JP4618334B2 - Measuring method, measuring device, program - Google Patents

Description

  The present invention relates to a measuring device for measuring sound, for example, for sound correction, a method thereof, and a program executed by such a measuring device.

  For example, when listening to audio signals that are reproduced by a multi-channel audio system by outputting from multiple speakers, the audio will depend on the listening environment, such as the structure of the listening room and the listening position of the listener relative to the speakers, for example. The sound field (sound) that the listener feels varies as the balance and sound quality change. This leads to the fact that the listener at the listening position cannot feel an appropriate sound field depending on the state of the listening environment.

  Incidentally, such a problem is remarkable in an environment such as an automobile interior. Since the position of the listener is almost limited to the seat position in the car interior, the distance from the speaker is also biased, and the sound field balance is greatly disturbed due to the time difference of the arrival time of the sound from the speaker. . In addition, since the interior of a car is relatively narrow and almost sealed, reflected sounds and the like are synthesized in a complex manner and reach the listener, thereby disturbing the sound field. Furthermore, because of the limitation of the speaker mounting position, the speaker is not often arranged so as to directly reach the listener's ear, and the change in sound quality due to this greatly affects the sound field.

  Therefore, it is known to perform acoustic correction so that the user can listen to a good sound field that is as close as possible to the original sound source in a listening environment where the audio system is actually used. For this acoustic correction, for example, for the audio signal to be output from each speaker, the delay time is adjusted so as to correct the time difference of the sound reaching the listener's ear, or the sound signal reaches the listener's ear. Necessary signal processing such as equalizing correction is performed so that changes in sound quality and listening level at the stage are corrected.

In order to efficiently perform such acoustic correction, for example, it is preferable that the user (listener) perform the adjustment automatically by the apparatus, instead of making the adjustment based solely on the sense of hearing.
That is, first, the acoustic characteristics in the listening environment are measured by the acoustic correction device, and signal processing parameters for acoustic correction are set for the audio output system of the audio system based on the measurement result. . If an audio signal that has been signal-processed according to the parameters set in this way is output from the speaker, it is possible to obtain a good sound field that has been corrected and adapted to the listening environment, even if the user does not need to adjust the sound field. You can listen to the audio source.

Further, the above-described measurement of acoustic characteristics is performed, for example, as follows.
First, in the listening space, a microphone is arranged at a listening position corresponding to the position of the listener's ear. Then, the measurement sound is output from the speaker by the acoustic correction device, the output measurement sound is collected by the microphone, and the sound signal obtained by collecting the sound is sampled. In the acoustic correction device, for example, as described above, signal processing parameters for acoustic correction are obtained based on the result of performing, for example, frequency analysis processing on the sampled sound.

JP 2001-346299 A

  However, for example, pink noise is generally used as the measurement sound for the measurement as described above. For this reason, the user listens to a noise sound when performing measurement. Since noise sounds are never comfortable as a kind of sound, it is not preferable in consideration of the user.

Therefore, the present invention is a measurement method comprising a first measurement procedure and a second measurement procedure following the first measurement procedure, wherein the first measurement procedure is obtained based on different fundamental sound components. A first output procedure for outputting a plurality of required phonemes to separate speakers such that output periods overlap each other, and a plurality of spatial transmission paths for the plurality of phonemes emitted from the separate speakers, respectively Based on a first sound collection procedure for obtaining a sound signal by collecting sound and an analysis result obtained by executing a predetermined frequency analysis process on the sound signal collected by the first sound collection procedure. A setting procedure for setting a signal output for each of the separate speakers according to a sound pressure level, and the second measurement procedure includes a plurality of required plural steps obtained based on different fundamental sound components. For phonemes, the above A second output procedure for applying a characteristic corresponding to the sound pressure level set in the predetermined procedure and outputting the obtained plurality of phoneme signals to the separate speaker; and the plurality of the sound emitted from the separate speaker. A second sound collection procedure for obtaining a sound signal by collecting a phoneme signal through the plurality of spatial transmission paths, and a predetermined frequency analysis process for the sound signal collected by the second sound collection procedure. And a measurement procedure for obtaining a measurement result for a required measurement item for each of the plurality of spatial transmission paths based on the analysis result obtained as described above, and the phoneme is a predetermined sample represented by a power of 2 It is assumed that it is obtained on the basis of a fundamental sound component that is a sine wave to which an integer number of periods is applied to the number N, and one phoneme output in the first output procedure and the second output procedure is a predetermined above-mentioned Above an integer number of cycles 1 / (2 ^P of the fundamental component ) When a frequency component having a frequency (P is a natural number) is used as a virtual fundamental component, an arbitrary number of harmonic components that have a frequency on a predetermined octave number with respect to the virtual fundamental component are selected. The signal is output as a signal formed by combining the harmonic components.

In addition, as a measuring apparatus, a first output means for outputting a plurality of required phonemes obtained based on different fundamental sound components to separate speakers so that their output periods overlap with each other, and a release from the separate speakers. A first sound collecting means for collecting a plurality of sounded phonemes through a plurality of spatial transmission paths to obtain a sound signal; and a sound signal picked up by the first sound collecting means for a predetermined Based on the analysis result obtained by executing the frequency analysis processing, the setting means for setting the signal output for each of the separate speakers according to the sound pressure level, and the requirements obtained based on different fundamental sound components A second output means for applying a characteristic according to the sound pressure level set by the setting means to the plurality of phonemes and outputting the obtained plurality of phoneme signals to the separate speakers; A second sound collecting means for collecting a plurality of phoneme signals emitted from a speaker through the plurality of spatial transmission paths to obtain a sound signal; and a sound collected by the second sound collecting procedure. Measurement means for obtaining a measurement result for a required measurement item for each of the plurality of spatial transmission paths based on an analysis result obtained by executing a predetermined frequency analysis process on the signal, It is obtained on the basis of a fundamental component that is a sine wave to which an integer number of cycles is applied to a predetermined number of samples N expressed as a power, and is output by the first output means and the second output means. One phoneme is 1 / ( 2 ^{P of the} fundamental component to which a predetermined integer number of periods applies. ) When a frequency component having a frequency (P is a natural number) is used as a virtual fundamental component, an arbitrary number of harmonic components that have a frequency on a predetermined octave number with respect to the virtual fundamental component are selected. The signal is output as a signal formed by combining the harmonic components.

A program for causing a computer to execute a measurement procedure including a first measurement procedure and a second measurement procedure following the first measurement procedure, wherein the first measurement procedure includes different fundamental sound components. A first output procedure for outputting a plurality of required phonemes obtained as a basis to separate speakers so that output periods overlap each other, and a plurality of the phonemes emitted from the separate speakers respectively A first sound collection procedure for obtaining a sound signal by collecting sound through a spatial transmission path, and an analysis obtained by executing a predetermined frequency analysis process on the sound signal collected by the first sound collection procedure. Based on the result, a setting procedure for setting the signal output for each of the separate speakers according to the sound pressure level is executed, and the second measurement procedure is obtained based on different fundamental sound components. A second output procedure for applying a characteristic according to the sound pressure level set in the setting procedure to the plurality of phonemes and outputting the obtained plurality of phoneme signals to the separate speakers; and the separate speakers A second sound collecting procedure for obtaining a sound signal by collecting the plurality of phoneme signals emitted from the sound through the plurality of spatial transmission paths; and a sound signal collected by the second sound collecting procedure. And a measurement procedure for obtaining a measurement result for a required measurement item for each of the plurality of spatial transmission paths based on an analysis result obtained by executing a predetermined frequency analysis process for the phoneme. It is obtained on the basis of a fundamental component that is a sine wave in which an integer number of cycles is applied to a predetermined number of samples N expressed as a power, and is output in the first and second output procedures. One phoneme is a predetermined number 1 / (2 ^P of the fundamental component periodic number of the number applies ) When a frequency component having a frequency (P is a natural number) is used as a virtual fundamental component, an arbitrary number of harmonic components that have a frequency on a predetermined octave number with respect to the virtual fundamental component are selected. The signal is output as a signal formed by combining the harmonic components.

  Therefore, according to the present invention, unlike a pink noise, for example, a sound that can be heard audibly can be heard as a measurement sound, so that the user does not feel uncomfortable. . In addition, regardless of the use of such sound as measurement sound, the frequency analysis process for obtaining the measurement result is simplified, for example, by eliminating the need for window function processing as described above. Thus, for example, it is possible to simplify the corresponding program or to suppress the expansion of the hardware circuit scale. In addition, this results in an analysis result with higher reliability. For example, a better result can be obtained for acoustic correction using the analysis result.

Hereinafter, embodiments of the present invention will be described.
In this embodiment, the measurement apparatus according to the present invention will be described by taking as an example a case where the measurement apparatus is mounted on an acoustic correction apparatus that corrects a sound field reproduced by an audio system that supports multi-channel. That is, the present invention is applied to a measuring apparatus that measures the acoustic characteristics of a listening environment that uses the audio system for acoustic correction.
In addition, the acoustic correction apparatus according to the present embodiment is not originally provided for an audio system, but can be retrofitted to an existing audio system. In other words, there is no particular limitation on the audio system to which the acoustic correction apparatus of the present embodiment can be connected as long as a certain standard is met.
In addition, in this way, according to the indefinite audio system connected to the acoustic correction device, in this embodiment, it is possible to specify the multi-channel method supported by the audio system itself. It is assumed that the situation is not possible.
Therefore, the acoustic correction apparatus according to the present embodiment performs the preparatory measurement at a stage prior to performing the main measurement. That is, first, by the preliminary measurement, the channel configuration (speaker configuration) of the actually connected audio system is mainly specified. Note that the signal level to be output from the speaker of each channel at the time of the main measurement is determined according to the result of the preliminary measurement at this time. Then, based on the measurement result obtained by performing the main measurement, a required parameter in the signal processing is changed and set so that the sound field correction is performed.
And the measurement sound of this Embodiment demonstrated below shall be used at the time of a preparatory measurement.

First, the basic concept of the measurement sound used in this embodiment will be described with reference to FIG.
In this embodiment, in order to obtain a measurement sound, a basic sine wave is defined as shown in FIG. This basic sine wave is “a variable N indicating the number of samples is set to a predetermined value represented by a power of 2 (2 ⁿ : n is a natural number), and exactly one period is equal to the number of samples N. It is a specific sine wave on condition that it is “applicable”.
The number of samples N in the present invention is not particularly limited as long as it is a number that is a power of 2, but in the present embodiment, in the following description, 2 12 (n = 12) ) And N = 4096.
The sampling frequency Fs is 48 KHz. Thereby, the frequency of the fundamental sine wave defined in the embodiment is 48000 / 4096≈11.72 Hz. In addition, although 11.72 Hz is an approximate value to the last, it may be considered as 48000/4096 = 11.72 Hz below for convenience of explanation.

In this embodiment, another sine wave is obtained as follows based on the basic sine wave defined as described above.
Here, it is assumed that 4096 sample points corresponding to the sample number N (= 4096) of the basic sine wave are t0 to t4095 according to the time series. Then, based on the sample points t0 to t4095 of the basic sine wave, 4096 samples are collected as sample points [t0, tm, t2m, t3m,...] (If t4095 is exceeded, t0 is reached. A sine wave that circulates back is generated.
In this case, if m = 1, the sample points [t0, t1, t2, t3,. Then, if m = 2, sample points [t0, t2, t4, t6,...] Are collected, and as a result, as shown in FIG. On the other hand, a sine wave having a double cycle is obtained. That is, a sine wave in which exactly two cycles are applied to the number of samples 4096 is obtained.
Similarly, if m = 3 and sample points [t0, t3, t6, t9,...] Are collected, as shown in FIG. A sine wave having a period of 3 times and exactly 3 periods corresponding to the number of samples 4096 is obtained.
Further, if m = 4 and sample points [t0, t4, t8, t12,...] Are collected, as shown in FIG. This is a sine wave in which four periods are applied to the number of samples 4096.
In this way, by collecting the sample points like the sample points [t0, tm, t2m, t3m,...] By changing the value of the variable m (m is an integer), based on the basic sine wave, It is possible to create a sine wave in which m cycles are applied to the number of samples N (= 4096).
Hereinafter, a sine wave in which m cycles are applied to the number of samples N (= 4096) will be referred to as an “m-th order sine wave”. Incidentally, since the basic sine wave is m = 1, it becomes a primary sine wave. In the case of the present embodiment, this basic sine wave (primary sine wave (m = 1)) is 11.72 Hz. Therefore, for example, the secondary sine wave is 11.72 × 2 = 23.44 Hz, and the tertiary sine wave. Is 11.72 × 3 = 35.16 Hz, and the frequency of the m-th order sine wave is expressed by 11.72 Hz × m.

  As is well known, I / O buffers for DSPs (Digital Signal Processors), CPUs (Central Processing Units), etc. are created, or FFT (Fast Fourier Transform) operations are performed. In this case, it is preferable to set the number of samples represented by a power of 2 for the data to be processed. It is based on this that the number of samples N is the number of samples expressed as a power of 2 as described above.

Further, it is assumed that frequency analysis such as FFT is performed on the time series of the basic sine wave that is just applied to the number of samples N (= 4096) expressed by a power of 2, and the amplitude value is obtained. Then, it has a value only at 11.72 Hz which is the frequency of the mth-order sine wave, and theoretically becomes −∞ on the logarithmic axis at other frequencies. In other words, if the frequency of 11.72 Hz is the main lobe, side lobes caused by the frequency components included in the main lobe signal will not occur.
The same can be said for the second-order or higher m-order sine wave. This is because these second-order and higher-order m-th order sine waves are also waveforms that exactly fit an integer period with respect to the number of samples N, as can be understood from FIG.
Since side lobes do not occur in this way, it is not necessary to perform processing of a window function other than a rectangle, for example, in order to perform FFT on an unknown general signal sequence.

In the present embodiment, based on this, an audio signal as a “phoneme” generated based on an mth-order sine wave is used as a measurement sound source (measurement sound source) for preparatory measurement. . That is, the audio signal as the “phoneme” is used to reproduce and output the measurement sound from the speaker of the audio system. Then, when the measurement sound is output from the speaker, the sound signal picked up by the microphone is sampled as a response signal, and the frequency analysis is performed by FFT. At this time, the number of samples N and the sampling frequency Fs when sampling the response signal are N = 4096 and Fs = 48 KHz, similarly to the m-th order sine wave.
If the measurement sound output and collected sound sampling and analysis procedures are used, side lobes corresponding to the frequency of the mth-order sine wave are not generated as described above. With respect to the frequency of the signal component reproduced and output as sound, the response can be measured very accurately. As a result of the frequency analysis, when the amplitude of the frequency other than the measurement sound is obtained, the side lobe corresponding to the frequency of the m-th order sine wave cannot be generated as described above. It can be considered that the background noise level is measured. In other words, as a result of the frequency analysis, the amplitude of the frequency component as the measurement sound and the amplitude of the frequency component regarded as background noise other than the measurement sound are clearly separated without particularly processing the window function. Become. For example, a necessary measurement result as a preparatory measurement can be obtained based on a result of comparing the amplitudes of the measurement sound and the background noise.

By the way, as preparatory measurement, for each speaker (channel) that may be output as an audio system, one m-order sine wave that is appropriately selected is output as measurement sound, and sampling is performed for analysis. You can follow the procedure. However, since the measurement sound of the present embodiment is a sine wave, it can be said that the human ear can recognize a sense of pitch as compared with a sound obtained by reproducing a signal such as pink noise. Therefore, in the present embodiment, the phoneme (measurement sound) obtained based on the m-th order sine wave is not timed out as a measurement sound. In combination with both the pitch direction and the pitch direction, it is output in a form that allows humans to recognize it as a melody.
As a result, the user who is listening to the measurement sound is listening to some kind of melody (musical piece), and may have an unpleasant impression, for example, when the user simply hears pink noise or the like. In addition, entertainment will be enhanced.

In order to output a melodic measurement sound based on the m-th order sine wave, the present embodiment forms phonemes as follows.
In the present embodiment, as a basic idea, phonemes used for melodic measurement sounds are obtained as shown in FIG.
First, in FIG. 2, for example, m = 9 to 19 is selected as the variable m indicating the m-th order sine wave. This is because the frequency of the phoneme is easy to hear as a melody (musical sound) in the audible band, and the number of pitches required (the melody to be created, the number of phonemes appropriate for the measurement sound, the range, etc.) And the range set in consideration of the processing capability of the device that actually generates the phoneme (measurement sound), but is merely an example.
In addition, as a frequency f obtained based on the m-th order sine wave,

f = (48000/4096) × m × 2 ^k (Equation 1)

Define Then, corresponding to each of the 9th to 19th sine waves (m = 9 to 19), the frequency f when k = 1 is defined as a bass sound (fundamental sound). Thus, as shown in FIG. 2, the bass sound is 210.94 Hz corresponding to the ninth-order sine wave (m = 9), 234.38 Hz, 11 corresponding to the tenth-order sine wave (m = 10). Corresponding to the second sine wave (m = 11), 257.81 Hz, corresponding to the eighteenth sine wave (m = 18), corresponding to the 421.88 Hz, nineteenth sine wave (m = 19). Is 445.31Hz.

  Further, each bass sound defined as described above is associated with a frequency f corresponding to k = 2 or more for a variable k (k is an integer) as a harmonic order. In this case, five frequencies f corresponding to harmonic orders k = 2, k = 3, k = 4, k = 5, and k = 6 are associated with one bass sound. According to the above formula 1, the frequency f is a harmonic having a frequency that is higher by the octave number represented by the numerical difference (k−1) of the harmonic order k relative to the bass sound (k = 1) (hereinafter, octave harmonics). Also called). For example, the frequency of the octave harmonic of the harmonic order k = 2 is 421.88 Hz with respect to the frequency (210.94 Hz) of the bass sound corresponding to the ninth-order sine wave (m = 9), and the harmonic order k = 3 octave harmonic frequency is four times 843.75Hz, k = 6 octave harmonic frequency is 32 times 6750.00Hz, 2 octaves above the bass sound, respectively It can be seen that there is a relationship of 5 above octave.

In the present embodiment, one phoneme sets an appropriate relationship for the level of each octave harmonic (k = 2 to 6) with respect to the bass sound (k = 1), and then uses these octave harmonics as the bass sound. It is made to form by synthesizing.
In this way, by synthesizing not only the frequency component of the base sound (k = 1) but also the frequency component as its octave harmonic as one phoneme used for the measurement sound, the above-mentioned is also performed first. This means that the tone color of the phoneme can be set by setting the level relationship of each frequency component. As a result, a timbre element is added to the measurement sound as a melody by a combination of phonemes, so that the sequence of phonemes output as the measurement sound becomes more musical.
When a phoneme composed of components of the bass sound (k = 1) and its octave harmonics (k = 2 to 6) is analyzed, for example, by frequency analysis, the frequency of the bass sound and the octave harmonics (k = A total of six amplitudes with a frequency of 2 = 6) are detected. This means that a plurality of frequencies are measured simultaneously. Thus, the fact that a plurality of frequencies are measured at the same time leads to an increase in the density of presence of the frequency to be measured within a certain frequency band range. For example, some speakers have a so-called dip characteristic in which the sound pressure level sharply decreases in a specific frequency band. If the loudspeaker is like this, if the frequency of the measurement sound falls within the band where dip occurs, sufficient amplitude will not be observed as an analysis result. It can no longer be obtained. Therefore, as in this embodiment, if different frequencies are synthesized simultaneously as the phoneme of the measurement sound, even if a certain frequency component in the phoneme is within the dip band, it is outside the other dip band. The frequency component can be observed with a sufficiently large amplitude, and a measurement result that does not impair the reliability can be obtained.

For confirmation, for each of the octave harmonics for which the harmonic order k ≧ 2 with respect to the bass sound (k = 1), similarly to the bass sound, Therefore, the phoneme composed of the base sound and its octave harmonics does not deviate from the definition of “a waveform in which the integer number of cycles corresponds to the number of samples N”.
In addition, the bass sound is indispensable as an element of the frequency component forming the phoneme, but includes, for example, all five octave harmonics corresponding to the harmonic orders in the range of 2 ≦ k ≦ 6 shown in FIG. There is no need.

In this case, there are 11 different pitches as phonemes with the base tone frequency corresponding to the order m = 9 to 19 in FIG. However, considering that the output sequence of the phoneme as the measurement sound is melodic, the pitch (frequency) of each phoneme should have a pitch difference corresponding to a scale of a certain temperament, for example. It will be.
Therefore, in this case, let us consider a case where a 12-tone average temperament is adopted as the temperament. In this case, attention is paid to the fact that the bass sound corresponding to m = 19 is 445.31 Hz. For example, if A = 445 Hz is defined as a standard scale based on the absolute pitch name, the base sound corresponding to this order m = 19 is 445.313 Hz, and its error is slight, so this order m = 19. It can be said that the corresponding bass sound may be treated as the A sound.

Assuming that the bass sound frequency 445.313 Hz corresponding to the order m = 19 is the sound of A, as a result, the bass sound that can be handled as a sound that falls within this scale is as follows.
Bass sound corresponding to order m = 10 (234.38Hz) → A #
Bass sound corresponding to order m = 12 (281.25 Hz) → C #
Bass sound corresponding to order m = 15 (351.56Hz) → F
Bass sound corresponding to order m = 16 (375.00Hz) → F #
Bass sound corresponding to order m = 17 (398.44Hz) → G
Bass sound corresponding to order m = 18 (421.88Hz) → G #

If the frequency 445.313 Hz is regarded as the sound of A as described above, the sound of A # is 235.896 Hz, the sound of C # is 280.529 Hz, and the sound of F is 353.445, as shown in FIG. The sound of Hz and F # is 374.462 Hz, the sound of G is 396.728 Hz, and the sound of G # is 420.319 Hz. Each of the bass sounds corresponding to the above orders m = 10, m = 12, m = 15, m = 16, m = 17, m = 18 is A #, C #, F, F #, G, It is close to the average temperament approximate sound frequency of G #. Therefore, each bass sound can be regarded as a sound of A #, C #, F, F #, G, and G #, respectively.
Therefore, in the case of FIG. 2, the phoneme obtained by synthesizing the octave harmonic based on the bass sound (234.38 Hz) corresponding to the order m = 10 is defined as A #, and the same applies to the order m = 12. A phoneme based on the base tone (281.25Hz) is C #, a phoneme based on the base tone (351.56Hz) corresponding to the order m = 15 is F, and a base tone (375.00Hz) corresponding to the order m = 16 is used. The base phoneme is F #, the phoneme based on the base tone (398.44 Hz) corresponding to the order m = 17 is G, and the phoneme based on the base tone (421.88 Hz) corresponding to the order m = 18 is G #. A phoneme based on the bass sound (445.31 Hz) corresponding to the order m = 19 is used as A.
In practice, it is confirmed that the scale obtained by the phonemes selected in this way is not audibly uncomfortable under the application of outputting the measurement sound melodyally. Yes.

  FIG. 3 shows frequency characteristics for each phoneme corresponding to the seven pitch names A #, C #, F, F #, G, G #, and A selected by the method described with reference to FIG. As can be seen from this figure, the measurement target obtained by these phonemes is the pitch name that becomes the highest frequency component from 235.896 Hz of the bass sound (k = 1) corresponding to the pitch name A # that becomes the lowest frequency component. It can be said that 42 (= 7 × 6) measurement target frequencies exist almost uniformly in the band range up to 14250.00 Hz of octave harmonics (k = 6) corresponding to A. This means that there are a sufficient number of measurement target frequencies in the frequency range of the measurement target, and the existence thereof is not biased in terms of bandwidth. As a result, for example, a stable and highly reliable measurement result can be obtained in spite of the speaker dip described above.

  The method of selecting phonemes in the present embodiment is based on the method described with reference to FIG. However, for example, as described above, the scale of the phoneme that can be used to form the scale as it is in accordance with the description of FIG. There are only six sounds F, F #, G, G #, and A. In consideration of creating a melody from a phoneme sequence as a measurement sound, it is preferable to obtain as many musical scales as possible.

Therefore, in the present embodiment, based on the method described with reference to FIG. 2, actually, a phoneme that can be used as the melody of the measurement sound is determined as shown in FIG.
First, a sine wave having a frequency half that of the basic sine wave shown in FIG. 1 is defined as a virtual basic sine wave. Then, the virtual bass sound shown in FIG. 4 is defined as the m-order sine wave for the virtual fundamental sine wave.
In this case, as the frequency f obtained based on the m-th order sine wave,

f = (48000/4096) × m × 2 ^(k−1) (Formula 2)

The virtual bass sound has a frequency f obtained by substituting k = 0 for each m-th order sine wave. Similarly to the above, the frequency obtained by substituting k = 1 is used as the bass sound. In other words, the virtual bass sound is such that the term (2 ^(k−1) ) in Equation 2 is halved by setting k = 0, and is 1 for a fundamental sine wave of k = 1. The frequency is / 2.
Here, based on the virtual bass sound, 26 frequencies in a range from 105.469 Hz corresponding to m = 18 to 251.953 Hz corresponding to m = 43 are set as candidates.

  In this case, the octave harmonics correspond to k = 1, k = 2, k = 3, k = 4, k = 5, and k = 6 for each virtual bass sound (k = 0). The frequency to be used is made to correspond.

  Here, the virtual bass sound is an mth-order sine wave with respect to a virtual sine wave having a wavelength (1/2 period) twice that of the original basic sine wave shown in FIG. Therefore, the odd-order sine wave (when m is an odd number) as the frequency of the virtual bass sound does not fit in an integer number of cycles with respect to the number of samples N. In addition, the virtual bass sound with k = 0 is generated based on a virtual sine wave having a wavelength twice that of the original basic sine wave. Therefore, the waveform data cannot be generated on the basis of the basic sine wave. Therefore, in the present embodiment, the virtual bass sound itself should be excluded from the actual phoneme components.

Then, as a phoneme element corresponding to each order m of the sine wave, what can be obtained as a real sound is an octave harmonic from k = 1 or higher. Therefore, an actual bass sound that forms a phoneme is an octave harmonic of k = 1 which is the lowest value among k = 1 to 6.
A comparison is made between the list of bass sounds that are octave harmonics of k = 1 and the list of bass sounds of k = 1 shown in FIG. Then, in the case of FIG. 4, each m-th order of the base sound of k = 1 shown in FIG. 2 is based on the virtual base sound having a frequency ½ of the original fundamental sine wave. It can be seen that in addition to the frequency, the intermediate frequency is also obtained as the bass sound. That is, the number of bass sounds in a predetermined frequency range is increased almost twice.

In this case, paying attention to the fact that the bass sound corresponding to m = 38 is 445.31 Hz, it is defined that A = 445 Hz is the standard as the scale based on the absolute pitch name. Correspondingly, based on the result of comparing the relationship between the frequency of the bass sound (k = 1) shown in FIG. The scale shown in the column of approximate absolute pitch names can be associated. That means
Bass sound corresponding to order m = 19 (222.656Hz) → A
Bass sound corresponding to order m = 20 (235.896Hz) → A #
Bass sound corresponding to order m = 21 (249.923Hz) → B
Bass sound corresponding to order m = 24 (280.529Hz) → C #
Bass sound corresponding to order m = 27 (314.883Hz) → D #
Bass sound corresponding to order m = 30 (353.445Hz) → F
Bass sound corresponding to order m = 32 (374.462Hz) → F #
Bass sound corresponding to order m = 34 (396.728Hz) → G
Bass sound corresponding to order m = 36 (420.319Hz) → G #
Bass sound corresponding to order m = 38 (445.313Hz) → A
Bass sound corresponding to order m = 40 (466.164Hz) → A #
Bass sound corresponding to order m = 42 (493.883Hz) → B
Can be defined as
In this way, assuming a virtual bass sound, based on the frequency of the bass sound, which is an octave higher harmonic by one octave, A, Twelve sounds of A #, B, C #, D #, F, F #, G, G #, A, A #, and B can be used. That is, as compared with the basic method shown in FIG. 2, the number of phoneme pitches necessary for creating a melody is also increased.
For confirmation, in this case as well, for each of the above 12 sounds, one octave harmonics from k = 2 to 6 are synthesized with k = 1 bass sound. The point at which phonemes are generated is as described with reference to FIG.
The virtual bass sound here is assumed to be a sine wave having the frequency (f) of the m-th order sine wave obtained by substituting k = 0 in Equation 2. Therefore, as a concept of the present invention, the virtual bass sound is not limited to a sine waveform having a frequency ½ of the m-th order sine wave of the basic sine wave as shown in FIG. In other words, the virtual bass sound is the m-th order sine wave frequency obtained by substituting an arbitrary negative natural number smaller than 0 for the variable k. In other words, the fundamental tone (k = 1) of the virtual bass tone is the frequency of 1 / (2 ^P ) (P is a natural number) of the basic sine wave (specific frequency component) shown in FIG. It can be said that it is what has.

FIG. 5 schematically shows a basic procedure example in the case where measurement is performed using the phoneme selected as the melody element of the measurement sound as described above.
FIG. 5A shows a measurement sound output sequence. This indicates the timing of outputting a phoneme signal to the audio signal output system in order to output the phoneme as the measurement sound from the speaker.
As an example in this case, first, a phoneme corresponding to the pitch F is continuously output twice as a measurement sound in a period t0 to t3 → a period t3 to t6. Here, since one phoneme is composed of a frequency component of a sine wave with an integer number of periods corresponding to the number N of time-series samples, an output period of one phoneme (periods t0 to t3, periods t3 to t6). Also corresponds to the number N of time-series samples.
Further, in this case, after the output of the phoneme having the pitch F at the time point t6, the phoneme corresponding to the pitch A # is continuously repeated twice from the period t6 to t9 to the period t9 to t12. Output.
That is, here, a phoneme with one pitch is output by looping a signal with the number of samples N twice.
In this embodiment, since the number of samples N = 4096 and the sampling frequency Fs = 48 KHz, the time length corresponding to the number of samples N is
4096/48000 ≒ 0.085 (seconds)
It becomes.

The sound of the phoneme output in the space from the speaker as described above reaches the microphone installed at the appropriate sound collection position at the timing shown in FIG. 5B, and this reached sound is transmitted by the microphone. Sound will be collected.
As can be seen by comparing the sound collection timing in FIG. 5B and the measurement sound output sequence in FIG. 5A, the phoneme as the measurement sound output from time t0 has passed the delay time Td. From time t1, sound collection on the microphone side is started. The delay time Td is, for example, a so-called system delay time from when a signal as a phoneme is input to the audio signal output system to when the sound is emitted from the speaker, and reaching the microphone after the audio is output from the speaker. The time of the spatial transmission delay that occurs in accordance with the positional relationship (distance) between the speaker and the microphone is determined.

In this case, the sound collection timing of the phoneme corresponding to the pitch F is the period t1 to t7 as shown in FIG. Note that the length of the sound collection period of the periods t1 to t7 corresponds to the phoneme output periods t0 to t6 as the pitch F. The sound collection period of the periods t1 to t7 is divided into two equal parts by the periods t1 to t4 and the periods t4 to t7, and each of the equally divided periods corresponds to the number of samples N.
In addition, the sound collection timing of the phoneme corresponding to the pitch A # is performed during the period t7 to t13. The periods t7 to t13 can also be regarded as being divided into two equal parts by the periods t7 to t10 and the periods t10 to t13, which correspond to the number of samples N.

In order to measure the sound signal obtained by collecting the sound with the microphone, it is necessary to sample the sound signal and obtain it as a response signal. This sample timing is shown in FIG.
First, in response to the phoneme corresponding to the pitch F output by the number of consecutive samples N in the period t0 to t6, the sample delay is started from the time t0 that is the output start time of the phoneme corresponding to the pitch F. Sampling is started from time t2 shifted by time Tdrs. Sampling started from time t2 is ended at time t5 when a time corresponding to the number of samples N has elapsed from time t2. That is, here, sampling is performed by the number N of samples. The timings of the periods t2 to t5 are within the periods t1 to t7 in which the sound of the phonemes corresponding to the pitch F is collected. Thereby, depending on the sampling in the periods t2 to t5, sampling data based on the number N of samples whose phonemes corresponding to the pitch F are to be measured can be obtained.
In the same manner as in the case of the pitch F, the next sampling timing is sampled from the time point t8 shifted by the sample delay time Tdrs from the time point t6 that is the output start time of the phoneme corresponding to the pitch A #. Start. At this time t11, sampling of the number N of samples is finished. Thereby, depending on the sampling in the period t8 to t11, sampling data based on the number N of samples with the phoneme corresponding to the pitch A # output in the period t6 to t12 as the measurement target is obtained.

Here, in FIG. 5, the sample delay time Tdrs corresponds to the start time of the sampling period for obtaining the sampling data for measuring the phoneme from the time when the output of a certain phoneme is started, It will be decided.
This sample delay time Tdrs should be set so as to obtain a sampling period in which only the phonemes to be measured can be sampled reliably. For example, considering the correspondence to the phoneme corresponding to the pitch F in FIG. 5, only the phoneme corresponding to the pitch F is reliably sampled as a measurement target during the sampling period t2 to t5. The phonemes that are not subject to measurement, such as phonemes corresponding to the pitch A # collected after the time t7, are surely fall within the period t1 to t7 so as not to be sampled. In this way, it should be set. In this case, the sampling period t8 to t11 corresponding to the phoneme corresponding to the pitch A # is also determined by the sample delay time Tdrs having the same time length as that corresponding to the phoneme corresponding to the pitch F. Thereby, only the phoneme corresponding to the pitch A #, which is obtained as the collected sound signal in the period t7 to t13, can be obtained as the measurement target.
Further, in actuality, the sample delay time Tdrs is obtained by estimating the delay time Td assumed to occur in the environment assuming the environment in which the acoustic correction apparatus of the present embodiment is used, and the obtained delay time. It can be set based on Td. For example, if the acoustic correction apparatus is compatible with an in-vehicle audio system, it is possible to obtain one delay time Td from a general environment in a car.

Note that, for example, the audio signal sampled in the sampling period of the period t2 to t5 includes a period of the first half and the second half of the sample number N, with the time point t4 being a continuous point of the sample number N as a boundary. By sampling by the number of samples N, only frequency components that can be accommodated by an integer number of cycles with respect to the number of samples N are obtained as sampling data. That is, a frequency analysis result in which no side lobe is generated with respect to the main lobe is obtained. Incidentally, even when sampling is performed with the number of samples N, when a phoneme that should not be measured is sampled (in the case of FIG. 5, for example, the time t7 is included in the sampling period of the period t2 to t5). When the phoneme corresponding to the pitch F is sampled in the first half and the phoneme corresponding to the pitch A # is sampled in the second half), side lobes will be generated.
This also shows that the output period of phonemes corresponding to one sampling period must be longer. In the case of the present embodiment, the phoneme output period and sampling period have the minimum number of time-series samples N. In addition, if the relationship between the sampling period and the phoneme output period is satisfied, when the sampling period is expressed by the number of samples N × a (a is a natural number), the corresponding phoneme output period is as follows. The number of samples N × (a + b) (the variable b is a natural number of 1 or more) is set.

FIG. 6 schematically shows an example of the band characteristics obtained by performing frequency analysis by FFT on the response signal sampled by the procedure shown in FIG. In this case, for example, the result of sampling and analyzing by FFT for a single measurement target sound using only phonemes corresponding to one pitch is shown.
Assuming that the sound to be measured by a single phoneme is collected and sampled and FFT is performed, as shown in the figure, the base sound (k = 1), the second-order octave harmonic (k = 2), the second Some amplitude values for the third-order octave harmonic (k = 3), the fourth-order octave harmonic (k = 4), the fifth-order octave harmonic (k = 5), and the sixth-order octave harmonic (k = 6) Will be obtained.

  Here, in the present embodiment, a measurement sound having a sine wave that falls within an integer number of cycles with respect to the number of samples N is output and collected, and the sound signal of the collected phonemes is similarly sampled. Sampling is performed by N. Therefore, as can be understood from the above description, for example, assuming that the sampling data is an ideal audio signal using only phonemes, the frequency analysis result by FFT indicates that the measurement target frequency forming the phoneme is the main lobe. It only has a value and no side lobe is generated.

  However, as an actual frequency analysis result by FFT, for example, as shown in FIG. 6, the amplitude is detected at the frequency around the measurement target frequency, which is the base sound and the next octave harmonic. It becomes a state. If frequency analysis is performed by FFT on a phoneme-only signal, there should be no amplitude other than the frequency that forms the phoneme, so the amplitude of the frequency other than the frequency to be measured is considered to be so-called background noise in the measurement environment. It will be good. In the present embodiment, as described above, such an analysis result can be obtained without performing window function processing.

Based on the analysis result shown in FIG. 6, for example, in this embodiment, the ratio between the measurement target frequency and the level of background noise existing in the neighboring frequency is obtained. That is, the S / N ratio is obtained by setting the amplitude of the measurement target frequency as the signal (S) and the amplitude of the background noise as the noise (N).
The method for calculating the S / N ratio here is not particularly limited as long as it is based on the amplitude of the measurement target frequency and the amplitude of the background noise, but here the noise to be compared with the measurement target frequency As the level, the frequency of background noise having the largest amplitude value in the vicinity frequency of the measurement target frequency is adopted. For example, taking the bass sound shown in FIG. 6 as an example, assuming that the amplitude value of the bass sound is L1, the amplitude value of the background noise of the neighboring frequency is lower than the bass sound as shown in the figure. Let L2a be L2 higher than the amplitude value L2a on the frequency side higher than the bass sound. At this time, L2 is adopted as the amplitude of the background noise for calculating the S / N ratio, and for example, the S / N ratio is obtained by calculating L2 / L1.
Such calculation of the S / N ratio is performed in the same manner for each order octave harmonic other than the bass sound, for example. Thereby, it is possible to obtain information on the S / N ratio in the six target frequency bands corresponding to the bass sound and the second to sixth harmonics.

In addition, as a technique for obtaining an S / N ratio (evaluation of noise) other than the above, one of the methods is to weight the amplitude value of each target frequency according to Log (exponential function), and then noise. The comparison with the amplitude value of the frequency is mentioned. At this time, the weighting coefficient may be changed according to a predetermined rule by adapting for each target frequency.
Alternatively, it is also conceivable to obtain an average value for the amplitude value of noise that is a neighboring frequency of the target frequency, and calculate the S / N ratio from this average value and the amplitude value of the target frequency.
Further, in calculating the S / N ratio, it is conceivable to employ a method of comparing with a linear axis instead of comparing with an amplitude value as a dB value.

  According to the method described above with reference to FIG. 4, phonemes corresponding to a total of 12 pitches are obtained in outputting the measurement sound in a melody manner. Actually, when creating a melody by measurement sound (measurement sound melody), a phoneme corresponding to an arbitrary pitch can be selected from the phonemes corresponding to the above 12 pitches and combined. That's right.

FIG. 7 shows an example of an output pattern of phonemes when a measurement sound melody is created using phonemes corresponding to 12 pitches selected by the method described with reference to FIG. 4 as selection candidates.
In this case, one unit of the measurement sound melody output period shown in FIG. 7 is divided into a primary analysis mode, a secondary analysis mode, and a non-analysis mode as time elapses. Further, in the output period of one phoneme here, the number of samples N is continuous twice as in the case described with reference to FIG. In the case of this embodiment where the output period Ta is the number of samples N = 4096 and the sampling frequency Fs = 48 KHz,
4096/48000 × 2 = 0.17 (seconds)
It becomes.
Also, the sampling timing (sampling period) corresponding to the output of the measurement sound melody is assumed to be sampled by the number of samples N as described with reference to FIG. 5, and also as described with reference to FIG. This is determined according to the determined sample delay time Tdrs. That is, here, the sampling timing is set so that only the phonemes output in each output period Ta are sampled, and the phonemes output in the preceding and subsequent output periods Ta are not sampled.

  Further, in FIG. 7, during the period in which the phoneme is output, the channel of the speaker to be measured that is selected as the output of the sound of the phoneme is shown. Here, the speaker channels are center channel (C), front left channel (L), front right channel (R), left surround channel (Ls), right surround channel (Rs), left back surround channel (Bsl), There are 7 right back surround channels (Bsr). That is, the maximum channel configuration that can actually be handled by the acoustic correction apparatus according to this embodiment is an audio system having such a 7-channel configuration.

  According to the measurement sound output sequence shown in FIG. 7, first, in the primary analysis mode, the output period Ta continues four times. First, in the first output period Ta, only the phonemes corresponding to the pitch G # are output by the center channel (C). In the next second output period Ta, phonemes corresponding to the pitch F from the front left channel (L) and phonemes corresponding to the pitch G # are output from the front right channel (R). In the next third output period Ta, phonemes corresponding to the pitch C # are output from the left surround channel (Ls), and phonemes corresponding to the pitch F # are output from the right surround channel (Rs). In the final fourth output period Ta, phonemes corresponding to the pitch C # are output from the left back surround channel (Bsl), and phonemes corresponding to the pitch G # are output from the right back surround channel (Bsr).

Also, in the secondary analysis mode, the output period Ta is assumed to be four consecutive times, and for each output period, as shown in the figure, a phoneme corresponding to a specific pitch is output from a specific speaker. Output from the channel.
In this case, the output period Ta is continued four times even in the non-analysis mode, and for each output period, a phoneme corresponding to a specific pitch is extracted from a specific speaker channel as shown in the figure. It is output.

  According to the output sequence shown in FIG. 7, for example, first, in each stage of the primary analysis mode and the secondary mode, a measurement sound (phoneme) with some pitch is associated with each speaker corresponding to seven channels. It will be output. As a result, it is possible to perform measurement without leaking all the speakers in both the primary analysis mode and the secondary analysis mode within the range of the channel configuration that can be handled as the acoustic correction apparatus of the present embodiment. .

Depending on the output period Ta, phonemes corresponding to different pitches are output from a plurality of speakers. In other words, it is a chord as a way of hearing in space. In this way, in the present embodiment, the output of the measurement sound is made musical by combining the phonemes in both the time direction and the scale direction to obtain a required output pattern.
Even if the output of the phoneme that is the measurement sound is in the chord state, if the sound is picked up and subjected to frequency analysis by FFT, the frequency components that form each phoneme that forms the chord (the base sound and the octave) Harmonics) can be obtained, so there is no problem in the measurement process.
In addition, since there is a period of output as a chord in this way, the melody formed by the measurement sound can be more musical and the entertainment for the user is also increased. Become.

  In the primary analysis mode, the phoneme level to be output from each speaker in the secondary analysis mode is determined based on the result of frequency analysis (FFT) of the phoneme to be measured output from each speaker. To be. Thereby, in the secondary analysis mode, the measurement sound (phoneme) can be output from each speaker at a level appropriate for the preparation measurement. And also as a secondary analysis mode, as shown in FIG. 7, the phoneme of a measuring object is output from each speaker, a frequency analysis (FFT) is performed, and the measurement result as a preparatory measurement based on these analysis results To get to.

In obtaining the measurement result by the processing in the primary analysis mode and the secondary analysis mode, for example, as described above with reference to FIG. 6, the measurement target frequency is in the amplitude value and its peripheral frequency. A value of the S / N ratio calculated based on the amplitude value of the background noise can be used.
Based on the information on the S / N ratio, various determinations and settings can be made as measurement results as follows.
First, by comprehensively using the value of the S / N ratio corresponding to each frequency component forming the phoneme output for each speaker, the reproduction frequency band characteristic of the speaker can be estimated. Further, since the output sound pressure level with respect to a certain input level also changes corresponding to the aperture size of the speaker, the aperture size of the speaker can also be estimated. Also, as a matter of course, for example, although a phoneme is output with a sufficient gain for a certain speaker, the S / N ratio analyzed from the response signal of this phoneme is below a certain level, When it is considered that the level of (S) is hardly obtained, it can be determined that the speaker is not connected. That is, the audio channel configuration of the audio system can be determined.
In addition, as an example of the present embodiment, a case where it is applied to a preliminary measurement that is a preliminary step for the main measurement is described. It is possible to estimate and set the level (gain) of (in the case of the actual measurement, not necessarily the measurement sound by the phoneme of the present embodiment). Further, as a process in the primary analysis mode, it can be used to set a synthesis balance of frequency components and a phoneme output level (gain) for phonemes to be output from each speaker in the secondary analysis mode.
Further, for example, if the noise amplitude value is very large and the S / N ratio is below a certain level, it can be determined that the environment does not provide a significant measurement result. In response to obtaining such a determination result, for example, the measurement may be interrupted, and a message that prompts the user to improve the listening environment may be transferred to, for example, display. Conceivable.

In FIG. 7, in the non-analysis mode following the secondary analysis mode, as shown in the figure, the center channel (C), the front left channel (L), the front right channel ( A phoneme as a pitch G # is output from the three speakers R). At the same time, phonemes corresponding to the pitch F are output from the speakers of the left surround channel (Ls) and the right surround channel (Rs), and the speakers of the left back surround channel (Bsl) and the right back surround channel (Bsr) are output. To output phonemes corresponding to the pitch C #.
In this non-analysis mode, the response signal for the phoneme output as described above is not sampled. That is, in the non-analysis mode, frequency analysis and measurement based on the phonemes output at this time are not performed.
Here, in the measurement sound melody output period formed by the continuation of the primary analysis mode, the secondary analysis mode, and the non-analysis mode, the output of the phonemes in FIG. As can be seen from the pattern, the pitch changes with the time of the output period Ta as the minimum note, but in the non-analysis mode, the triads by C #, F, and G # are all notes. By outputting the sound like a melody, you can feel the end. That is, the non-analysis mode is not actually used for measurement, but outputs a phoneme for the purpose of enhancing musical elements of the measurement sound melody. From this, it can be said that in this embodiment, it is not always necessary to sample and analyze all the phonemes output from the speaker as response signals.

FIG. 8 is a flowchart showing the flow of processing as a preparatory measurement according to the measurement sound melody output sequence shown in FIG.
In the procedure shown in this figure, first, background noise is checked in step S101. When the background noise is checked, the phoneme is not output, and the audio signal collected by the microphone at this time is sampled to perform frequency analysis (FFT). Thereby, first, the presence or absence of background noise can be checked by looking at the amplitude of background noise. In a typical listening environment, there can be no background noise at all. Therefore, if it is recognized that there is no background noise as a result of the background noise check in step S101, it is estimated that the sound collecting microphone is not connected to the sound correction device. It will be good. Therefore, as an actual example of the acoustic correction apparatus according to the embodiment, for example, when a determination result indicating that there is no background noise is obtained as a result of the background noise check in step S101, for example, the user is prompted to connect a microphone. Messages are displayed and output by voice. If it is determined that the background noise exists as a result of the background noise check, the microphone is connected, and the process proceeds to step S102.

The procedure of step S102 corresponds to the first output period Ta in the primary analysis mode. That is, the phoneme corresponding to the pitch G # is output from the speaker of the center channel (C). For this purpose, first, phonemes of phonemes G # with the number of samples N (= 4096) are generated. The generated phonemes are continuously output so as to be looped twice. Thereby, the audio signal as a phoneme corresponding to the pitch G # is reproduced and output with a time length corresponding to twice the number of samples N, that is, a time length corresponding to the output period Ta.
The next step S103 is a procedure for executing the measurement process as the first analysis mode corresponding to the phoneme output in step S102. That is, sampling is performed at the timing when the sample delay time Tdrs has elapsed from the output time of the phoneme in step S102, and a response signal is obtained. Then, FFT is performed on the response signal, the S / N ratio is calculated as shown in FIG. 6, and a required determination result is made or set based on the S / N ratio. That is, the measurement result corresponding to the primary analysis process is performed to obtain the measurement result. For example, since the response signal obtained in step S103 is output from the center channel (C) speaker, it is output from the center channel (C (ch)) speaker in the next secondary analysis mode. The gain of the audio signal is set according to the sound pressure level of the measurement sound to be generated.

Step S104 corresponds to the second output period Ta in the primary analysis mode, and generates two phonemes (number of samples N) respectively corresponding to the pitches F and G # according to step S102. In this case, the signals are output from the front left channel (L) and the front right channel (R) in a loop twice.
In step S105, the phonemes output in step S104 are sampled according to step S103, and the measurement process as the primary analysis mode is executed to obtain the measurement result.

Step S106 corresponds to the third output period Ta in the primary analysis mode, and generates two phonemes (number of samples N) corresponding to the pitches C # and F in accordance with Step S102. In this manner, the output is performed from the left surround channel (Ls) and the right surround channel (Rs) so as to be looped twice.
In step S107, the phonemes output in step S106 are sampled in accordance with step S103, and the measurement process as the primary analysis mode is executed to obtain the measurement result.

The next step S108 corresponds to the fourth (last) output period Ta in the first analysis mode, and in accordance with step S102, two phonemes (number of samples N) corresponding to the pitches C # and G #. ) And are output from the left back surround channel (Bsl) and the right back surround channel (Bsr) so as to be looped twice.
In step S109, according to step S103, the phonemes output in step S105 are sampled, and the measurement process as the primary analysis mode is executed to obtain the measurement result.
By the procedure up to step S109, the measurement result as the primary analysis mode corresponding to each of the seven audio channels is obtained. That is, at this stage, for example, the gain of the audio signal to be output from the speaker of each audio channel is set in the next secondary analysis mode.

The subsequent steps S110 to S117 correspond to the secondary analysis mode here. Step S110, which is the first in the secondary analysis mode, corresponds to the first output period Ta in the secondary analysis mode, and here again, according to step S102, the pitch A # based on the number of samples N A phoneme corresponding to is generated and output in succession twice.
Then, in the next step S111, in accordance with the previous step S103, the phoneme output in step S110 is sampled to obtain a response signal, and the FFT is performed. The measurement process as the next analysis mode is executed. Also in this case, in the measurement process, information on the S / N ratio calculated based on the amplitude value of the target frequency obtained by FFT and the amplitude value of the frequency regarded as background noise is used. As the measurement result, for example, first, it is determined whether or not there is a speaker (which becomes a center channel in the case of step S111) that is supposed to output a phoneme (measurement sound). If it is determined that there is a speaker, the sound pressure level to be output from the center channel speaker during the main measurement, that is, the signal level of the measurement sound is set. In this setting, a determination result such as whether or not a clip has occurred in the audio signal output from the speaker is also used.

The subsequent step S112 corresponds to the second output period Ta in the secondary analysis mode, and two phonemes (number of samples) respectively corresponding to the pitches D # and A # according to the description of the previous step S102. N) is generated and output from the front left channel (L) and the front right channel (R) in a loop twice.
In step S113, the phonemes output in step S113 are sampled in accordance with the previous step S103, and the measurement process as the secondary analysis mode is executed to obtain the measurement result.

Step S114 corresponds to the third output period Ta in the secondary analysis mode, and generates two phonemes (number of samples N) corresponding to the pitches F # and D # according to step S102. Then, each is looped twice to output from the left surround channel (Ls) and the right surround channel (Rs).
In step S115, the phonemes output in step S114 are sampled in accordance with step S103, and the measurement process as the secondary analysis mode is executed to obtain the measurement result.

Step S116 corresponds to the last (fourth) output period Ta of the secondary analysis mode, and two phonemes (number of samples N) corresponding to the pitches G and A # according to step S102. Are generated and output from the left surround channel (Ls) and the right surround channel (Rs) so as to be looped twice.
In step S117, the phonemes output in step S116 are sampled in accordance with step S103, and the measurement process as the secondary analysis mode is executed to obtain the measurement result.
At the stage through the steps so far, the measurement sound output as the second analysis mode, the acquisition of the response signal by sampling, and the analysis by FFT are completed, for example, for each speaker of 7 channels. The presence or absence (that is, the audio channel configuration in the audio system) is determined, and the output level of the measurement sound at the time of the main measurement for each speaker is also set.

  When the measurement sound output sequence of FIG. 7 is followed, the procedure of step S118 corresponding to the non-analysis mode is performed following the secondary analysis mode. That is, a phoneme corresponding to each of the pitches G #, F, and C # is generated, and for the phonemes corresponding to the pitch G #, the center channel (C), the front left channel (L), and the front right channel (R ). The phonemes corresponding to the pitch F # are output from the left surround channel (Ls) and the right surround channel (Rs). The phonemes corresponding to the pitch C # are output from the left back surround channel (Bsl) and the right back surround channel (Bsr). Further, the output of phonemes corresponding to these pitches is performed simultaneously at the timing corresponding to the output period Ta, and as can be seen from FIG. 7, it corresponds to the output period Ta that is continuous four times. Thus, continuous output twice by the number of samples N is repeated four times.

Following the measurement sound output in the non-analysis mode according to the procedure of step S118, as shown as the procedure of step S119, comprehensive determination processing is performed based on the analysis and measurement results so far. For example, the analysis and measurement so far have been performed individually for each phoneme output in the output period Ta. Therefore, for example, an error occurs in the measurement result due to some factor for a certain channel. Even if this is the case, it may be impossible to identify the occurrence of an error based only on the analysis and measurement results of the channel.
Therefore, as step S119, the presence / absence of a local error as described above is determined by comparing and referring to the entire analysis result and measurement result thus far. Alternatively, in consideration of the balance of the parameters set for each channel, the resetting is performed so that these parameters become more optimal.

  FIG. 9 shows a configuration example of the entire system including the sound correction apparatus according to the present embodiment and an audio system connected to the sound correction apparatus. As described above, the acoustic correction apparatus according to the present embodiment is a so-called retrofit kit, and the corresponding model has versatility within a certain range of conditions. In this figure, an example is given in which the audio system that can be connected to the sound correction apparatus 2 of the present embodiment is included in an AV (Audio Video) system that can perform not only audio reproduction but also video reproduction. .

First, the AV system 1 in this case includes a media playback unit 11, a video display device 12, a power amplifier unit 13, and a speaker 14.
The media playback unit 11 plays back, for example, media on which data as video / audio content is recorded, and plays back and outputs a video signal and an audio signal. Here, the media playback unit 11 outputs a digital video signal and audio signal.
In this case, the type, format, and the like of media to be played back by the media playback unit 11 are not particularly limited. For example, a DVD (Digital Versatile Disc) can be considered in the present situation. In the case of supporting a DVD as a specific configuration of the media playback unit 11, data as video / audio content recorded on the loaded DVD is read, and for example, video data and audio data to be played back simultaneously are output. To get. Here, in the current DVD format, the video data and the audio data are in a code format that is compression-encoded according to a method compliant with the DVD standard. Therefore, the compression-encoded video data and audio data are decoded. Processing is performed. Then, the digital video signal and the digital audio signal obtained by the decoding process are output at the timing when the reproduction times are synchronized.

  Note that the media playback unit 11 may have a so-called multimedia configuration that enables playback of, for example, an audio CD in addition to a DVD. Further, it may be configured as a single television tuner that receives and demodulates a television broadcast and outputs a video signal and an audio signal. Alternatively, the configuration may be such that the function of the television tuner and the playback function of the package media are combined.

When the media playback unit 11 supports multi-audio channels, the audio signal played back and output from the media playback unit 11 is output by a plurality of signal lines corresponding to each audio channel. To be.
For example, the media playback unit 11 may have a center channel (C), a front left channel (L), a front right channel (R), a left surround channel (Ls), a right surround channel (Rs) as illustrated in FIG. ), The left back surround channel (Bsl), and the right back surround channel (Bsr) corresponding to 7 channels, audio signals are output by 7 systems corresponding to each of these channels. Is done.

When viewed as a configuration of only the AV system 1, the video signal output from the media playback unit 11 is input to the video display device 12. The audio signal is input to the power amplifier unit 13.
The video display device 12 displays an image based on the input video signal. Here, the display device actually used as the video display device 12 is not particularly limited. For example, in the present situation, a CRT (Cathode Ray Tube), an LCD ((Liquid Crystal Display), a PDP ( Various display devices such as Plasma Display Panel) can be used.

The power amplifier unit 13 amplifies the input audio signal and outputs a drive signal for driving the speaker. The power amplifier unit 13 in this case includes a plurality of power amplifier circuit systems corresponding to the audio channel configuration to which the AV system 1 is compatible, and amplifies an audio signal corresponding to each channel by each of these power amplifier circuits. Then, a drive signal is output to the speaker 14 corresponding to the channel. Accordingly, a plurality of speakers 14 are provided according to the audio channel configuration supported by the AV system 1. For example, when the AV system 1 corresponds to the above-described seven channels, the power amplifier unit 13 includes seven power amplifier circuit systems. Also, as the speaker 14, seven corresponding to each channel are provided, and each is arranged at an appropriate position in the listening environment.
Then, by supplying the drive signal obtained by amplifying the audio signal of each channel by the power amplifier unit 13 to the speaker 14 of the appropriate channel, the sound of the corresponding channel is output from the speaker 14 to the space. As a result, the sound of the content is reproduced and output so as to form a sound field corresponding to the multi-channel configuration. For confirmation, the sound reproduced and output from the speaker in this way is synchronized with the image displayed on the video display device 12 in accordance with the video signal (so-called lip sync). It will be a thing.

  Further, in the AV system 1 itself, for example, the media playback unit 11, the video display device 12, the power amplifier unit 13, and the speaker 14 may be configured as component AV systems that are separated from each other. A unit type device configuration in which at least two of the parts are integrated may be employed.

When the audio correction device 2 according to the present embodiment is retrofitted to such an AV system 1, as shown in the figure, the audio signal output from the media playback unit 11 is processed. To be entered. In this case, for example, as shown in FIG. 7, for example, as the acoustic correction device 2, the center channel (C), the front left channel (L), the front right channel (R), the left surround channel (Ls), Since seven channels of the right surround channel (Rs), the left back surround channel (Bsl), and the right back surround channel (Bsr) can be supported, the maximum of these seven channels can be supported, for example, 7 One audio signal input terminal is provided. In an actual AV system, a subwoofer channel is usually provided in addition to the above seven audio channels, but is omitted here for the sake of simplicity.
Further, for example, when the AV system 1 supports only the L and R stereo channels, the input terminals corresponding to the above seven audio channels for the L and R audio signals output from the media playback unit 11. Of these, connection may be made so as to input to each input terminal corresponding to the front left channel (L) and the front right channel (R).
In the acoustic correction apparatus 2, the audio signal output terminal is also provided so as to be able to output the audio signals of the seven channels at the maximum. The audio signal output of the acoustic correction apparatus 2 is connected to the audio signal input terminal corresponding to each channel in the power amplifier unit 13.

As described above, in the media playback unit 11, for example, when audio information read from the media has been compression-encoded, it is decoded and output as a digital audio signal. This means that the audio signal information to be handled by the acoustic correction apparatus 2 of the present embodiment should be an audio signal in a format after being demodulated for compression coding and the like. As a result, it is not necessary for the acoustic correction apparatus 2 of the present embodiment to include an encoder / decoder for a compression-encoded audio signal.
Further, as the measurement sound to be output from the acoustic correction device 2 to the power amplifier unit 13, it is only necessary to generate a signal according to the format after encoding and decoding. An encoder / decoder process for compression encoding or the like is not required.

In this case, the sound correction device 2 can also input / output a video signal. In this case, the video signal system is connected so that the video signal output from the media playback unit 11 is input and output to the video display device 12.
Further, in the sound correction apparatus 2, the video signal is also processed with respect to the digital video signal in the format after compression coding in the same manner as the audio signal.

The sound measuring apparatus 2 to which the video signal and the audio signal are input in this way is roughly composed of a frame buffer 21, a sound field correction / measurement function unit 22, a control unit 23, and a memory unit 24. .
First, the sound field correction / measurement function unit 22 has two functions. One has a measurement function for measuring the listening environment in order to set parameter values for sound field control necessary for sound field correction. When this measurement function is executed, a measurement sound signal is output to the power amplifier circuit 13 so that the measurement sound is output from an appropriate audio channel as necessary.
In addition, according to the parameter value for sound field control set according to the measurement result by the measurement function, the audio signal for each channel input from the media reproducing unit 11 is subjected to necessary signal processing, and the power amplifier unit 13 To be output. As a result, the sound field formed by the sound of the content output from the speaker is corrected so as to be favorable at an appropriate listening position.

  By the way, the signal processing for the sound field control as described above means that the audio signal input from the media playback unit 11 passes through the DSP in the acoustic correction device 2. In this way, when the audio signal passes through the DSP, a time lag occurs with respect to the reproduction time with the video signal output from the media reproduction unit 11 in the same manner. The frame buffer 2 is provided to eliminate this time lag and achieve so-called lip sync. That is, for example, the control unit 23 controls the video signal input from the media playback unit 11 to be written in the frame buffer 21 for each frame, for example, temporarily held, and then output to the video display device 12. Execute. As a result, the audio correction device 2 outputs a video signal and an audio signal in which the time lag described above is eliminated and the reproduction time is appropriately synchronized.

In addition to the above-described write / read control for the frame buffer 21, the control unit 23 executes control for various functional parts in the sound correction device 2 and various processes.
In this case, the memory unit 24 includes, for example, a nonvolatile memory element, and writing / reading is performed under the control of the control unit 23. In the present embodiment, as essential information stored in the memory unit 24, one example is waveform data of a basic sine wave (see FIG. 1A) for generating a measurement sound as a phoneme. The other is sequence data having a structure as control information for outputting a measurement sound melody with a phoneme pattern of a predetermined phoneme as shown in FIG. 7, for example.
In actuality, for example, various necessary setting information to be referred to by the control unit 23 and other necessary information other than sequence data may be written and stored in the memory unit 24. .

  The microphone 25 is to be attached to the sound correction device 2, and when the sound correction device 2 performs measurement, the microphone 25 is connected to the sound correction device 2 in order to collect the measurement sound output from the speaker 14. To be connected.

FIG. 10 shows an internal configuration example of the sound field correction / measurement function unit 22. As shown in this figure, the sound field correction / measurement function unit 22 includes a microphone amplifier 101, a main measurement processing block 103, a preparation measurement processing block 106, and a sound field correction processing block 110. Here, the sound field correction processing block 110 performs processing for sound field correction, whereas the parts on the microphone amplifier 101, the main measurement processing block 103, and the preparation measurement processing block 106 side are parts for executing the measurement process. It is. Based on the result of this measurement process, the values of various required parameters for the sound field correction process in the sound field correction process block 110 are changed and set.
In addition, switches 102 and 109 are provided to switch the measurement mode between the main measurement and the preparation measurement. In addition, a switch 120 is provided to switch between the measurement mode and the sound field correction mode. These switches 102, 109, and 120 are switched so that the terminal Tm2 or Tm3 is alternatively connected to the terminal Tm1. The controller 23 controls the switching operation.

As an explanation of the sound field correction / measurement function unit 22 shown in FIG. 10, first, an operation corresponding to the preparation measurement mode will be described.
In the preparation measurement mode, first, the control unit 23 connects the terminal Tm2 to the terminal Tm1 of the switch 120. For the switches 102 and 109, the terminal Tm3 is connected to the terminal Tm1. Thereby, a signal path in the sound field correction / measurement function unit 22 corresponding to the preparation measurement mode is formed as the measurement mode.

The preparation measurement processing block 106 includes an analysis processing unit 107 and a measurement sound processing unit 108 as illustrated. In the measurement sound processing unit 108, for example, as shown in FIG. 11, waveform data of a basic sine wave is input, a phoneme corresponding to a predetermined pitch is generated, and this is used as a measurement sound for preparation measurement. This is a part for outputting in the signal format.
The phoneme generation process by the measurement sound processing unit 108 follows, for example, the phoneme formation method shown in FIG. Further, as can be understood from the description from FIG. 7, in the present embodiment, the measurement sound can be output corresponding to each of multiple audio channels. Therefore, in FIG. 10, for the sake of simplicity of illustration, only one signal output line from the measurement sound processing unit 108 is shown, but in reality, as shown in FIG. It is assumed that there is a signal output line for the measurement sound corresponding to each channel.
Further, the measurement sound processing unit 108 generates a frequency corresponding to which pitch as a phoneme, and outputs the generated phoneme from a signal line corresponding to which channel in the control content described in the sequence data. Followed.
The waveform data of the basic sine wave is read from the memory unit 24 under the control of the control unit 23 at a required timing, and is input to the measurement sound processing unit 108. In addition, the sequence data is not directly input to the measurement sound processing unit 108. First, the control unit 23 reads the sequence data from the memory unit 24, interprets it, and sends it to the measurement sound processing unit 108. On the other hand, the pitch (frequency) corresponding to the phoneme to be generated and the audio channel to be output are indicated.

Further, the process for forming one phoneme in the measurement sound processing unit 108 can be shown by the block configuration shown in FIG.
As the measurement sound processing unit 108, first, waveform data of a basic sine wave is input, and the m-th order sine wave processing 201 is a base tone of a phoneme having a frequency corresponding to a designated pitch, and a required order m. To generate an mth-order sine wave. The frequency of the m-order sine wave generated in this way is expressed by, for example, (Equation 2). In addition, what value is selected as the order m, that is, what frequency is set as the base sound depends on the control of the control unit 23 based on the contents of the sequence data.

  Here, as the waveform data (basic waveform component data) of the basic sine wave used by the m-order sine wave processing 201 for generating the m-order sine wave, a waveform corresponding to one cycle as shown in FIG. Although it may be data, at a minimum, it can be said that there is only a quarter period of waveform data. That is, since the waveform data is a sine wave, it is possible to easily form a complete sine waveform for one cycle by a simple calculation if there is at least a quarter cycle. Further, by setting the waveform data for the minimum quarter period, the basic waveform component data has a smaller data amount, and the capacity of the memory unit 24 can be saved.

The m-order sine wave generated by the process of the m-order sine wave generation process 201 becomes the base tone of the phoneme with the octave order k = 1 as understood from the above description. The m-th order sine wave waveform data generated by the m-th order sine wave generation processing 201 is transferred so as to be branched to the level adjustment processing 203-1 and the octave harmonic generation processing 202. It is.
In the octave harmonic generation process 202, the multiplication by a predetermined multiple (2 times, 4 times, 8 times, 16 times, 32 times) is performed based on the m order sine wave as the base sound taken from the m order sine wave generation processing 201. By executing the process, in this case, five octave harmonics with octave orders k = 2, k = 3, k = 4, k = 5, and k = 6 are generated. As the multiplication process, for example, the concept shown in FIG. 1 may be applied. That is, based on the m-th order sine wave as the base sound, the sine wave of the base sound is subjected to thinning sampling according to the order of the octave harmonic as shown in FIGS. To be.
The octave harmonics with these octave orders k = 2, k = 3, k = 4, k = 5, and k = 6 are level adjustment processing 203-2, 203-3, 203-4, 203-5, respectively. 203-6.
In this way, in the six level adjustment processes 203-1 to 203-6, the bass sound (k = 1) and the octave harmonics corresponding to the octave orders k = 2 to 6 are obtained as m-order sine waves, respectively. Entered.
In these level adjustment processes 203-1 to 203-6, processing is executed so that required amplitude values are set for the bass sound and octave harmonics that are input and captured. The amplitude value set by the level adjustment processing 203-1 to 203-6 may be a fixed value determined in advance, or may be varied according to the control of the control unit 23. Good.

The bass sound and the octave harmonics whose levels have been adjusted by the level adjustment processes 203-1 to 203-6 are synthesized by the synthesis process 204 and output as one phoneme (speech signal waveform). The phoneme obtained by the synthesis process 204 has a tone color according to the amplitude balance of the base sound and the octave harmonics according to the level adjustment results of the level adjustment processes 203-1 to 203-6. Become.
The phonemes generated according to the process shown in FIG. 12 correspond to the number of samples N, for example. Therefore, for example, in order to output phonemes in the output period Ta as shown in FIG. 7, the measurement sound processing unit 108 outputs the phonemes generated according to the process of FIG. 12 twice in succession. The
Further, the measurement sound processing unit 108 can simultaneously generate phonemes corresponding to different pitches by executing the processes shown in FIG. 12 in parallel, for example. Also, a sound signal as a phoneme generated by the process shown in FIG. 12 can be output as a measurement sound signal from an output line corresponding to one or more appropriate audio channels.

In FIG. 10, a measurement sound signal composed of phonemes output from the measurement sound processing unit 108 of the preparation measurement processing block 106 is a power amplifier via a switch 109 (terminal Tm3 → Tm1) via a switch 120 (Tm2 → Tm1). This is input to the unit 13. In the power amplifier unit 13 shown in FIG. 9, the input sound signal of the measurement sound is amplified and output from the speaker 14.
As can be understood from the above description, when the measurement sound processing unit 108 outputs the sound signals of the measurement sound (phonemes) simultaneously through a plurality of channels, the power amplifier unit 13 outputs these channels. Each is amplified and output from the speaker 14 of the corresponding channel.
As a result, the measurement sound is output as real sound from the speaker 14 to the surrounding space.

In the main measurement and the preparatory measurement, as shown in FIG. 9, the microphone 25 for collecting the measurement sound as a target is connected to the sound correction device 2. The audio signal from the microphone 25 is input to the microphone amplifier 101 in the sound field correction / measurement function unit 22 as shown in FIG.
The microphone 25 is installed so that sound is picked up at a listening position (listening position) where it is desired to obtain the best corrected sound field in the listening environment. For example, if the system shown in FIG. 9 is an in-vehicle device and the user wants to obtain an appropriate sound field when listening to the driver's seat, the user sits in this driver's seat. In this state, the microphone 25 is installed so that sound is collected at a position where the ear is almost present.

Assuming that the measurement sound is output from the speaker 14 in response to the measurement sound signal being output from the measurement sound processing unit 108 under the preparation measurement mode as described above, the microphone 25 Thus, ambient sound including the measurement sound is collected. The sound signal of the collected sound is amplified by the microphone amplifier 101 and input to the analysis processing unit 107 of the preparation measurement processing block 106 via the terminals Tm1 → Tm3 of the switch 102.
The analysis processing unit 107 samples the input audio signal, for example, at the timing described above with reference to FIG. 5 to obtain a response signal, and performs frequency analysis by FFT, for example. The frequency analysis result is captured by the control unit 23, for example, so that a required measurement result based on the frequency analysis result is obtained as described with reference to FIG.

  In this measurement mode, the control unit 23 sets the switch 120 to the measurement mode by maintaining the connection state between the terminal Tm1 and the terminal Tm2, and the switches 102 and 109 are both connected to the terminal Tm1. Switch to Tm2 connection. As a result, the sound field correction / measurement function unit 22 forms a signal path corresponding to the main measurement mode as the measurement mode.

In the measurement in the main measurement mode, the main measurement processing block 103 functions instead of the preparation measurement processing block 106. The main measurement processing block 103 also includes an analysis processing unit 104 and a measurement sound processing unit 105. At the time of actual measurement, the measurement sound processing unit 105 generates a required signal waveform and outputs it as measurement sound. In this measurement, measurement sound other than the measurement sound by the phoneme used in the preparation measurement is also used.
At this time, the level of the measurement sound output from the speaker of each channel is set according to the measurement result of the previous preparation measurement. Furthermore, since the presence / absence of the speaker and (channel configuration) are also determined by the previous preparation measurement, the measurement sound is not output to the channel corresponding to the speaker that is not present in the AV system. Thereby, the processing burden as the measurement sound processing unit 105 is efficiently reduced. The measurement sound level setting and the measurement sound output setting according to the channel configuration are performed by the control unit 23 controlling the measurement sound processing unit 105 according to the preparation measurement result.

  In this way, when the measurement sound signal is output from the measurement sound processing unit 105 of the main measurement processing block 103, the ambient sound including the measurement sound is generated by the microphone 25 in the same manner as in the preparation measurement. The collected sound is input from the microphone amplifier 101 to the analysis processing unit 104 via the terminals Tm1 → Tm2 of the switch 102.

The analysis processing unit 104 also samples the input sound signal at a required timing according to the measurement sound output to obtain a response signal, and performs frequency analysis by FFT, for example. Then, for example, the control unit 23 takes in the frequency analysis result to obtain a required measurement result as the main measurement. That is, for example, a set value of a required parameter for sound field correction is determined.
Here, the analysis processing unit 104 of the main measurement processing block 103 and the analysis processing unit 107 of the preparation measurement processing block 106 have a common function in that, for example, frequency analysis is performed by FFT. Further, the main measurement process and the preparation process are not executed concurrently. Therefore, the analysis processing units 104 and 107 may be combined into one and shared between the main measurement process and the preparation process.

  Subsequently, in order to enter the sound field correction mode, the switch 120 is connected to the terminal Tm3 with respect to the terminal Tm1. Note that the switches 102 and 109 are for switching between the main measurement mode and the preparation measurement mode in the measurement mode, and at this time, the terminal switching state may be indefinite.

  In the sound field correction mode, the source sound signal is input to the sound field correction processing block 110. The source audio signal here is an audio signal reproduced and output by the media reproducing unit 11, and as described above, a plurality of audio signals of a maximum of 7 channels are input. There is a case. The sound field correction processing block 110 in this case is provided with a delay processing unit 111, an equalizer unit 112, and a gain adjustment unit 113, but each of these parts is also independent for each of up to seven channels of audio signals. It is configured to be capable of processing.

In the sound field correction processing block 110, the delay processing unit 111 is configured to be able to output the audio signals of the input channels after being delayed by different delay times. The delay processing unit 111 corrects the disturbance of the sound field caused by the time difference of the arrival sound from the speaker to the listening position according to the difference in distance from the listening position from each speaker.
Further, the equalizer unit 112 can set and output arbitrary equalizer characteristics independently for each input audio signal of each channel. Depending on the equalizer unit 112, the sound quality that varies depending on the relationship between the position of the speaker and the listening position, the state of the obstacle between the speaker and the listening position, and the variation in the reproduction acoustic characteristics of the speaker is corrected.
The gain adjusting unit 113 can set and output the gain independently for each input audio signal of each channel. Depending on the gain adjusting unit 113, the positional relationship between the speaker and the listening position, the state of an obstacle existing between the speaker and the listening position, the volume that varies from channel to channel is corrected depending on the distance between the speaker and the listening position. To do.
The sound field correction processing block 110 having such a signal processing function is configured as a DSP corresponding to an audio signal, for example.

As a measurement result of the main measurement described above, the control unit 23 determines the relationship between the time difference of the arrival sound to the listening position between the audio channels, the sound quality change when the sound of each audio channel reaches the listening position, and the level It is assumed that information such as a variation state is obtained.
Then, as the sound field correction parameter, for example, based on the information on the relationship of the time difference of the arrival sound to the listening position between the audio channels, each delay processing unit 111 is informed so that this time difference is eliminated. Set the delay time for each audio channel.
Further, based on the information on the sound quality change at the stage when the sound of each audio channel reaches the listening position, the equalizer characteristic for each audio channel is set to the equalizer unit 112 so as to compensate for this sound quality change. . Also, based on the information on the variation in the sound level of each audio channel that has reached the listening position, the gain is set for each audio channel in the gain adjustment unit 113 so that this variation is eliminated.

  The source audio signal input to the sound field correction processing block 110 is subjected to signal processing by the delay processing unit 111, the equalizer unit 112, and the gain adjustment unit 113 set as described above, and then the power amplifier unit. 13 is amplified and output from the speaker 4 as real sound. The sound field formed by the sound output in this way is improved and better than before correction, for example, by listening at an appropriate listening position.

Here, the structure of the sequence data is illustrated in FIG. Note that the data structure shown in this figure is merely an example.
The sequence data shown in this figure has a structure formed by connecting event units. One event is data corresponding to one phoneme. Each event stores, for example, information on sound generation time, bass sound, harmonic structure, channel, and analysis mode.
The pronunciation time information defines the output timing for the phoneme to which the current event corresponds. With this, how many times the sample number N is continuously output for the phoneme, and the phoneme, It is specified at what timing in time to output. As the output timing, for example, it is conceivable that the start point of the phoneme output as the measurement sound melody is defined as a base point (0) and designated by the integration of the number of samples with respect to the base point. As the resolution of the phoneme output timing in this case, the time corresponding to one cycle of the sampling frequency is the highest.
As the bass sound information, designation as to which m-th order sine wave is to be used as the base sound is performed.

The harmonic structure information designates the balance with respect to the bass sound for each amplitude value of the octave harmonics with octave orders k = 2-6. Thereby, the tone color of a phoneme is determined. Note that the balance of the amplitude values of the octave harmonics may take into account that, for example, a good measurement result suitable for the measurement conditions can be obtained instead of considering only the timbre of the phoneme.
The harmonic structure is generated in accordance with the information on the harmonic structure in the first analysis mode, but in the second analysis mode, for example, at this stage according to the measurement result in the first analysis mode. In order to obtain a better measurement result, it may be changed adaptively.

  The channel information specifies an audio channel on which the phoneme should be output. For example, considering that there is a case where phonemes corresponding to the same pitch are output from a plurality of channels at the same time, it is possible to define the information of this channel as one that can be described by specifying a plurality of channels. preferable. In this way, it is possible to perform control for simultaneously outputting phonemes corresponding to the same pitch from a plurality of channels by one event without creating a plurality of events corresponding to the number of channels.

The analysis mode information specifies the analysis mode to which the phoneme corresponds. For example, according to the example shown in FIGS. 7 and 8, it is indicated which mode is the primary analysis mode, the secondary analysis mode, or the non-analysis mode. The control unit 23 determines whether or not to analyze the voice obtained by outputting this phoneme according to the mode indicated by the information of the analysis mode. Accordingly, for example, a measurement result corresponding to either the first order or the second order is obtained. It is also conceivable to include information designating the sample delay time Tdrs in the analysis mode information.
Based on such sequence data, the control unit 23 executes control on the preparatory measurement processing block, so that phonemes are output according to the pitch and output timing according to the description content of the sequence data. As described with reference to FIG. 7, the measurement sound is output melodyally.

FIG. 14 is a flowchart showing a control process for preparatory measurement that is executed by the control unit 23.
First, the control unit 23 reads required sequence data from the memory unit 21 in step S201. Thereafter, the control unit 23 can analyze and process the content of the read sequence data.

  In the next step S202, background noise is checked. This is a process for realizing the same operation as step S101 shown in FIG. Steps S203 and after are processing when it is determined that the microphone 25 is connected as a background noise check result.

Step S203 and subsequent steps are processing for processing events according to the interpretation of the sequence data.
First, in step S203, it is determined whether or not there is a phoneme that has reached the output start timing among phonemes that have not yet started output by referring to information on the sound generation time of an event that has not yet been processed. If a negative determination result is obtained on the assumption that there is no phoneme reaching the output start timing, the process of step S204 is skipped and the process proceeds to step S205, but it is affirmed that there is a phoneme reaching the output start timing. If the determination result is obtained, the process of step S204 is executed.

  In step S204, for the phoneme determined to be output in step S203, information on the base sound and the harmonic structure described in the event corresponding to the phoneme is used to actually generate the phoneme. Execute the process. Then, the generated phoneme is output by the number of repetitions of the number of samples N based on the sound generation time information described in the event corresponding to the phoneme. Also, a channel for outputting phonemes as an audio signal is determined according to channel information described in the same event.

  Each time a phoneme is started to be output by the processing of step S204 so far, a sampling processing event occurs at a timing corresponding to the sample delay time Tdrs. In step S205, it is determined whether or not there is any sampling processing event to be generated in this way that has reached the start timing. If it is determined that there is no sample processing event that has reached the start timing, the process from step S206 to S207 is skipped and the process proceeds to step S208. On the other hand, if a positive determination result is obtained in step S205 that there is a sampling process event that has reached the start timing, the process proceeds to step S206.

  In step S206, a sampling process is executed for the audio signal collected by the microphone 25 at a timing according to the sample delay time Tdrs with a predetermined sampling number. In the embodiment described so far, the sampling process with the number N of samples is executed. In the next step S207, the response signal obtained by sampling in step S206 is subjected to frequency analysis by FFT according to the analysis mode specified by the event corresponding to the phoneme, and the analysis result is used. Then, a process for obtaining a measurement result corresponding to the designated analysis mode is executed.

In step S108, it is determined whether or not the sequence is completed, that is, whether or not the event processing for the sequence data read in step S201 and the sampling processing and analysis processing corresponding thereto have been completed. If a negative determination result is obtained, the process returns to step S203. If a positive determination result is obtained, the process proceeds to step S209.
In step S209, a comprehensive determination process similar to the procedure in step S119 of FIG. 8 is executed.

By the way, in this embodiment, what kind of melody is used as the measurement sound melody is determined by the sequence data. In the simplest form, only one sequence data is prepared in advance so as to be stored in the memory unit 24, and when the measurement sound melody is output, the sequence data may be performed based on the sequence data. ,It turns out that. Further, as a plurality of sequence data stored in the memory unit 24 and prepared, it is conceivable to select and use sequence data according to a selection operation by a user or a predetermined condition in preparation measurement.
Moreover, as sequence data, for example, not only preset data stored in the memory unit 24 at the time of factory shipment, but also acquired from the outside, for example, at a stage after the sound correction device 2 reaches the user's hand. It is also conceivable to store (download) the data in the memory unit 24.
In addition, in the sequence data, the measurement sound output sequence corresponding to the non-analysis mode can be edited so as to arbitrarily change the melody, phoneme tone, and the speaker that outputs the phoneme according to the user's operation. It is also possible to do so. In this way, entertainment properties are further enhanced. However, if the phoneme output corresponding to the analysis mode is inadvertently changed, significant measurement may not be possible, so the output sequence of the measurement sound corresponding to the analysis mode cannot be edited by the user. Is preferred.

  Further, in the above embodiment, the basic waveform component data is held, and all necessary phonemes are generated based on the basic waveform component data. In this case, since the source for obtaining the necessary phoneme is only one basic waveform component data, there is an advantage that the storage capacity of the storage area in the acoustic correction device such as the memory unit 24 is not compressed. However, if there is room for such a storage area, waveform data of all phonemes required to create the measurement sound melody are created and stored in advance as sound source data, and the measurement sound melody At the time of output, a configuration may be considered in which the sound source data is read from the storage area and reproduced and output.

Further, according to the concept shown in FIGS. 2 and 4, only phonemes that can form a scale are adopted as elements of the measurement sound melody. However, phonemes that do not fall within the scale can be measured frequencies as long as they are based on mth order sine waves that fit in integer cycles with respect to the number of samples N. There is no. Rather, for example, by using a phoneme that does not fit in such a scale in the measurement sound melody, it is possible to make the measurement sound melody more musically effective, so it is actively used. It is good.
In consideration of not performing frequency analysis on the response signal in the non-analysis mode, in the non-analysis mode, a measurement sound based on an mth-order sine wave that fits in an integer period with respect to the number of samples N is output. It can be said that there is no need. Therefore, in response to the non-analysis mode, if a waveform other than the m-th order sine wave is used, a melody with more diverse tones can be obtained as a series of measurement sound output sequences. Musicality and entertainment will be further enhanced. For example, if a sampled sound of a real musical instrument is used as a waveform other than that based on the m-th order sine wave, the measured sound melody becomes more musical.

  In the embodiment, as the microphone 25 that picks up the measurement sound, for example, if one non-directional monaural one is used, a sufficiently significant measurement can be performed. For example, a plurality of microphones can be used. Can be obtained at an appropriate position, or a stereo-compatible microphone or a plurality of binaural microphones can be used to obtain a more reliable measurement result.

Further, for example, in the acoustic correction device 2 of the present embodiment shown in FIG. 10, the operations as the measurement sound processing unit 108 and the analysis processing unit 107, which are assumed to be responsible for the preparation measurement processing block 106, that is, phoneme generation processing The control processing for constructing the measurement sound melody (outputting the generated phoneme at the timing corresponding to the sequence data), sampling the collected sound signal at the required timing, and performing the FFT, etc. You may comprise by. Moreover, what is necessary is just to be implement | achieved as a process which CPU performs according to a program as the acoustic correction apparatus 2 having a microcomputer. When this configuration corresponds to FIG. 10, the control unit 23 is configured as a microcomputer, and the functional unit lock as the preparation measurement processing block 106 is actually executed by the CPU in the control unit 23. Software processing.
Further, the functions as the main measurement processing block 103 and the sound field correction processing block 110 may be configured as hardware or may be configured by software.

In the description of the embodiments so far, the measurement sound based on the mth-order sine wave has been described as being used in the preparation measurement for sound field correction. Depending on the measurement conditions, this measurement can be used without any problem. In addition, the purpose of the measurement is not limited to sound field correction as long as sound in a human audible frequency band is appropriate as a measurement target.
In the above embodiment, FFT is used to perform frequency analysis on the response signal of the measurement sound based on the mth-order sine wave. For example, DFT (Discrete Fourier Transform) is used. It is also possible to adopt other frequency analysis.

It is explanatory drawing which shows the basic concept about the phoneme used as the element of a measurement sound in embodiment of this invention. It is explanatory drawing which shows the basic concept about the phoneme formation method and selection of the phoneme suitable for a measurement sound melody. It is a figure which shows the frequency characteristic of the phoneme selected based on the concept shown in FIG. It is explanatory drawing which shows the concept about the selection of the phoneme which is actually employ | adopted in this Embodiment, and the selection of the phoneme suitable for a measurement sound melody. It is a timing chart which shows the basic sequence about measurement sound (phoneme) output and sampling in an embodiment. It is a figure which shows the example of a frequency analysis result about the response signal in embodiment. It is a figure which shows the actual example of the output pattern of the measurement sound melody in embodiment. It is a flowchart which shows the example of a procedure of a phoneme production | generation and output process according to the output pattern of the measurement sound melody shown in FIG. 7, and an analysis and a measurement process. 1 is a block diagram illustrating a configuration example of an entire system including an acoustic correction system according to an embodiment and an AV system. It is a block diagram which shows the structural example of the acoustic correction system of embodiment. It is a block diagram which shows the example of an actual signal output form about the measurement sound process part in a preparatory measurement process block. It is a block diagram which shows the phoneme production | generation process corresponding to one phoneme in the measurement sound process part in a preparation measurement process block. It is a figure which shows the structural example of sequence data. It is a block diagram which shows the processing operation | movement which a control part (microcomputer) performs for a preparatory measurement.

Explanation of symbols

  1 AV system, 2 sound correction device, 11 media playback unit, 12 video display device, 13 power amplifier unit, 14 speaker, 21 frame buffer, 22 sound field correction / measurement function unit, 23 control unit, 24 memory unit, 25 microphone , 101 Microphone amplifier, 102, 120 switch, 103 Main measurement processing block, 104, 107 Analysis processing unit, 105, 108 Measurement sound processing unit, 106 Preparation measurement processing block, 110 Sound field correction processing block, 110 Sound field correction processing block , 111 delay processing unit, 112 equalizer unit, 113 gain adjustment unit, 201 mth-order sine wave generation processing, 202 octave harmonic generation processing, 203-1 to 203-6 level adjustment processing, 204 synthesis processing

Claims (9)

A measurement method comprising a first measurement procedure and a second measurement procedure following the first measurement procedure,
The first measurement procedure is as follows.
A first output procedure for outputting a plurality of required phonemes obtained based on different fundamental sound components to separate speakers so that the output periods overlap each other;
A first sound collection procedure for obtaining a sound signal by collecting the plurality of phonemes emitted from the separate speakers through a plurality of spatial transmission paths;
Based on the analysis result obtained by executing a predetermined frequency analysis process on the audio signal collected in the first sound collection procedure, the signal output for each of the separate speakers is in accordance with the sound pressure level. Execute the setting procedure and to set
The second measurement procedure is as follows.
A plurality of required phonemes obtained based on different fundamental sound components are subjected to characteristics according to the sound pressure level set in the setting procedure, and the obtained plurality of phoneme signals are output to the separate speakers. 2 output procedures;
A second sound collection procedure for obtaining a sound signal by collecting the plurality of phoneme signals emitted from the separate speakers via the plurality of spatial transmission paths;
Based on an analysis result obtained by executing a predetermined frequency analysis process on the audio signal collected in the second sound collection procedure, a measurement result on a required measurement item is obtained for each of the plurality of spatial transmission paths. To obtain the measurement procedure and
The phoneme is obtained on the basis of a fundamental component that is a sine wave in which an integer number of periods is applied to a predetermined number of samples N expressed as a power of 2,
One phoneme output in the first output procedure and the second output procedure is 1 / ( 2 ^{P of the} fundamental sound component to which a predetermined number of cycles of the integer is applied. ) When a frequency component having a frequency of (P is a natural number) is set as a virtual fundamental component, an arbitrary number of harmonic components that have a frequency on a predetermined octave number with respect to the virtual fundamental component are selected. A measurement method that is output as a signal formed by combining harmonic components.   One phoneme output in at least one of the first output procedure and the second output procedure is a sine wave whose period number is 1 with respect to a predetermined sample number N expressed by a power of 2 The measurement method according to claim 1, wherein the measurement is output using basic waveform component data for a predetermined period that is equal to or more than ¼ period.   In at least one of the first sound collection procedure and the second sound collection procedure, the audio signal emitted from the speaker and collected is sampled at a predetermined timing with the number of samples N as the minimum sample unit. The measurement method according to claim 1, wherein a sampling procedure is executed.   The measurement method according to claim 1, wherein at least one of the first output procedure and the second output procedure outputs the next required phoneme at a required timing after outputting the required phoneme.   The specified phoneme is output at a specified output start timing based on control information specifying a phoneme output pattern in at least one of the first output procedure and the second output procedure. The measuring method as described in.   In at least one of the first output procedure and the second output procedure, among the phonemes, a phoneme having a specific frequency component set as one reference frequency, and this reference frequency is converted into a predetermined scale. The measurement method according to claim 1, wherein a phoneme having a specific frequency component having a frequency that can be another pitch in the scale is output when one pitch is formed. Prior to the first measurement procedure, a determination procedure for determining the presence or absence of background noise is provided,
The measurement method according to claim 1, wherein the first measurement procedure is executed when it is determined that background noise exists as a result of the background noise check in the determination procedure. First output means for outputting a plurality of required phonemes obtained based on different fundamental sound components to separate speakers so that the output periods overlap each other;
First sound collecting means for collecting the plurality of phonemes emitted from the separate speakers through a plurality of spatial transmission paths to obtain an audio signal;
Based on the analysis result obtained by executing a predetermined frequency analysis process on the sound signal collected by the first sound collecting means, the sound output level corresponding to the signal output for each of the separate speakers is determined. Setting means for setting;
A plurality of required phonemes obtained based on different fundamental sound components are subjected to characteristics according to the sound pressure level set by the setting means, and the obtained plurality of phoneme signals are output to the separate speakers. Two output means;
A second sound collecting means for collecting the plurality of phoneme signals emitted from the separate speakers through the plurality of spatial transmission paths to obtain a sound signal;
Based on an analysis result obtained by executing a predetermined frequency analysis process on the audio signal collected in the second sound collection procedure, a measurement result on a required measurement item is obtained for each of the plurality of spatial transmission paths. Measuring means to obtain,
The phoneme is obtained on the basis of a fundamental component that is a sine wave in which an integer number of periods is applied to a predetermined number of samples N expressed as a power of 2,
One phoneme output by the first output means and the second output means is 1 / ( 2 ^{P of the} fundamental sound component to which a predetermined number of periods is applied. ) When a frequency component having a frequency of (P is a natural number) is set as a virtual fundamental component, an arbitrary number of harmonic components that have a frequency on a predetermined octave number with respect to the virtual fundamental component are selected. A measuring device that is output as a signal formed by combining harmonic components. A program for causing a computer to execute a measurement procedure comprising a first measurement procedure and a second measurement procedure following the first measurement procedure,
The first measurement procedure is as follows.
A first output procedure for outputting a plurality of required phonemes obtained based on different fundamental sound components to separate speakers so that the output periods overlap each other;
A first sound collection procedure for obtaining a sound signal by collecting the plurality of phonemes emitted from the separate speakers through a plurality of spatial transmission paths;
Based on the analysis result obtained by executing a predetermined frequency analysis process on the audio signal collected in the first sound collection procedure, the signal output for each of the separate speakers is in accordance with the sound pressure level. Execute the setting procedure to set
The second measurement procedure is as follows.
A plurality of required phonemes obtained based on different fundamental sound components are subjected to characteristics according to the sound pressure level set in the setting procedure, and the obtained plurality of phoneme signals are output to the separate speakers. 2 output procedures;
A second sound collection procedure for obtaining a sound signal by collecting the plurality of phoneme signals emitted from the separate speakers via the plurality of spatial transmission paths;
Based on an analysis result obtained by executing a predetermined frequency analysis process on the audio signal collected in the second sound collection procedure, a measurement result on a required measurement item is obtained for each of the plurality of spatial transmission paths. To obtain the measurement procedure and
The phoneme is obtained on the basis of a fundamental component that is a sine wave in which an integer number of periods is applied to a predetermined number of samples N expressed as a power of 2,
One phoneme output in the first output procedure and the second output procedure is 1 / ( 2 ^{P of the} fundamental sound component to which a predetermined number of cycles of the integer is applied. ) When a frequency component having a frequency (P is a natural number) is used as a virtual fundamental component, an arbitrary number of harmonic components that have a frequency on a predetermined octave number with respect to the virtual fundamental component are selected. A program that causes a computer to execute a procedure that is output as a signal formed by combining harmonic components.

Images