatus shown in the first and second embodiments, the microphone 1, A / D2, D / A 9, the driver 10, and the speaker 11, but the DSP 3b. Configuration is different. Therefore, the configuration of the DSP 3a having a different configuration will be described.

  As shown in FIG. 8, the DSP 3b temporarily stores the input data converted into a digital signal, reads the input data stored in the memory 4b, calculates an effective value, and calculates the calculated effective value. The effective value calculation unit 5b to be held, the filter processing unit 6b that reads the effective value held in the effective value calculation unit 5b and performs a product-sum operation using the effective value and the differential filter coefficient, and the calculated product-sum operation A determination unit 7b for determining whether or not the value exceeds a predetermined threshold with a predetermined probability or more; and a mute means 8b for canceling mute of input data read from the memory 4b based on a determination result of the determination unit 7b. It becomes.

  The determination unit 7b performs control for switching whether the mute unit 8b holds the mute or cancels the mute using the product-sum operation value calculated by the filter unit 6b. The determination unit 7b determines whether or not the product-sum operation value exceeds a predetermined threshold value with a predetermined probability or more. When the product-sum operation value exceeds a threshold value with a predetermined probability or more, the mute means 8b holds mute. If the product-sum operation value exceeds a predetermined probability and does not exceed the threshold value, the mute means 8b instructs the mute means 8b to cancel mute.

Next, details of the determination process by the DSP 3b will be described with reference to the flowchart of FIG. In FIG. 9, S11, S21,... Indicate operation steps. First, the input signal converted into a digital signal in A / D2 is temporarily stored in the memory 4b (S11). Next, the effective value calculating unit 5b, using the input data of the input is stored in the memory data di (i = 0,1, ··, n) k pieces from past input data d n in (k <n) The effective value is calculated in order, and the effective value is calculated in the same manner for the input data d n−1 , d n−2 ,..., And n−k + 2 effective values are obtained and held (S21). . Next, the filter processing unit 6b performs a product-sum operation with m differential filters in order from the past effective value in the effective value held in the effective value calculation unit 5b, thereby obtaining n−k−m + 3 product sums. An operation value is calculated (S31). Next, the determination unit 7b determines whether or not the product-sum operation value exceeds a predetermined threshold with a predetermined probability (S41, S51). Specifically, the number of product-sum calculation values exceeding a predetermined threshold value is counted (S41), and if the number exceeds a predetermined threshold value (YES in S51), the input data used for the calculation is an impulse sound. And the mute unit 8b is instructed to hold the mute (S61). If the number does not exceed the threshold value (NO in S51), it is determined that the input data used for the calculation is voice, and the mute unit 8b is instructed to cancel mute (S71).

  According to the above embodiment, the sum of products operation using the differential filter is performed using the effective value of the sound input from the microphone, and the input sound is an impulse sound due to the probability that the calculated value exceeds the predetermined threshold value. Therefore, when an impulse sound such as applause is input from the microphone, speech is not permitted, and there is no possibility of hindering the progress of the conference.

It is a functional block diagram which shows the Example of the audio conference apparatus concerning this invention. It is a flowchart which shows operation | movement of the Example of the audio conference apparatus concerning this invention. It is a graph which shows the characteristic of an impulse sound (a) and a human voice (b). It is a graph which shows the characteristic of the absolute value of an impulse sound (a) and a human voice (b). It is a graph which shows the characteristic of the absolute value of an impulse sound (a) and a human voice (b), and the average value of a predetermined area. It is a functional block diagram which shows another Example of the audio conference apparatus concerning this invention. It is a flowchart which shows operation | movement of the said Example. It is a functional block diagram which shows another Example of the audio conference apparatus concerning this invention. It is a flowchart which shows operation | movement of the said Example.

Explanation of symbols

1 Microphone 2 Analog to Digital Converter 3 DSP
3a DSP
3b DSP
9 Digital-analog converter 10 Speaker driver

Claims (8)

  1. An analog-to-digital converter that converts the input signal from the microphone into a digital signal;
    Input data storage means for temporarily storing the digital signal converted by the analog-digital converter;
    A first computing means for calculating an absolute value of the input data stored in the input data storing means;
    A sound comprising: second calculating means for calculating an average value of the absolute values; and determining means for determining that the input signal is an impulse sound when a probability that the absolute value exceeds the average value is a predetermined value or more. Conference equipment.
  2.   The sound according to claim 1, wherein the determination means determines that the input signal is an impulse sound when the absolute value exceeds a constant multiple of the average value with a predetermined probability or more. Conference equipment.
  3. An analog-to-digital converter that converts the input signal from the microphone into a digital signal;
    Input data storage means for temporarily storing the digital signal converted by the analog-digital converter;
    A first calculation means for calculating an effective value of the input data stored in the input data storage means;
    A second computing means for performing a second derivative using the result of the quadratic curve approximation using the effective value;
    An audio conference apparatus comprising determination means for determining that the input signal is an impulse sound when a result of the second derivative operation exceeds a predetermined threshold with a predetermined probability or more.
  4. An analog-to-digital converter that converts the input signal from the microphone into a digital signal;
    Input data storage means for temporarily storing the digital signal converted by the analog-digital converter;
    A first calculation means for calculating an effective value of the input data stored in the input data storage means;
    A second computing means for performing a product-sum operation with a differential filter coefficient using an effective value of input data;
    An audio conference apparatus comprising: determination means for determining that the input signal is an impulse sound when a result of the product-sum operation exceeds a predetermined threshold with a predetermined probability.
  5.   5. The audio conference apparatus according to claim 1, further comprising a mute unit for attenuating the signal level of the input signal determined to be an impulse.
  6.   6. The audio conference apparatus according to claim 5, wherein the input data storage unit, the calculation unit, the mute unit, and the determination unit are included in the digital signal processor.
  7.   6. The audio conference apparatus according to claim 5, comprising a plurality of microphones, and corresponding to each input signal input from each microphone, there is an input data storage unit, a calculation unit, a mute unit, and a determination unit.
  8. A digital-to-analog converter for converting the output of the mute means into an analog signal;
    The audio conference apparatus according to claim 5, further comprising a speaker that outputs an analog signal converted by the digital / analog converter.
JP2007264422A 2007-10-10 2007-10-10 Audio teleconference device Pending JP2009092994A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2007264422A JP2009092994A (en) 2007-10-10 2007-10-10 Audio teleconference device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2007264422A JP2009092994A (en) 2007-10-10 2007-10-10 Audio teleconference device

Publications (1)

Publication Number Publication Date
JP2009092994A true JP2009092994A (en) 2009-04-30

Family

ID=40665027

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2007264422A Pending JP2009092994A (en) 2007-10-10 2007-10-10 Audio teleconference device

Country Status (1)

Country Link
JP (1) JP2009092994A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110264447A1 (en) * 2010-04-22 2011-10-27 Qualcomm Incorporated Systems, methods, and apparatus for speech feature detection
JP2013531444A (en) * 2010-07-09 2013-08-01 グーグル インコーポレイテッド Method and apparatus for indicating the presence of transient noise in a call
US8898058B2 (en) 2010-10-25 2014-11-25 Qualcomm Incorporated Systems, methods, and apparatus for voice activity detection
KR101848320B1 (en) * 2016-11-24 2018-04-13 (주) 트라이너스 Method for gun sound detection and taking photograph

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110264447A1 (en) * 2010-04-22 2011-10-27 Qualcomm Incorporated Systems, methods, and apparatus for speech feature detection
US9165567B2 (en) * 2010-04-22 2015-10-20 Qualcomm Incorporated Systems, methods, and apparatus for speech feature detection
JP2013531444A (en) * 2010-07-09 2013-08-01 グーグル インコーポレイテッド Method and apparatus for indicating the presence of transient noise in a call
CN103262517A (en) * 2010-07-09 2013-08-21 谷歌公司 Method of indicating presence of transient noise in a call and apparatus thereof
US8818799B2 (en) 2010-07-09 2014-08-26 Google Inc. Method of indicating presence of transient noise in a call and apparatus thereof
KR101537080B1 (en) * 2010-07-09 2015-07-15 구글 인코포레이티드 Method of indicating presence of transient noise in a call and apparatus thereof
US8898058B2 (en) 2010-10-25 2014-11-25 Qualcomm Incorporated Systems, methods, and apparatus for voice activity detection
KR101848320B1 (en) * 2016-11-24 2018-04-13 (주) 트라이너스 Method for gun sound detection and taking photograph

Similar Documents

Publication Publication Date Title
TWI433553B (en) Automatic volume and dynamic range adjustment for mobile audio devices,and methods thereof
JP5248625B2 (en) System for adjusting the perceived loudness of audio signals
CA2571417C (en) Advanced periodic signal enhancement
US7248709B2 (en) Dynamic volume control
KR100860805B1 (en) Voice enhancement system
JP4166706B2 (en) Adaptive beamforming method and apparatus using feedback structure
KR101223830B1 (en) Hearing aid and a method of detecting and attenuating transients
CN1637857B (en) Noise eliminating system, sound identification system and vehicle navigation system
JP4764995B2 (en) Improve the quality of acoustic signals including noise
US7885818B2 (en) Controlling an apparatus based on speech
US8284947B2 (en) Reverberation estimation and suppression system
US8170879B2 (en) Periodic signal enhancement system
US9197181B2 (en) Loudness enhancement system and method
US8977545B2 (en) System and method for multi-channel noise suppression
US9336785B2 (en) Compression for speech intelligibility enhancement
CN101583996B (en) A method and noise suppression circuit incorporating a plurality of noise suppression techniques
US7680652B2 (en) Periodic signal enhancement system
DE102008027848B4 (en) Echo cancellers, echo cancellation and computer readable storage media
KR20080059147A (en) Robust separation of speech signals in a noisy environment
CN1117343C (en) Method and device for detecting voice sections, and speech velocity conversion method and device utilizingsaid method and device
JPWO2005125273A1 (en) Howling detection apparatus and method
CN102187388A (en) Methods and apparatus for noise estimation in audio signals
EP2292020A1 (en) Hearing assistance apparatus
JP4287762B2 (en) Howling detection method and apparatus, and acoustic apparatus including the same
HU0101288A2 (en) Acoustic noise reduction process on the useful signals in telecommunications, particularly in the transmission of speech