JP2016123059A - Signal generator and terminal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a signal for calculating a highly accurate correction factors efficiently, while suppressing increase in the amount of calculation, at the time of sound field correction.SOLUTION: A signal generation unit (30) includes a smoothing unit (31), a selection unit (32) and an averaging unit (40). The smoothing unit (31) outputs N (N is an integer of 2 or more) smoothing signals by smoothing the frequency characteristics of voice, and consists of N stoichiometric bandpass filters having center frequencies different from each other. The N smoothing signals are outputted from the N stoichiometric bandpass filters. The averaging unit (40) averages the smoothing signals selected, and outputs averaged signals. The selection unit (32) selects one smoothing signal for averaging in the averaging unit, from the N smoothing signals.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a signal generation device that generates a signal for calculating a correction coefficient for correcting a sound field.

  2. Description of the Related Art Conventionally, in an audio system, sound field correction has been performed in which an equalizer or the like is used to correct the frequency characteristics of sound output from a speaker based on the analysis result of inspection sound collected through sound space. As a conventional technique relating to sound field correction, Patent Document 1 discloses a technique for automatically correcting a sound field by playing a normal music source without using a special inspection sound.

Japanese Patent Laid-Open No. 5-184000 (published July 23, 1993)

  An example of processing for calculating a correction coefficient used in a parametric equalizer (hereinafter abbreviated as PEQ) for performing sound field correction will be described with reference to FIG. FIG. 8 is a schematic diagram illustrating an example of a signal generated in the process of calculating the correction coefficient. In this example, the inspection sound is a TSP (Time Stretched Pulse) signal or the like. The number of PEQ bands is assumed to be ten.

  (Process 1) First, fast Fourier transform (hereinafter abbreviated as FFT) is performed on the inspection sound that is propagated through the acoustic space and collected by a microphone or the like. An example of a signal obtained by this processing is shown in FIG. (Process 2) Next, the signal shown in FIG. 8A is smoothed by a generally known smoothing method. An example of a signal obtained by this processing is shown in FIG. The example shown in the figure is smoothed by a 1/3 octave band filter, and the number of bands after smoothing is 32. (Process 3) Next, in order to match the number of bands with the PEQ, the signal shown in FIG. 8B is averaged. An example of a signal obtained by this processing is shown in FIG. The example shown in the figure is obtained by averaging the gains of two adjacent bands for 20 bands excluding the first 8 bands and the last 4 bands in the signal shown in FIG. 8B. is there. For example, the average of the 15th and 16th bands of the signal shown in FIG. 8B is the 4th band of the signal shown in FIG. Similarly, the fifth band of the signal shown in FIG. 8C is obtained by averaging the 17th and 18th bands of the signal shown in FIG. 8B. (Processing 4) Finally, a correction coefficient used in PEQ is generated based on the signal shown in FIG.

  In the processing described above, the correction coefficient used in PEQ can be generated while suppressing an increase in calculation amount. However, since the averaging is performed in the above-described processing 3, the generated correction coefficient is not necessarily accurate. May not be good. This is because the peaks and dips that were steep before averaging may become gradual or disappear after averaging. As a specific example, there is a steep rise from the 15th band to the 16th band in the circle shown in FIG. 8B and a prominent peak at the 16th band. There is no prominent peak in the vicinity of the fourth band in the circle shown in (c). In this case, the correction coefficient generated based on the signal shown in FIG. 8C does not take into account the peak of the 16th band of the signal shown in FIG. The sound field cannot be corrected with high accuracy.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to efficiently generate a signal for calculating an accurate correction coefficient for performing sound field correction while suppressing an increase in calculation amount. An object of the present invention is to provide a signal generation device and the like for generation.

  In order to solve the above problems, a signal generation device according to an aspect of the present invention includes a smoothing unit that smoothes frequency characteristics of speech and outputs N (N is an integer of 2 or more) smoothed signals. A smoothing unit comprising: a selecting unit that selects one smoothed signal from the N smoothed signals; and an averaging unit that averages the selected smoothed signal and outputs an averaged signal. Is composed of N constant ratio band-pass filters having different center frequencies, and outputs the N smoothed signals from each of the N constant ratio band-pass filters. A smoothing signal corresponding to the averaging process of the averaging unit is selected.

  According to the above configuration, N smoothing signals are generated by the N constant ratio band pass filters, and one is selected from the generated N smoothing signals. That is, one smoothing signal is selected from N selection branches. Conventionally, the smoothed signal had to be used regardless of whether or not a highly accurate correction coefficient can be generated from the generated smoothed signal. According to the present invention, from the N selection branches, Therefore, it is possible to select a smoothed signal that can generate a correction coefficient with relatively high accuracy.

  This selection is performed according to the averaging process of the averaging unit as described above. This process is typically a process for obtaining an average value of gains of two adjacent bands in the smoothed signal as each gain of the averaged signal. In this case, it is preferable to select a signal having the smallest gain difference between two adjacent bands among the N smoothed signals.

  Further, the above configuration is not complicated, and the process of selecting one smoothed signal after generating the N smoothed signals by the N constant ratio band pass filters requires a large amount of calculation. It is not a thing.

  Therefore, according to the present invention, with a simple configuration, it is possible to generate a signal for calculating a correction coefficient with higher accuracy than before while suppressing an increase in the amount of calculation.

  The signal generation device according to each aspect of the present invention may be realized by a computer. In this case, the signal generation device is realized by a computer by operating the computer as each unit included in the signal generation device. A control program for the signal generation device to be executed and a computer-readable recording medium on which the control program is recorded also fall within the scope of the present invention. Furthermore, the signal generation device according to each aspect of the present invention may be realized as an integrated circuit (IC chip), and in this case, a chip including the integrated circuit falls within the scope of the present invention.

  According to the present invention, with a simple configuration, it is possible to generate a signal for generating a correction coefficient for accurately correcting a sound field while suppressing an increase in the amount of calculation. As a result, it is possible to efficiently generate a correction coefficient with higher accuracy than before.

1 is a schematic diagram showing a schematic configuration of an audio system according to an embodiment of the present invention. It is a block diagram which shows the structural example of the speaker apparatus and measurement apparatus which are contained in the said audio system. It is a block diagram which shows the structural example of the signal generation part contained in the said audio system. It is a schematic diagram which shows an example of the fixed ratio type | mold band pass filter which comprises the smoothing part contained in the said signal generation part. It is a flowchart which shows an example of the flow of a process until it produces | generates a correction coefficient in the said audio system. (A) is a schematic diagram which shows an example of the frequency characteristic of the test | inspection sound collected with the said measuring apparatus. (B) And (c) is a schematic diagram which shows an example of the signal which smoothed the signal shown to (a) in the said smoothing part. (D) is a schematic diagram which shows an example of the signal produced | generated when the signal shown to (b) is averaged in the averaging part contained in the said signal generation part. (E) is a schematic diagram which shows an example of the signal which averaged the signal shown to (c) in the said averaging part. It is a block diagram which shows the structural example of the modification of the said speaker apparatus and the said measuring apparatus. It is a schematic diagram explaining a prior art. (A) is a schematic diagram which shows an example of the frequency characteristic of the test | inspection sound which collected. (B) is a schematic diagram which shows an example of the signal which smoothed the signal shown to (a). (C) is a schematic diagram which shows an example of the signal which averaged the signal shown to (b).

  An embodiment of the present invention will be described below with reference to FIGS.

  In an audio system with multiple speakers, such as a 2.0ch stereo audio system, the user can adjust the frequency and phase characteristics of the playback sound of each speaker appropriately according to the sound field. It is difficult to obtain a simple acoustic space. For this reason, such an audio system usually includes an automatic sound field correction system that automatically corrects sound field characteristics to create an optimal acoustic space. The audio system 1 according to the present embodiment includes such an automatic sound field correction system, and performs sound field correction of a space where the audio system 1 is arranged.

  The schematic configuration of the audio system 1 will be described with reference to FIG. FIG. 1 is a schematic diagram showing a schematic configuration of the audio system 1. As shown in the figure, the audio system 1 includes a speaker device 10 and a measuring device 20.

  The speaker device 10 is a sound output device including a speaker that outputs input or recorded sound to a space. Typical examples of the speaker device 10 are a television, Bluetooth (registered trademark) Audio, and a personal computer. In FIG. 1, the speaker device 10 is illustrated as a device that includes two speakers and outputs stereo (2.0 ch) sound. However, the audio system 1 is a 5.1 ch surround system, a 7.1 ch surround system, or the like. Also good.

  The measuring device 20 is a device including a microphone that is disposed at a predetermined position in the space and collects sound that is output from the speaker device 10 and propagates through the space. A typical example of the measuring device 20 is a dedicated terminal such as a remote controller or a portable terminal such as a smartphone.

  The audio system 1 generates a correction coefficient used for sound field correction. This correction coefficient is a coefficient for correcting the frequency characteristics, phase characteristics, and the like of the sound output from the speaker device 10. In generating the correction coefficient, the audio system 1 first outputs a test sound such as a TSP signal from each speaker of the speaker device 10. And the test | inspection sound output from the speaker apparatus 10 is collected in the measuring apparatus 20 arrange | positioned in the viewing-and-listening position of the user in space. And the audio system 1 produces | generates a correction coefficient based on the result of having analyzed the frequency characteristic etc. of the test | inspection sound collected by the measuring apparatus 20. FIG.

  Next, configuration examples of the speaker device 10 and the measurement device 20 will be described with reference to FIG. FIG. 2 is a block diagram illustrating a configuration example of the speaker device 10 and the measurement device 20.

  The speaker device 10 includes an audio input unit 101, an inspection sound reproduction unit 102, a reproduction sound selection unit 103, a correction unit 104, a coefficient generation unit 105, an analysis unit 106, an audio output unit 107, and a configuration example shown in FIG. A speaker 108 is provided. The coefficient generation unit 105 and the analysis unit 106 are not necessarily provided in the speaker device 10 and may be provided in the measurement device 20. Furthermore, the coefficient generation unit 105 and the analysis unit 106 may be provided in an external device that can communicate with the speaker device 10 and the measurement device 20.

  The audio input unit 101 receives an audio source, and includes an A / D (Analog / Digital) converter, a decoder, and the like.

  The inspection sound reproducing unit 102 reproduces the inspection sound. Note that the inspection sound may be stored in a storage unit (not shown).

  The reproduction sound selection unit 103 is a switcher for switching sound, and normally selects and outputs the sound from the audio input unit 101. When the inspection sound needs to be output, the reproduction sound selection unit 103 outputs the inspection sound from the inspection sound reproduction unit 102. Select and output.

  The speaker device 10 may be configured not to include the inspection sound reproduction unit 102 and the reproduction sound selection unit 103. In the case of this configuration, the inspection sound is input to the audio input unit 101 and output to the correction unit 104.

  The correction unit 104 corrects the delay, volume, frequency characteristics, and the like of the sound input to the correction unit 104 using the correction coefficient generated by the coefficient generation unit 105. In particular, the correction unit 104 includes a PEQ having a predetermined number of bands in order to correct the frequency characteristic. When outputting the inspection sound, the correction unit 104 outputs the inspection sound to the audio output unit 107 without correcting the inspection sound.

  The coefficient generation unit 105 generates a correction coefficient based on the result of the analysis performed by the analysis unit 106. The coefficient generation unit 105 may store the generated correction coefficient in a storage unit (not shown).

  The analysis unit 106 analyzes the inspection sound, which is output from the sound input unit 201 of the measurement apparatus 20 and propagated through the acoustic space, by performing FFT or the like, and analyzes the speaker characteristics of the speaker 108 or the audio system 1. The acoustic space characteristic etc. of the space where is arranged is calculated. In particular, the frequency characteristics of the inspection sound collected through propagation in the acoustic space are smoothed and averaged, and a coefficient generation signal for generating a correction coefficient used in the PEQ included in the correction unit 104 is generated. The analysis unit 106 may store the processing result in a storage unit (not shown).

  The audio output unit 107 is connected to a speaker 108 that outputs audio from the speaker device 10 to the outside, and includes a D / A converter, an amplifier, and the like.

  Next, the measuring apparatus 20 includes a microphone 200 (sound collecting unit) and an audio input unit 201 as in the configuration example shown in FIG. The microphone 200 collects sound output from the speaker device 10 and propagated through the space, and the sound input unit 201 is configured by an amplifier, an A / D converter, and the like. The voice input unit 201 outputs the test sound input from the microphone 200 and collected through the acoustic space to the analysis unit 106. The voice input unit 201 may record the inspection sound in a storage unit (not shown).

  The connection from the measuring device 20 to the speaker device 10 may be a wired connection or a wireless connection.

  Next, a configuration example of generating a coefficient generation signal in the analysis unit 106 will be described with reference to FIG. FIG. 3 is a block diagram illustrating a configuration example of a portion of the analysis unit 106 that generates a coefficient generation signal. Note that, as described above, the measurement apparatus 20 may include the analysis unit 106.

  The analysis unit 106 includes a frequency analysis unit 50 and a signal generation unit 30 (signal generation device). The frequency analyzing unit 50 performs frequency analysis by performing FFT or the like on the input inspection sound that has been collected through the acoustic space.

  The signal generation unit 30 includes a smoothing unit 31, a selection unit 32, and an averaging unit 40. The smoothing unit 31 includes N (N is an integer of 2 or more) constant ratio band pass filters. The N fixed ratio band pass filters are composed of a plurality of bands having different center frequencies. A frequency obtained by shifting the center frequency of one band of one of the N constant-ratio bandpass filters by 1 / N times the bandwidth of the band is the other N-1 constant-ratio filters. The center frequency of the band corresponding to the one band of the bandpass filter is set. The smoothing unit 31 smoothes the frequency characteristics of the test sound collected through the acoustic space using these N constant ratio band pass filters, and generates N smoothed signals. A typical example of the fixed ratio band pass filter is an 1 / M (M is an integer of 1 or more) octave band filter. Note that the number of bands of the N smoothed signals to be generated is the same and is larger than the number of bands of PEQ provided in the correction unit 104. One smoothed signal is composed of signals of each band.

  Here, a specific example of the fixed ratio band pass filter constituting the smoothing unit 31 will be described with reference to FIG. FIG. 4 is a schematic diagram illustrating an example of two constant ratio bandpass filters that form the smoothing unit 31. The constant ratio band pass filter shown in this example is a 1/1 octave band filter. The 1/1 octave band filter shown in FIG. 4 (a) consists of nine bands. Each band has an octave width and the center frequency is 63 Hz, 125 Hz,..., 16 kHz. Next, the 1/1 octave band filter shown in (b) of FIG. 4 is the same as the 1/1 octave band filter shown in (a) of FIG. The band is shifted (shifted) by 1/2 band. As shown in the figure, the center frequency of each band is 89 Hz, 177 Hz,..., 23 kHz.

  In this example, the example in which the smoothing unit 31 is configured by two 1/1 octave band filters has been described. However, when the smoothing unit 31 is configured by N 1/1 octave band filters, for example, The smoothing unit 31 may be configured by N 1/1 octave band filters in which the center frequency of each band having an octave width is shifted by 1 / N band (1 / N octave).

  Refer to FIG. 3 again. The selection unit 32 selects one of the N smoothed signals generated by the smoothing unit 31. Specifically, one smoothing signal for generating a coefficient generation signal capable of generating a correction coefficient for correcting the sound field with relatively high accuracy is selected from the N smoothing signals. In this selection, the selection unit 32 considers the averaging process performed by the averaging unit 40. When the averaging unit 40 performs a process of obtaining the average value of the gains of two adjacent bands in the smoothed signal as each gain of the coefficient generation signal, as a typical selection method by the selecting unit 32, the smoothing unit Among the N smoothed signals generated by 31, it is preferable to select a smoothed signal having the smallest maximum gain difference between two adjacent bands. In the case of this selection method, the smoothed signal having the smallest change in gain between adjacent bands is selected. In other words, a smoothed signal with relatively few steep rising and falling edges is selected, and as a result, a smoothed signal in which peaks and dips that existed before averaging tend to remain relatively after averaging. Can be elected.

  The averaging unit 40 averages the smoothed signal selected by the selection unit 32 and generates a coefficient generation signal. Since the coefficient generation signal is a signal for generating a correction coefficient used in the PEQ included in the correction unit 104, the number of bands is the same as that of the PEQ. The averaging by the averaging unit 40 may be an average that is generally known as long as the number of bands of the smoothed signal is matched with the PEQ. A typical averaging process is a process for obtaining an average value of gains of two adjacent bands in the smoothed signal as each gain of the coefficient generation signal. In calculating the average value, a generally known arithmetic expression or arithmetic table may be used.

  Next, a processing flow until the correction coefficient is generated in the audio system 1 will be described with reference to FIG. FIG. 5A is a flowchart illustrating an example of a process flow until the correction coefficient is generated in the audio system 1. As shown in the figure, first, the test sound output from the speaker 108 of the speaker device 10 is collected and recorded by the microphone 200 of the measuring device 20 (step S1). Next, the analysis unit 106 measures the delay of the inspection sound acquired from the voice input unit 201 (step S2). Subsequently, the analysis unit 106 performs gain difference measurement using the delay measurement result (step S3). Further, the analysis unit 106 calculates the frequency characteristic of the inspection sound output from the speaker 108 (step S4). Finally, the coefficient generation unit 105 generates a correction coefficient (step S5).

  Since steps S1 to S3 and S5 are generally known processes, detailed description thereof is omitted. Step S4 is a characteristic part of the present invention and will be described in detail with reference to FIG. 5B and FIG. In addition, the smoothing part 31 demonstrated below shall be comprised with two 1/3 octave band filters. The two 1/3 octave band filters are referred to as “band filter A” and “band filter B”. The center frequency of each band of the band filter B is obtained by shifting (shifting) the center frequency of each band of the band filter A by ½ band (1/6 octave), and the number of both bands is the same. Shall.

  FIG. 5B is a flowchart showing details of step S4. As shown in the figure, first, the frequency analysis unit 50 performs FFT on the inspection sound output from the voice input unit 201 of the measuring apparatus 20 (step S41). An example of a signal obtained by this processing is shown in FIG.

  Next, the smoothing unit 31 smoothes the signal obtained in step S41 with the band filter A and the band filter B, and generates two smoothed signals (step S42). An example of signals obtained by this processing is shown in FIGS. 6B and 6C. FIG. 6B is a schematic diagram showing a smoothed signal (hereinafter referred to as a smoothed signal A) obtained by smoothing the signal shown in FIG. The smoothed signal A has 32 bands, has a steep rise from the 15th band to the 16th band, and has a prominent peak at the 16th band (see the inside of the circle labeled X1). . Similarly, FIG. 6C is a schematic diagram showing a smoothed signal (hereinafter referred to as a smoothed signal B) obtained by smoothing the signal shown in FIG. The smoothed signal B has 32 bands, has a steep rise from the 15th band to the 16th band, and has a prominent peak at the 16th band (see the inside of the circle labeled Y1). .

  What should be noted here is that the rise of the 15th to 16th bands of the smoothed signal B is not as steep as the rise of the 15th to 16th bands of the smoothed signal A. In other words, the gain difference between the 15th and 16th bands of the smoothed signal B (about 8 dB) is smaller than the gain difference between the 15th and 16th bands of the smoothed signal A (about 14 dB).

  Next, the selection unit 32 selects one of the two smoothed signals A and B generated in step S42 (step S43). In this selection, the selection unit 32 considers the averaging process of the averaging unit 40. When the process of the averaging unit 40 is a process of averaging the gains of two adjacent bands in the smoothed signal, the selecting unit 32 selects two adjacent bands of the smoothed signal A and the smoothed signal B. A smoothed signal having the smallest maximum gain difference is selected. The maximum value of the gain difference between two adjacent bands of the smoothed signal A is the gain difference between the 15th band and the 16th band (about 14 dB). On the other hand, the maximum value of the gain difference between two adjacent bands of the smoothed signal B is the gain difference between the 15th band and the 16th band (about 8 dB). Therefore, in this case, the selection unit 32 selects the smoothed signal B.

  Finally, the averaging unit 40 averages the smoothed signal B selected in step S43 to generate a coefficient generation signal (step S44). (E) in FIG. 6 shows two adjacent signals from the smoothed signal B in FIG. 6 (c) except for the first 8 bands and the last 4 bands by the averaging unit 40. It is a schematic diagram which shows the signal for coefficient generation obtained by averaging the gain of two bands. For example, the average of the 15th band (about 4 dB) and the 16th band (about 13 dB) of the smoothed signal B is the fourth band of the coefficient generation signal shown in FIG. 8 dB).

  As a reference, it is obtained when the gains of two adjacent bands are averaged for signals of 20 bands excluding the first 8 bands and the last 4 bands in the smoothed signal A in FIG. 6B. The resulting signal is shown in FIG. For example, the average of the 15th band (about -3 dB) and the 16th band (about 12 dB) of the smoothed signal A is the fourth band (about 5 dB) of the signal shown in FIG. .

  What should be noted here is that the portion that was a steep peak in (c) of FIG. 6 (within the circle labeled Y1) still has a steep peak after averaging ((e ), See the inside of the circle labeled Y2). On the other hand, the steep peak existing in the circle labeled X1 in FIG. 6B becomes gentle after averaging, and no steep peak remains (the sign X2 in FIG. (Refer to the enclosed circle.) Therefore, the coefficient generation signal obtained by averaging the smoothed signal B selected by the selection unit 32 can generate a correction coefficient with higher accuracy than the signal obtained by averaging the smoothed signal A. . Thus, according to the signal generation unit 30, it is possible to generate an appropriate coefficient generation signal for performing accurate sound field correction.

  Next, a configuration example of an audio system 1a that is a modification of the audio system 1 will be described with reference to FIG. FIG. 7 is a block diagram illustrating a configuration example of the speaker device 10a and the measurement device 20a (terminal device) included in the audio system 1a. As described above, instead of the speaker device 10 including the analysis unit 106, the measurement device 20 may include the analysis unit 106. The speaker device 10a shown in FIG. 7 has a configuration in which the analysis unit 106 is removed from the speaker device 10, and the measurement device 20a has a configuration in which the analysis unit 106 is added to the measurement device 20. In the measurement apparatus 20 a, the inspection sound that has propagated through the acoustic space and collected by the microphone 200 is input to the analysis unit 106 via the voice input unit 201. Then, as shown in FIG. 3, the frequency analysis unit 50 of the analysis unit 106 performs FFT or the like on the inspection sound to perform frequency analysis. The frequency characteristic of the inspection sound output from the frequency analysis unit 50 is input to the signal generation unit 30 of the analysis unit 106. Since the subsequent processing flow is the same as described above, a description thereof will be omitted.

  The coefficient generation unit 105 and the analysis unit 106 described above may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or by software using a CPU (Central Processing Unit). It may be realized. When implemented using a CPU, the coefficient generation unit 105 / analysis unit 106 records a CPU that executes instructions of a program, which is software that implements each function, and the program and various data that can be read by a computer (or CPU). ROM (Read Only Memory) or storage device (these are referred to as “recording medium”), RAM for developing the program, and the like are provided. And the objective of this invention is achieved when a computer (or CPU) reads the said program from the said recording medium and runs it. As the recording medium, a “non-temporary tangible medium” such as a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. The program may be supplied to the computer via an arbitrary transmission medium (such as a communication network or a broadcast wave) that can transmit the program. The present invention can also be realized in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.

  The present invention is not limited to the above-described embodiment, and various modifications are possible within the scope shown in the claims, and the present invention is also applied to an embodiment obtained by appropriately combining technical means disclosed in the embodiment. It is included in the technical scope of the invention.

  The present invention can be applied to an apparatus for correcting a sound field and an audio system including the apparatus.

  DESCRIPTION OF SYMBOLS 1 Audio system, 1a audio system, 10 Speaker apparatus, 10a Speaker apparatus, 20 Measurement apparatus, 20a Measurement apparatus (terminal apparatus), 30 Signal generation part (signal generation apparatus), 31 Smoothing part (Constant ratio type band pass filter) , 32 selection unit, 40 averaging unit, 50 frequency analysis unit, 105 coefficient generation unit, 106 analysis unit, 200 microphone (sound collection unit), 201 voice input unit

Claims (5)

A smoothing unit that smoothes the frequency characteristics of speech and outputs N (N is an integer of 2 or more) smoothed signals;
A selection unit for selecting one smoothed signal from the N smoothed signals;
An averaging unit that averages the selected smoothed signal and outputs the averaged signal;
The smoothing unit includes N constant ratio bandpass filters having different center frequencies, and outputs the N smoothed signals from each of the N constant ratio bandpass filters.
The selection unit selects a smoothed signal corresponding to the averaging process of the averaging unit. Each of the N constant ratio band pass filters is composed of a plurality of bands,
The frequency obtained by shifting the center frequency of one band of one constant ratio bandpass filter by 1 / N times the bandwidth of the band is the one of the N-1 constant ratio bandpass filters. The signal generation device according to claim 1, wherein the signal generation device is set as a center frequency of a band corresponding to the band.   3. The signal generation device according to claim 1, wherein the N constant ratio band-pass filters are 1 / M (M is an integer of 1 or more) octave band filters. 4. The averaging process of the averaging unit is a process of calculating an average value of gains of two adjacent bands in the smoothed signal,
The said selection part selects the smoothing signal with the smallest maximum value of the difference of the gain of two adjacent bands among the said N smoothing signals, The any one of Claim 1 to 3 characterized by the above-mentioned. The signal generation device according to item. A signal generation device according to any one of claims 1 to 4,
A sound collection unit for collecting sound;
A frequency analysis unit that performs frequency analysis of the collected sound,
A terminal device, wherein a frequency characteristic output from the frequency analysis unit is input to the signal generation device.

Images