JP6119336B2 - Frequency characteristic correction apparatus and program - Google Patents

Description

  The present invention relates to a frequency characteristic correction apparatus and program, and can be applied to, for example, a smartphone.

  Currently, IP telephones using VoIP (Voice over IP) are in widespread use. There are some which realize this as an application such as a personal computer (PC), which is generally called a softphone. In recent years, portable information terminals called smartphones have become widespread, and cases where softphones are realized as applications on smartphones have become mainstream.

  However, smartphones, like PCs, are not specialized equipment for voice communications, so the frequency characteristics of voice signal output devices (earpieces, speakers) vary from model to model, and the models are the same. However, the variation among products is large, and it is necessary to correct the frequency characteristics for each product.

  In order to solve such a problem, for example, a frequency characteristic automatic correction method as described in Patent Document 1 can be applied. The conventional method as described in Patent Document 1 reproduces a signal having a frequency characteristic such as white noise (M-sequence signal in Patent Document 1) from a speaker and captures it with a measurement microphone provided at a predetermined position. Then, the captured signal is analyzed to obtain frequency characteristics, etc., parameters to be set in the equalizer are determined, and the frequency correction equalizer is set.

JP-A-7-74565

  However, the conventional technique as described in Patent Document 1 requires a specific reference signal that is a premise for analyzing the frequency, or requires installation work of equipment for measurement, etc. The environment and application conditions were very limited. For example, it is unsuitable as a method applied when a smartphone user corrects the frequency characteristics of the smartphone.

  The present invention has been made in view of the above points, and an object of the present invention is to provide a frequency characteristic correction apparatus and program capable of automatically correcting the frequency characteristic of a sound signal without the user being aware of it. .

According to a first aspect of the present invention, there is provided a frequency characteristic correction apparatus for correcting a frequency characteristic of an output sound signal that is a sound signal output from a sound output device, wherein: (1) the frequency characteristic of the output sound signal is corrected; A frequency characteristic correcting means for outputting to the sound output device side; and (2) a frequency characteristic of an input sound signal obtained by the sound capturing means capturing sound that is output from the sound output device and arrives via an echo path. and the correlation value between the frequency characteristics of the above SL output sound signal, the correlation value is not less than a predetermined threshold value, the frequency characteristic and the output sound signal of the input sound signal from which the sound capturing means is obtained by capturing And correction characteristic information forming means for forming correction information of frequency characteristics executed by the frequency characteristic correction means based on the difference from the frequency characteristics.

A frequency characteristic correction program according to a second aspect of the present invention is a frequency characteristic correction program for correcting a frequency characteristic of an output sound signal, which is a sound signal output from a sound output device, comprising: (1) A frequency characteristic correcting unit that corrects the frequency characteristic of the output sound signal and outputs the corrected sound signal to the sound output device; and (2) a sound capturing unit that captures the sound that is output from the sound output device and arrives through the echo path. There the correlation value between the frequency characteristic of the frequency characteristic and the upper Symbol output sound signal of the input sound signal obtained by capturing, the correlation value is not less than a predetermined threshold value, the sound capturing means is obtained by capturing based on the difference between the frequency characteristic of the frequency characteristic of the input sound signal and the output sound signal, functions as a correction characteristic information forming means for forming a correction information of the frequency characteristics of the frequency characteristic correcting means is executed And characterized in that.

  ADVANTAGE OF THE INVENTION According to this invention, the frequency characteristic correction apparatus and program which can correct | amend the frequency characteristic of a sound signal automatically without a user being conscious can be provided.

It is a block diagram which shows the functional structure of the frequency characteristic correction apparatus of 1st Embodiment. It is explanatory drawing which shows the various frequency characteristics used for operation | movement description of the frequency characteristic correction apparatus of 1st Embodiment. It is a block diagram which shows the functional structure of the frequency characteristic correction apparatus of 2nd Embodiment. It is a block diagram which shows the functional structure of the frequency characteristic correction apparatus of 3rd Embodiment.

(A) First Embodiment Hereinafter, a first embodiment of a frequency characteristic correcting apparatus and a program according to the present invention will be described with reference to the drawings.

(A-1) Configuration of First Embodiment FIG. 1 is a block diagram showing a functional configuration of a frequency characteristic correction apparatus according to the first embodiment.

  The frequency characteristic correcting apparatus according to the first embodiment can be configured as hardware such as a dedicated IC chip, or can be configured as software such as a CPU and a program executed by the CPU. Functionally, it can be represented in FIG. As an example of a software configuration, an example of an application mounted on a smartphone can be given.

  In FIG. 1, the frequency characteristic correction apparatus 1 according to the first embodiment includes a frequency characteristic correction unit 2 and a correction characteristic information formation unit 3.

  The frequency characteristic correction unit 2 uses the correction characteristic given from the correction characteristic information forming unit 3 as the frequency characteristic of the received signal ri (n) (where n is a parameter representing time) that is a digital signal received from the opposite device. Correction is made according to the information F (k) to obtain a reception signal (hereinafter referred to as an output reception signal) ro (n) toward a speaker (not shown). The frequency characteristic correction unit 2 is composed of a digital filter, for example. The output reception signal ro (n) is converted into an analog signal by a digital / analog conversion unit (not shown), and then output by a speaker (not shown).

  In addition to the reception signal ri (n), the correction characteristic information forming unit 3 is provided with a transmission signal s (n) obtained by converting a signal captured by a microphone (not shown) into a digital signal by an analog / digital conversion unit (not shown). The correction characteristic information forming unit 3 includes the frequency characteristic of the reception signal ri (n) and the transmission signal s (n) (in other words, an echo signal) obtained by capturing the reception signal output by the speaker with a microphone. The correction characteristic information F (k) is formed on the basis of the frequency characteristic. The correction characteristic information F (k) is information that matches the initial frequency characteristic of the reception signal ri (n) with the frequency characteristic of the transmission signal s (n).

  In general, the microphone has a flat frequency characteristic, but the frequency characteristic of the speaker is not flat. In addition, the speaker has different frequency characteristics for each model, and even if the model is the same, the variation among products is large. Therefore, when the initial received signal ri (n) (analog signal thereof) is sounded and output from the speaker, the frequency characteristics of the sound that has been sounded and output are considerably different from the frequency characteristics of the received signal ri (n). .

  Therefore, in the first embodiment, the frequency characteristic of the received signal ri (n) is corrected so that the frequency characteristic of the sound output from the speaker matches the frequency characteristic of the original received signal ri (n). The output received signal ro (n) is formed with correction according to the information F (k).

  The correction characteristic information forming unit 3 includes a reception side power normalization unit 10, a reception side FFT (fast Fourier transform) unit 11, a transmission side power normalization unit 12, a transmission side FFT unit 13, a correlation determination unit 14, and a comparative analysis. Part 15.

  Note that the transmission signal s (n), the reception signal ri (n), and the output reception signal ro (n) are each referred to as the number of samples (hereinafter referred to as a frame) for a fixed time in the correction characteristic information forming unit 3. ) And is processed every time.

  The reception side power normalization unit 10 normalizes the reception signal ri (n) with the signal power of the reception signal itself, and the transmission side power normalization unit 12 transmits the transmission signal s (n). The signal is normalized by the signal power of the signal itself. Therefore, the normalized received signal Ri (n) output from the receiver power normalization unit 10 and the normalized transmission signal S (n) output from the transmitter power normalization unit 12 are: The power will be complete.

  The receiving side FFT unit 11 converts the normalized received signal Ri (n) into frequency information (frequency characteristics) RF (k) (k represents a frame) in units of frames, and the transmitting side FFT unit 13 converts the normalized transmission signal S (n) into frequency information SF (k) in units of frames. In addition, the conversion method to frequency information is not limited to FFT. The receiving-side FFT unit 11 has a function of holding frequency information RF (k) to RF (k−N) for the most recent predetermined number of frames.

  The correlation determination unit 14 has the highest correlation with the frequency information SF (k) on the transmitting side of the current frame among the held frequency information RF (k) to RF (k−N) on the receiving side. Frequency information RF (x) is searched, and the frequency information RF (x) on the receiving side and the frequency information SF (k) on the transmitting side that have been searched are provided to the comparison analysis unit 15. Note that the frequency information RF (x) on the receiver side having the highest correlation may be handled as a search failure if the correlation is less than a predetermined threshold. This is because, for example, it is not appropriate to correct the frequency characteristics according to the first embodiment in the double talk period. In consideration of such points, instead of comparing with the threshold value, a double talk detection unit is provided so that the correction characteristic information forming unit 3 functions effectively only in the single talk period of the reception signal ri (n). good.

  The comparison analysis unit 15 subtracts the transmission side frequency information SF (k) for each frequency component from the reception side frequency information RF (x), and uses the obtained frequency information as the correction characteristic information F (k). This is given to the correction unit 2.

(A-2) Operation of the First Embodiment Next, the operation of the frequency characteristic correction apparatus 1 of the first embodiment will be described with reference to the drawings.

  In a softphone of a smartphone in which the frequency characteristic correction device 1 is mounted, when a call with the opposite telephone terminal is started, the frequency characteristic correction device 1 starts operating.

  The reception signal ri (n) from the opposing device is converted into an analog signal by a digital / analog conversion unit (not shown) after passing through the correction characteristic information forming unit 3 and is output as a sound from a speaker (not shown). The received voice that is output as a sound passes through a microphone (not shown) via an echo path, is captured by the microphone, is converted into a digital signal by an analog / digital converter, and is corrected in frequency characteristics as a transmission signal s (n). Input to the device 1.

  The received signal ri (n) is normalized by the signal power of the received signal itself by the received power normalizing unit 10 and then converted into frequency information RF (k) in frame units by the received FFT unit 11. Frequency information RF (k) to RF (k−N) for the most recent predetermined number of frames is held in the reception-side FFT unit 11.

  On the other hand, the transmission signal s (n) is normalized with the signal power of the transmission signal itself by the transmission side power normalization unit 12, and then the transmission side FFT unit 13 performs frequency information SF ( k).

  Then, among the frequency information RF (k) to RF (k−N) on the reception side held in the reception side FFT unit 11 by the correlation determination unit 14, the frequency information SF (k) on the transmission side of the current frame. ) And the frequency information RF (x) on the receiver side having the highest correlation is searched, and the frequency information RF (x) on the receiver side searched for and the frequency information SF (k) on the transmitter side are given to the comparison / analysis unit 15. It is done. In the comparison / analysis unit 15, the frequency information SF (k) on the transmission side is subtracted for each frequency component from the frequency information RF (x) on the reception side, and the obtained frequency information is used as the correction characteristic information F (k). The correction unit 2 is given.

  Thereby, the correction characteristic information F (k) is set as the frequency characteristic of the frequency characteristic correction unit 2, and the frequency characteristic of the reception signal ri (n) is corrected according to the correction characteristic information F (k).

  For example, the frequency information SF (k) on the transmission side of the current frame obtained from the transmission signal s (n) has the characteristics shown by the broken line in FIG. 2A, and is obtained from the reception signal ri (n). In addition, when the frequency information RF (x) on the receiving side having the highest correlation with the frequency information SF (k) on the transmitting side has characteristics as indicated by the solid line in FIG. 2A, the correction characteristic information F (k ) (= RF (x) −SF (k)) has a characteristic as shown in FIG.

  When the frequency characteristic of the reception signal ri (n) is corrected with the correction characteristic information F (k), the frequency component attenuated by the sound output from the speaker as seen after power normalization is amplified by the correction. Then, the frequency component amplified by the sound output from the speaker is attenuated by the correction. When the output reception signal ro (n) subjected to such correction is sounded and output, the frequency component amplified by the correction is The frequency component attenuated by the sound output by the sound and attenuated by the correction is amplified by the sound output by the speaker, and the frequency characteristic of the sound output from the speaker is the same as the frequency characteristic of the original received signal ri (n) It becomes.

  That is, with the above correction, the frequency characteristic of the sound output from the speaker regardless of the frequency characteristic variation for each speaker model or the product variation is the frequency characteristic of the original received signal ri (n). Will be the same.

  The operation in the frequency characteristic correcting apparatus 1 may be terminated when a predetermined period (for example, 10 seconds) has elapsed since the start of the call. Further, the operation in the frequency characteristic correction apparatus 1 may be executed during a call. Further, the operation in the frequency characteristic correction apparatus 1 may be terminated when the levels of all frequency components in the correction characteristic information F (k) are within a predetermined narrow range.

  In the above description, the operation of automatically enabling the operation in the frequency characteristic correction apparatus 1 is shown as the call starts. However, the user can select whether to enable the operation in the frequency characteristic correction apparatus 1 or not. The operation of the frequency characteristic correction apparatus 1 may be validated when a call is made in a state where the validation is selected (selection by a dip switch, selection from an application, etc.).

(A-3) Effect of First Embodiment According to the first embodiment, the user automatically compensates for the influence of the frequency characteristic of the speaker without being aware of the correction of the frequency characteristic during a call. It is possible to correct the frequency characteristic of the received signal and output the sound.

  In addition, according to the first embodiment, a special reference signal is not necessary due to the frequency characteristics of the received signal, and the received signal itself can be used as the reference signal, and a special reference signal such as a white noise signal is generated. A circuit can be dispensed with.

  It should be noted that providing a reference signal generating circuit and outputting the reference signal by sound generation to obtain the correction characteristic information F (k) constitutes a modified embodiment of the first embodiment. For example, a frequency characteristic correction function may be provided as one of setting functions immediately after purchasing a smartphone. In such a case, a white noise signal or the like can be used for correcting the frequency characteristic.

(B) Second Embodiment Next, a second embodiment of the frequency characteristic correcting apparatus and program according to the present invention will be described with reference to the drawings.

  FIG. 3 is a block diagram showing a functional configuration of the frequency characteristic correcting apparatus according to the second embodiment. The same or corresponding parts as those in FIG. 1 according to the first embodiment are indicated by the same reference numerals. Show.

  The frequency characteristic correction apparatus 1A of the second embodiment is mounted on a smartphone. The frequency characteristic correction apparatus 1A according to the second embodiment includes a second frequency characteristic correction unit 16 and a user attribute determination unit 17 in addition to the same configuration as the frequency characteristic correction apparatus 1 according to the first embodiment.

  The user attribute discriminating unit 17 takes in user information from an application that manages user information of the smartphone on which the frequency characteristic correction device 1A is mounted, obtains a user (user) attribute, The second correction characteristic information F2 set in the frequency characteristic correction unit 16 is obtained.

  The second frequency characteristic correction unit 16 is composed of, for example, a digital filter, corrects the frequency characteristic of the reception signal ri (n) from the opposite device according to the second correction characteristic information F2, and receives the corrected reception signal ri2 (n ) To the frequency characteristic correction unit 2 and the correction characteristic information formation unit 3.

  For example, the user attribute discriminating unit 17 includes a table that associates the age of the age with the second correction characteristic information F2, discriminates the age of the user from the user information, and the age The second correction characteristic information F2 suitable for the above is read from the table and given to the second frequency characteristic correction unit 16. As a result, the corrected reception signal ri2 (n) output from the second frequency characteristic correction unit 16 has frequency characteristics that are easily heard by people of that age. For example, if the user is a person in their 60s, the high frequency component is lifted.

  The attribute of the user is not limited to age. For example, it may be a combination of age and gender.

  According to the second embodiment, the correction of the frequency characteristic in consideration of the frequency characteristic of the speaker and the correction of the frequency characteristic in consideration of the user attribute can coexist without adversely affecting each other.

(C) Third Embodiment Next, a third embodiment of the frequency characteristic correcting apparatus and program according to the present invention will be described with reference to the drawings.

  FIG. 4 is a block diagram showing a functional configuration of the frequency characteristic correcting apparatus according to the third embodiment. The same or corresponding parts as those in FIG. 1 according to the first embodiment are indicated by the same reference numerals. Show.

  The frequency characteristic correction apparatus 1B according to the third embodiment includes a noise canceller unit 18 in addition to the same configuration as the frequency characteristic correction apparatus 1 according to the first embodiment.

  The noise canceller 18 is provided with a transmission signal s (n) obtained by converting a signal captured by a microphone (not shown) into a digital signal by an analog / digital converter (not shown). The noise canceller 18 removes a noise component included in the transmission signal s (n) and gives the transmission signal s2 (n) after the removal to the correction characteristic information forming unit 3.

  The correction characteristic information forming unit 3 basically uses the fact that the transmission signal that the reception signal wraps around through the echo path has a very high correlation with the reception signal, and the frequency of the reception signal and the transmission signal. Based on the idea that the difference in characteristics is based on the frequency characteristics of the speaker, the correction characteristic information F (k) is formed.

  However, when a smartphone equipped with the frequency characteristic correction device is under background noise, the frequency characteristic of background noise is also included in the transmission signal s (n). When the correction characteristic information F (k) is formed without removing the background noise component, the sound that is produced and output is a sound in which the frequency characteristic of the speaker and the frequency characteristic of the background noise are corrected. This is greatly different from ri (n). On the other hand, when the correction characteristic information F (k) is formed based on the transmission signal s2 (n) after the noise component is removed by the noise canceller 18 as in this embodiment, the first embodiment As described above, the sound that is generated and output is very close to the received signal ri (n) in which the frequency characteristic of the speaker is canceled.

  Whether or not to enable the function of the noise canceller unit 18 may be left to the user's selection.

  According to the third embodiment, the effects described in the first embodiment can be exhibited even in an environment with a large background noise.

(D) Other Embodiments In the description of each of the above embodiments, various modified embodiments have been referred to. However, modified embodiments exemplified below can be cited.

  In each of the above-described embodiments, the microphone for calling is applied as a microphone for obtaining the transmission signal s (n) for correcting the frequency characteristic. However, the microphone for correcting the frequency characteristic is referred to as a microphone for calling. It may be provided separately.

  In each of the above embodiments, the frequency information after being converted by the FFT unit has been shown to search for synchronization (correlation) between the transmission signal and the reception signal. The synchronization (correlation) with the received signal may be obtained. In this case, frame cutout and FFT conversion may be performed thereafter, and normalization based on signal power may be performed after obtaining the correlation.

  In each of the above embodiments, the case where there is one microphone is shown, but the present invention can also be applied to a case where there are a plurality of microphones. For example, the correction characteristic information may be obtained for each captured signal of each microphone, and average correction characteristic information obtained by averaging a plurality of correction characteristic information for each frequency component may be provided to the frequency characteristic correction unit 2.

  In each of the above-described embodiments, the frequency characteristic correction device of the present invention is mounted on a smartphone. However, the frequency characteristic correction device of the present invention can be applied to other voice communication devices such as a video conference device. Furthermore, the frequency characteristic correcting device of the present invention can be applied to devices that handle musical sound signals, acoustic signals, etc., such as music players and karaoke devices. What is expressed as “sound signal” in the scope of claims represents “sound signal”, “musical sound signal”, “acoustic signal”, and the like that are output by the sound output device.

  DESCRIPTION OF SYMBOLS 1, 1A, 1B ... Frequency characteristic correction apparatus, 2 ... Frequency characteristic correction part, 3 ... Correction characteristic information formation part, 10 ... Reception side power normalization part, 11 ... Reception side FFT part, 12 ... Transmission side power normalization , 13 ... Transmission side FFT unit, 14 ... Correlation determination unit, 15 ... Comparative analysis unit, 16 ... Second frequency characteristic correction unit, 17 ... User attribute determination unit, 18 ... Noise canceller unit.

Claims (3)

In the frequency characteristic correction device that corrects the frequency characteristic of the output sound signal, which is the sound signal output from the sound output device,
A frequency characteristic correcting means for correcting the frequency characteristic of the output sound signal and outputting to the sound output device side;
Is pronounced outputted from the sound output device, the correlation value between the frequency characteristic of the frequency characteristic of the input sound signal a sound arriving via the echo path the sound capturing means obtained by capturing the upper SL output sound signal, the When the correlation value is greater than or equal to a predetermined threshold , the frequency characteristic correcting unit executes the frequency based on the difference between the frequency characteristic of the input sound signal obtained by the sound capturing unit and the frequency characteristic of the output sound signal. A frequency characteristic correction device comprising correction characteristic information forming means for forming characteristic correction information. The correction characteristic information forming means is
An output side normalization unit that normalizes the output sound signal with the power of the output sound signal;
An output-side signal axis conversion unit that converts a normalized output sound signal that is a time axis signal into a frequency axis signal;
An input side normalization unit that normalizes the input sound signal with the power of the input sound signal;
An input-side signal axis converter that converts a normalized input sound signal that is a time axis signal into a frequency axis signal;
Correlated frequency axis signal search for searching a set having a correlation value equal to or greater than the predetermined threshold in the output side frequency axis signal from the output side signal axis converter and the input side frequency axis signal from the input side signal axis converter And
A characteristic for obtaining correction information of the frequency characteristic executed by the frequency characteristic correction means based on the difference in frequency characteristic between the output side frequency axis signal and the input side frequency axis signal searched by the correlation frequency axis signal search unit The frequency characteristic correction apparatus according to claim 1, further comprising a comparison analysis unit. A frequency characteristic correction program for correcting a frequency characteristic of an output sound signal, which is a sound signal output from a sound output device,
Computer
A frequency characteristic correcting means for correcting the frequency characteristic of the output sound signal and outputting to the sound output device side;
Is pronounced outputted from the sound output device, the correlation value between the frequency characteristic of the frequency characteristic of the input sound signal a sound arriving via the echo path the sound capturing means obtained by capturing the upper SL output sound signal, the When the correlation value is greater than or equal to a predetermined threshold , the frequency characteristic correcting unit executes the frequency based on the difference between the frequency characteristic of the input sound signal obtained by the sound capturing unit and the frequency characteristic of the output sound signal. A frequency characteristic correction program that functions as correction characteristic information forming means for forming characteristic correction information.

Images