JP5879199B2 - Reverberation response generation apparatus and program thereof - Google Patents

Description

  The present invention relates to a reverberation response generation apparatus that generates a reverberation response (room impulse response for each direction) and a program thereof, which are used when generating reverberation that simulates the sound of a predetermined acoustic space when listening to headphones.

  Binaural technology that perceives sound as coming from the direction by convolving the head related impulse response (HRIR) in the direction in which the signal arrives when listening to headphones. (For example, refer to Patent Document 1). The above-mentioned “head impulse response” indicates acoustic transmission characteristics from a sound source position to an ear position (a binaural tympanic membrane position or a binaural ear canal entrance) in a free sound field (anechoic space). This transfer function representation of the head impulse response is a head related transfer function (HRTF). The above-mentioned “impulse response” indicates, for example, a response at a sound receiving point when a pulse sound is emitted from the sound source position.

  The head impulse response is usually measured in a non-reflective environment such as an acoustic anechoic chamber. Since the head impulse response is determined by reflection, diffraction, etc. of sound waves in the vicinity of the body, head, and ears, it varies depending on the individual. Therefore, in order to obtain an optimal audibility in binaural technology, it is desirable to measure the head impulse response with the listener's own ear.

  On the other hand, such a head impulse response is not measured, and a microphone is installed at a binaural eardrum position or a binaural ear canal entrance of a mannequin called HATS (Head and Torso Simulator) The method of picking up and presenting sound signals directly at the positions of both ears is called binaural recording / playback.

  Here, in the binaural technique in which the head impulse response is convoluted with the acoustic signal as described above, there are roughly two methods for adding the reverberation of a certain acoustic space to the acoustic signal. The first method is a method in which a dummy head or HATS is arranged in an acoustic space and a head impulse response including reflection is measured. By convolving a head impulse response including such a reflection component with an acoustic signal, the listener can perceive the arrival of sound including reverberation. Here, in the following description, the head impulse response including the reflection component in the acoustic space is referred to as a binaural room impulse response (BRIR).

  In the second method, reverberation created using a reverberator (reverberation device) or reverberation created using a room impulse response (RIR) measured in an acoustic space, It is a method of adding separately from the direct sound produced | generated by the convolution of an impulse response and an acoustic signal. The above-mentioned “room impulse response (indoor impulse response)” is an acoustic transfer characteristic from the sound source position to the sound receiving point in the acoustic space, and more specifically, the shape of the acoustic space, the wall It shows the sound reflection component based on the sound absorption rate and the position of the sound source.

JP 2002-209300 A

  However, in the first method described above, the head impulse response and the reflection in the acoustic space are measured at the same time. Therefore, in order to obtain the optimum audibility, the listener himself / herself goes into the acoustic space, and the binaural room impulse is obtained. The response had to be measured. For this reason, for example, when the listener changes, it is necessary for the listener to go to the acoustic space and remeasure the binaural room impulse response each time, and there is a problem that the processing is complicated.

  The second method described above is based on the premise that reverberation added by the method is originally reproduced on a speaker. For example, when the reverberation is directly presented to both ears through headphones, the reverberation is not necessarily the original location (acoustic space). ) Has a problem that the desired reverberation to reproduce the reverberation is not always obtained.

  The present invention has been made in view of such a point, and generates a reverberation response capable of generating reverberation close to that heard in the same acoustic space even when the listener is switched when listening to headphones. It is an object of the present invention to provide a reverberation response generation device and a program thereof.

  In order to solve the above-mentioned problem, a reverberation response generation device according to claim 1 is a device that generates a reverberation response for generating reverberation in an acoustic space, and includes an impulse response convolution unit, a first impulse response conversion unit, The second impulse response conversion means, the spectrogram calculation means, the spectrogram conversion means, and the impulse response deconvolution means are provided.

  The reverberation response generating apparatus having such a configuration exhibits acoustic transfer characteristics for each of a plurality of directions from the sound source position in the acoustic space to an arbitrary sound receiving point with reference to the sound receiving point by the impulse response convolution means. A room direction impulse response according to the first direction, and a head impulse response indicating acoustic transfer characteristics in a plurality of directions based on the position of the ear from the sound source position in the anechoic space to the position of the head ear, The room impulse response for each first binaural direction in each direction is calculated by convolution for each matching direction. The reverberation response generating device calculates a spectrogram of the first binaural direction room impulse response by performing a short-time Fourier transform on the first binaural direction room impulse response by the first impulse response conversion means. In addition, the reverberation response generating device generates a binaural room impulse response indicating a sound transfer characteristic from the sound source position in the acoustic space to the position of the head ear located at the same position as the sound receiving point by the second impulse response conversion means. A spectrogram of the binaural room impulse response is calculated by performing a short-time Fourier transform.

  In addition, the reverberation response generating apparatus may use the spectrogram calculating unit to calculate the ratio of the amplitude of the spectrogram of the first binaural direction-specific room impulse response to the total sum of the spectrogram amplitudes of the first binaural direction-specific room impulse responses of all directions. Is multiplied by the amplitude of the spectrogram of the binaural room impulse response, and the spectrogram of the room impulse response of the second binaural direction in each direction is calculated by replicating the phase of the spectrogram of the binaural room impulse response. The reverberation response generating apparatus calculates the second binaural direction-specific room impulse response by performing inverse short-time Fourier transform on the spectrogram of the second binaural direction-specific room impulse response by the spectrogram conversion unit. Then, the reverberation response generation device deconvolves the second binaural direction-specific room impulse response and the head impulse response for each matching direction, so that the second direction-specific room impulse response in each direction is used as the reverberation response. calculate.

  In this way, the reverberation response generating device uses the spectrogram calculating means to calculate the ratio of the amplitude response of the first binaural direction-specific room impulse response to the total sum of the amplitude responses of the first binaural direction-specific room impulse responses in all directions. Binaural room impulses recorded by binaural recording by multiplying the amplitude response of the binaural room impulse response and replicating the phase response of the binaural room impulse response and performing the head impulse response and the deconvolution by the impulse response deconvolution means. The head impulse response can be separated from the response to generate a reverberation response (second direction-specific impulse response).

  A reverberation response generation device according to claim 2 is the reverberation response generation device according to claim 1, and includes a head impulse response division unit and a delay amount deletion unit.

  The reverberation response generating apparatus having such a configuration divides the head impulse response into the minimum phase component and the all-pass component by the head impulse response dividing means. Further, the reverberation response generation apparatus deletes the delay amount from the room impulse response for each second direction by the delay amount deleting unit. In addition, the reverberation response generation device deconvolutes the second binaural direction-specific room impulse response and the minimum phase component of the head impulse response for each matching direction by means of the impulse response deconvolution means. The room impulse response for the second direction is calculated. Then, the reverberation response generation device deletes the delay amount obtained as an approximate amount of the all-pass component of the head impulse response from the second direction-specific room impulse response by the delay amount deleting means.

  In this way, the reverberation response generating apparatus divides the head impulse response into the minimum phase component not including the time delay and the all-pass component including the time delay by the head impulse response dividing means. The deconvolution means can perform deconvolution with the second binaural direction-specific room impulse response using the minimum phase component of the head impulse response not including the time delay.

  In order to solve the above-mentioned problem, a reverberation response generation program according to claim 3 is configured to generate a reverberation response for generating reverberation in an acoustic space by using an impulse response convolution unit, a first impulse response conversion unit, The second impulse response converting means, the spectrogram calculating means, the spectrogram converting means, and the impulse response deconvolution means are allowed to function.

  The reverberation response generation program having such a configuration shows acoustic transfer characteristics for each of a plurality of directions from the sound source position in the acoustic space to an arbitrary sound receiving point with reference to the sound receiving point by the impulse response convolution means. A room direction impulse response according to the first direction, and a head impulse response indicating acoustic transfer characteristics in a plurality of directions based on the position of the ear from the sound source position in the anechoic space to the position of the head ear, The room impulse response for each first binaural direction in each direction is calculated by convolution for each matching direction. The reverberation response generation program calculates a spectrogram of the first binaural direction-specific room impulse response by performing a short-time Fourier transform on the first binaural direction-specific room impulse response by the first impulse response conversion means. In addition, the reverberation response generation program generates a binaural room impulse response indicating a sound transfer characteristic from the sound source position in the acoustic space to the position of the ear of the head arranged at the same position as the sound receiving point by the second impulse response conversion means. A spectrogram of the binaural room impulse response is calculated by performing a short-time Fourier transform.

  Further, the reverberation response generation program uses the spectrogram calculating means to calculate the ratio of the amplitude of the spectrogram of the first binaural direction-specific room impulse response in each direction to the total sum of the spectrogram amplitudes of the first binaural direction-specific room impulse responses of all directions. Is multiplied by the amplitude of the spectrogram of the binaural room impulse response, and the spectrogram of the room impulse response of the second binaural direction in each direction is calculated by replicating the phase of the spectrogram of the binaural room impulse response. The reverberation response generation program calculates the second binaural direction-specific room impulse response by inverse short-time Fourier transform of the spectrogram of the second binaural direction-specific room impulse response by the spectrogram conversion means. And the reverberation response generation program deconvolves the second binaural direction-specific room impulse response and the head impulse response for each matching direction, so that the second direction-specific room impulse response in each direction is used as the reverberation response. calculate.

  In this way, the reverberation response generation program uses the spectrogram calculating means to calculate the ratio of the amplitude response of the first binaural direction room impulse response in each direction to the sum of the amplitude responses of the first binaural direction room impulse responses in all directions. Binaural room impulses recorded by binaural recording by multiplying the amplitude response of the binaural room impulse response and replicating the phase response of the binaural room impulse response and performing the head impulse response and the deconvolution by the impulse response deconvolution means. The head impulse response can be separated from the response to generate a reverberation response (second direction-specific impulse response).

  According to the first and third aspects of the present invention, the generated room impulse response according to the second direction and the listener's head impulse response are convoluted with the acoustic signal, so that the listener goes into the actual acoustic space. The reverberation close to what the listener listens to in the same acoustic space can be generated. Moreover, according to the invention which concerns on Claim 1 and Claim 3, even if it is a case where a listener changes, the room impulse response according to the 2nd direction produced | generated and the listener's head impulse response after the said change are produced | generated. Can be generated in the acoustic signal to generate reverberation close to that of the listener who has received the change in the same acoustic space.

  According to the second aspect of the present invention, even when the head impulse response is deconvolved as it is due to factors such as displacement of the microphone and measurement error, even if the causality is not satisfied, the second binaural is satisfied while satisfying the causality. Deconvolution with the direction-specific room impulse can be performed, and the second direction-specific room impulse response indicating the acoustic transfer characteristics such as reflection in the acoustic space can be generated.

It is a block diagram which shows the whole structure of the reverberation response production | generation apparatus which concerns on 1st Embodiment of this invention. It is the schematic for demonstrating the room impulse response in the reverberation response production | generation apparatus which concerns on embodiment of this invention. It is the schematic for demonstrating the room impulse response according to direction in the reverberation response production | generation apparatus which concerns on embodiment of this invention. It is the schematic which shows the room impulse response according to direction in the reverberation response production | generation apparatus which concerns on embodiment of this invention. It is a flowchart which shows the measurement process of a head impulse response in the process sequence of the reverberation response production | generation apparatus which concerns on 1st Embodiment of this invention. It is a flowchart which shows the calculation process of the spectrogram of the 1st binaural direction room impulse response in the process sequence of the reverberation response production | generation apparatus which concerns on 1st Embodiment of this invention. It is a flowchart which shows the calculation process of the room impulse response according to a 2nd direction in the process sequence of the reverberation response production | generation apparatus which concerns on 1st Embodiment of this invention. It is the schematic which shows the procedure which produces | generates a reverberation using the room impulse response according to the 2nd direction produced | generated by the reverberation response production | generation apparatus which concerns on embodiment of this invention, and shows to a listener. It is a block diagram which shows the whole structure of the reverberation response production | generation apparatus which concerns on 2nd Embodiment of this invention. It is a flowchart which shows the calculation process of the minimum phase component of a head impulse response in the process sequence of the reverberation response production | generation apparatus which concerns on 2nd Embodiment of this invention. It is a flowchart which shows the calculation process of the room impulse response classified by 2nd direction in the process sequence of the reverberation response production | generation apparatus which concerns on 2nd Embodiment of this invention.

  A reverberation response generation device and a program thereof according to an embodiment of the present invention will be described with reference to the drawings. In the following description, the same configuration is given the same name and symbol, and detailed description is omitted.

[First Embodiment]
Hereinafter, the configuration of the reverberation response generation device 1 according to the first embodiment of the present invention will be described with reference to FIGS. The reverberation response generation device 1 generates a reverberation response used when generating reverberation. Here, the “reverberation response” indicates a room impulse response (reflection component) for each finite number (for example, N) of arrival directions with reference to an arbitrary sound receiving point in a predetermined acoustic space. Specifically, as shown in FIG. 1, it means the room impulse direction by the second direction.

  As shown in FIG. 1, the reverberation response generation device 1 receives a head impulse response, a first direction-specific room impulse response, and a binaural room impulse response as input, and generates a second direction-specific room impulse response. That is, as will be described later, the reverberation response generation device 1 uses the head impulse response and the first direction-specific room impulse response to detect the head from a binaural room impulse response that is a head impulse response including a reflection component. The impulse response is separated, and a second room-specific room impulse response including only a reflection component is generated. The room direction impulse response according to the second direction generated in this manner is convoluted with the acoustic signal together with the listener's head impulse response, as will be described later, thereby generating reverberation in the acoustic space. (See FIG. 8 described later).

  Here, as shown in FIG. 1, the reverberation response generation apparatus 1 includes an impulse response convolution means 10, a first impulse response conversion means 20, a second impulse response conversion means 30, a spectrogram calculation means 40, and a spectrogram conversion means. 50 and impulse response deconvolution means 60. Hereinafter, each element of the reverberation response generation device 1 will be described.

  The impulse response convolution means 10 calculates a room impulse response for each first binaural direction. Here, the “room impulse response by binaural direction” indicates a binaural room impulse response for each finite number (for example, N) of arrival directions with reference to an arbitrary sound receiving point in a predetermined acoustic space. Yes. Specifically, as shown in FIG. 1, the impulse response convolution means 10 is predetermined by convolving (convolution) the room direction impulse response for each first direction and the head impulse response for each matching direction. A first binaural room impulse response is calculated for each of the plurality of directions.

  Here, the head impulse response input to the impulse response convolution means 10 is obtained in advance before processing in the reverberation response generation device 1. This head impulse response is determined in advance, for example, by placing a dummy head or HATS at a predetermined position in the acoustic anechoic chamber, and using the ear position (the binaural membrane position or the binaural ear canal entrance) of these heads as a reference. Each is obtained for each of a plurality of directions, and each is output to the impulse response convolution means 10 as shown in FIG. The head impulse response described above may be obtained using not only a dummy head or HATS but also an actual human head.

  The specific method for obtaining the head impulse response is not particularly limited, and any method can be used. The head impulse response is, for example, the acoustic transfer function from the sound source at a predetermined position when the head is present to the microphone installed at the outer ear ear entrance, and the head center from the sound source at the predetermined position when there is no head. It can be calculated by dividing by the acoustic transfer function up to the microphone placed at the position and performing inverse discrete Fourier transform on this.

  Further, the direction for obtaining the head impulse response is not particularly limited. For example, 8 directions (45 ° intervals) in the horizontal direction with reference to the position of the ear (the binaural tympanic membrane position or the binaural ear canal entrance) such as the above-described dummy head. The head impulse responses for a total of 16 directions of two layers up and down can be obtained. The head impulse response can be obtained for each of the left and right ears, but here it may be obtained for one of the left and right ears. However, the head impulse response may be obtained for each of the left and right ears, and the head impulse response of the left and right ears may be selected and used with good conditions such as S / N.

  The first direction-specific room impulse response input to the impulse response convolution means 10 is obtained in advance before processing in the reverberation response generation apparatus 1. Here, “direction-specific room impulse response” indicates a room impulse response for each finite number (for example, N) of arrival directions with reference to an arbitrary sound receiving point in a predetermined acoustic space. The first direction room impulse response is obtained for each of a plurality of predetermined directions with reference to a predetermined position in the acoustic space, for example, and each is output to the impulse response convolution means 10 as shown in FIG. .

  Here, the signal obtained by removing the direct sound from the binaural room impulse response, that is, the response indicating the reflection component in the acoustic space is a head corresponding to the direction of arrival in the reflected sound coming from various directions as shown in FIG. It is assumed that the signal obtained by convolving the impulse response can be represented by the area on the boundary surface S near the head. Then, the arrival direction of the reflected sound coming from various directions is limited to a finite number (discretization), and the reflected sound coming from various directions is distributed to N arrival directions limited by a conventional method such as panning. Thus, these time-series signals can be approximated and used as direction-specific impulse responses. 2 indicates a dummy head DH used in binaural recording.

  On the other hand, as shown in FIGS. 3 and 4, the reverberation response generating apparatus 1 according to the present invention uses room impulse responses measured in a plurality of predetermined directions as the direction-specific impulse responses. FIG. 4 shows an example of the direction-specific room impulse response, in which the vertical axis indicates amplitude and the horizontal axis indicates time. The room-specific room impulse response can be represented as a waveform whose amplitude gradually decreases with time, as shown in FIG.

  A specific method for obtaining the first-direction room impulse response is not particularly limited, and an arbitrary method can be used. The first direction-specific room impulse response includes, for example, a method of analytically calculating a direction-specific room impulse response based on a predetermined position in the acoustic space by acoustic simulation, or a directional microphone installed at a predetermined position in the acoustic space. A method of actually measuring in a plurality of directions can be used.

  Also, the direction for obtaining the first direction room impulse response is not particularly limited, but a direction that equally divides the entire circumference is desirable. For example, as in the case of the head impulse response, the room impulse response according to the first direction corresponding to a total of 16 directions of 8 directions in the horizontal direction (45 degree intervals) and 2 layers in the vertical direction is obtained with reference to a predetermined position in the acoustic space Can do. That is, the first direction room impulse response and the head impulse response are obtained in the same direction.

  For example, as described above, the impulse response convolution means 10, when the first direction room impulse response and the head impulse response are input for 16 directions, respectively, performs the convolution for the 16 directions by performing the convolution for the corresponding directions. The first binaural direction-specific room impulse responses are calculated and output to the first impulse response conversion means 20.

  The first impulse response conversion means 20 converts the first binaural direction-specific room impulse response into a spectrogram. Specifically, the first impulse response conversion means 20 calculates a spectrogram of the first binaural direction-specific room impulse response by performing a short-time Fourier transform on the first binaural direction-specific room impulse response.

  For example, as described above, when the first binaural direction-specific room impulse responses for 16 directions are input, the first impulse response conversion means 20 calculates a spectrogram for 16 directions by performing a short-time Fourier transform, respectively. These are output to the spectrogram calculating means 40. The short-time Fourier transform is used to analyze changes in the frequency and phase of a time-varying signal such as speech, and a spectrogram is calculated by multiplying the function while shifting the window function and performing Fourier transform on the function. .

  The second impulse response conversion means 30 converts the binaural room impulse response into a spectrogram. Specifically, the second impulse response conversion means 30 calculates a spectrogram of the binaural room impulse response by performing a short-time Fourier transform on the binaural room impulse response. Then, the second impulse response conversion means 30 outputs a spectrogram of the calculated binaural room impulse response to the spectrogram calculation means 40 as shown in FIG.

  Here, the binaural room impulse response input to the second impulse response conversion means 30 is obtained in advance before processing in the reverberation response generation device 1. The binaural room impulse response is the same dummy head as that obtained by measuring the head impulse response input to the impulse response convolution means 10 at the same position as the predetermined position in the acoustic space from which the room direction impulse response according to the first direction is obtained. HATS can be placed and determined using conventional binaural recording techniques. The binaural room impulse response can be obtained for each of the left and right ears, but here it may be obtained for one of the left and right ears. However, a binaural room impulse response may be obtained for each of the left and right ears, and a binaural room impulse response of the left and right ears having a good S / N may be selected and used.

  The spectrogram calculating means 40 calculates a spectrogram of the second binaural direction room impulse response. Specifically, the spectrogram calculation means 40 calculates the ratio of the amplitude of the spectrogram of the first binaural direction-specific room impulse response in each direction to the sum of the spectrogram amplitudes of the first binaural direction-specific room impulse responses of all directions, binaural. By multiplying the amplitude of the spectrogram of the room impulse response and replicating the phase of the spectrogram of the binaural room impulse response, the spectrogram of the second binaural direction room impulse response is calculated for each of a plurality of directions. The duplication of the phase of the spectrogram of the binaural room impulse response means that the phase of the spectrogram of the binaural room impulse response is directly used as the phase of the spectrogram of the second binaural direction-specific room impulse response. In this way, the spectrogram calculating means 40 divides the binaural recorded binaural room impulse response by the ratio (ratio) of the second binaural direction-specific room impulse response in all directions to the total direction, and includes only the reflection component. A room impulse response by two binaural directions is generated.

  More specifically, the spectrogram calculating means 40 calculates the amplitude of the spectrogram of the second binaural direction-specific room impulse response by the following equation (1).

Here, in the formula (1), F _Gi is second binaural direction another room impulse response of the spectrogram, F _B spectrogram of binaural room impulse response, F _Di first direction-specific room impulse response spectrogram in each direction, k is an index of time, f is an index of frequency bin, N is the number of measurement directions of the room impulse response for each first direction, and i is an index of the number of measurement directions of the room impulse response for each first direction. In addition, the denominator of Equation (1) represents the sum of the amplitudes of spectrograms of the omnidirectional room impulse responses in the first direction.

  More specifically, the spectrogram calculating means 40 replicates the phase of the spectrogram of the second binaural direction room impulse response according to the following equation (2).

Here, in the formula (2), ∠F _Gi is second binaural direction another room impulse response spectrogram phase, ∠F _B binaural room impulse response of the spectrogram of the phase, k is the time index, f is the frequency bin , N indicates the number of measurement directions of the first direction room impulse response, and i indicates the index of the number of measurement directions of the first direction room impulse response.

  For example, as described above, when the spectrogram of the first binaural direction room impulse response is input for 16 directions, the spectrogram calculating unit 40 calculates the ratio of the respective amplitudes to the binaural as shown in the above formula (1). By multiplying the amplitude of the spectrogram of the room impulse response and replicating the phase of the spectrogram of the binaural room impulse response, a spectrogram of the second binaural direction-specific room impulse response for 16 directions is calculated, and these are sent to the spectrogram conversion means 50. Output.

  The spectrogram converting means 50 converts the spectrogram of the second binaural direction-specific room impulse response into the second binaural direction-specific room impulse response. Specifically, the spectrogram converting means 50 calculates the second binaural direction-specific room impulse response by performing inverse short-time Fourier transform on the spectrogram of the second binaural direction-specific room impulse response.

  For example, as described above, when the spectrogram of the room impulse response for each second binaural direction is input for 16 directions, the spectrogram conversion unit 50 performs the inverse short-time Fourier transform for each of the 16 bins in the second binaural direction. Separate room impulse responses are calculated and output to the impulse response deconvolution means 60.

  The impulse response deconvolution means 60 calculates a second room-specific room impulse response. As described above, the second room-specific room impulse response indicates a reverberation response. Specifically, the impulse response deconvolution means 60 deconvolves (deconvolves) the second binaural direction-specific room impulse response and the head impulse response for each matching direction, so that the plurality of directions are deconvolved. The room impulse response for the second direction is calculated.

  For example, as described above, when the second binaural direction room impulse response and the head impulse response are input for 16 directions as described above, the impulse response deconvolution means 60 performs the deconvolution in the corresponding directions. The room direction impulse response for the second direction for 16 directions is calculated.

  The reverberation response generating apparatus 1 having the above-described configuration uses the spectrogram calculating unit 40 to calculate the first binaural direction-specific room impulse response in each direction relative to the sum of the amplitude responses of the first binaural direction-specific room impulse responses in all directions. By multiplying the amplitude response ratio by the amplitude response of the binaural room impulse response and replicating the phase response of the binaural room impulse response, the impulse response deconvolution means 60 performs deconvolution with the head impulse response. The head impulse response can be separated from the recorded binaural room impulse response to generate a reverberation response (second direction-specific impulse response).

  Therefore, according to the reverberation response generation device 1, by convolving the generated room impulse response according to the second direction and the head impulse response of the listener into the acoustic signal, the listener does not go into the actual acoustic space, A reverberation similar to that heard by the listener in the same acoustic space can be generated. In addition, according to the reverberation response generation device 1, even when the listener is switched, the generated room impulse response according to the second direction and the head impulse response of the listener after the change are convoluted with the acoustic signal. Thus, it is possible to generate reverberation close to what the listener after the change listens to in the same acoustic space.

[Processing procedure of reverberation response generating apparatus 1]
Hereinafter, the processing procedure of the reverberation response generation device 1 will be described with reference to FIGS. 5 to 7 (refer to FIG. 1 as appropriate). In the following description, as shown in FIGS. 5 to 7, measurement processing of each impulse response (head impulse response, room direction impulse response according to first direction and binaural room impulse response) will be described. In the following description, as shown in FIGS. 5 to 7, the processing of each impulse response is shown in separate drawings, but the order of the drawings and the order of processing of each impulse response are not related. Therefore, for example, the timing of each impulse response measurement (step S10 in FIG. 5, step S20 in FIG. 6, and step S30 in FIG. 7) may be the same or different.

  First, as shown in FIG. 5, when a head impulse such as a dummy head is measured by a measuring unit (not shown) or the like, the reverberation response generating apparatus 1 inputs this to the impulse response convolution unit 10 (step S10). .

  Next, as shown in FIG. 6, the reverberation response generation apparatus 1 inputs the room response in the first direction, for example, by a measurement unit (not shown), etc., to the impulse response convolution unit 10 (step S20). ). Next, as shown in FIG. 6, the reverberation response generating apparatus 1 performs convolution of the first direction room impulse response and the head impulse response by the impulse response convolution means 10, and the first binaural direction-specific room impulse response. Is calculated (step S21). Next, as shown in FIG. 6, the reverberation response generation device 1 calculates the spectrogram by performing a short-time Fourier transform on the first binaural direction room impulse response by the first impulse response conversion means 20 (step S22). ).

  Next, as shown in FIG. 7, when the binaural room impulse response is measured by a measuring unit (not shown) or the like, the reverberation response generating apparatus 1 inputs the binaural room impulse response to the second impulse response converting unit 30 (step S30). . Next, as shown in FIG. 7, the reverberation response generating apparatus 1 calculates a spectrogram by performing a short-time Fourier transform on the binaural room impulse response by the second impulse response conversion unit 30 (step S31).

  Next, as shown in FIG. 7, the reverberation response generating apparatus 1 uses the spectrogram calculating unit 40 to calculate the ratio of the amplitude response (spectrum amplitude) of the first binaural direction-specific room impulse response in each direction to the binaural room impulse. By multiplying the amplitude of the spectrogram of the response, the amplitude of the spectrogram of the second binaural direction-specific room impulse response is calculated (step S32). Next, as shown in FIG. 7, the reverberation response generation apparatus 1 duplicates the phase response of the binaural room impulse response (the phase of the spectrogram) by the spectrogram calculating means 40, and obtains the spectrogram of the second binaural direction-specific room impulse response. Calculate (step S33).

  Next, as shown in FIG. 7, the reverberation response generation apparatus 1 performs a short-time Fourier transform on the spectrogram of the second binaural direction-specific room impulse response by the spectrogram conversion unit 50, thereby generating the second binaural direction-specific room impulse response. Is calculated (step S34). Next, as shown in FIG. 7, the reverberation response generation device 1 performs the second direction by deconvolution of the second binaural direction room impulse response and the head impulse response by the impulse response deconvolution means 60. Another room impulse response is calculated (step S35), and the process ends.

  The reverberation response generation apparatus 1 separates the head impulse response from the binaural recorded binaural room impulse response and generates a reverberation response (second direction-specific impulse response) by the procedure described above.

[Procedure for presenting reverberation to the listener]
Hereinafter, a procedure of generating reverberation using the second direction-specific room impulse response generated by the reverberation response generating device 1 and presenting (listening to) the listener will be described with reference to FIG. In the following description, as shown in FIG. 8, the reverberation response generation apparatus 1 generates room impulse responses for the second direction (N pieces) in the second direction in advance, and measurement means (not shown), etc. It is assumed that the head impulse responses of both ears in the N direction (N) of the listener L are measured in advance. In addition, it is assumed that the second direction room impulse response and the head impulse response are generated and measured in directions that coincide with each other with respect to the sound receiving point or the listener L.

  First, as shown in FIG. 8, the second direction-specific room impulse responses for the N directions generated by the reverberation response generation device 1 are convolved with the acoustic signal (step S100). Thereby, N different room impulse responses according to the second direction are convolved with the acoustic signal, and N signals are generated.

  Next, as shown in FIG. 8, the head impulse responses in the matching direction are convolved with the N signals generated in step S100 for the left and right ears, respectively (step S101). Thereby, N different head impulse responses of the left ear and N different head impulse responses of the right ear are convolved with the N signals, and 2N signals are generated.

  Next, as shown in FIG. 8, the 2N signals generated in step S101 are added for each left ear and each right ear and presented to the left and right ears of the listener L via headphones or the like, respectively. (Step S102).

  Through the above procedure, the reverberation of the acoustic space is generated, and the reverberation that reproduces the sound as if the listener L is in the acoustic space is presented to the listener L. Even when the listener L in FIG. 8 is replaced with another listener, the head impulse response used in step S101 is changed to the head impulse response of the other listener. The listener is presented with reverberation that reproduces the sound as if in an acoustic space.

[Second Embodiment]
Hereinafter, the configuration of a reverberation response generation device 1A according to the second embodiment of the present invention will be described with reference to FIG. Here, as shown in FIG. 9, the reverberation response generating apparatus 1 </ b> A includes an impulse response deconvolution unit 60 </ b> A instead of the impulse response deconvolution unit 60, and includes a head impulse response division unit 70 and a delay amount deletion unit 80. Except for this, it has the same configuration as the reverberation response generation device 1 according to the first embodiment described above. Therefore, about the structure which overlaps with above-described reverberation response production | generation apparatus 1, the same code | symbol is attached | subjected and description is abbreviate | omitted and description about operation | movement is also abbreviate | omitted. In FIG. 9, means that are changed or newly added from the reverberation response generation apparatus 1 shown in FIG. 1 are indicated by thick-line blocks.

  Similar to the impulse response deconvolution means 60, the impulse response deconvolution means 60A calculates a second room-specific room impulse response. Specifically, the impulse response deconvolution means 60A deconvolves the second binaural direction-specific room impulse response and the minimum phase component of the head impulse response for each matching direction, so that the second direction-specific room Calculate the impulse response. The minimum phase component will be described later.

  The head impulse response dividing means 70 divides each component of the head impulse response. Specifically, the head impulse response dividing means 70 divides the head impulse response into a minimum phase component and an all-pass component. Here, as described above, the head impulse response is a sound transfer characteristic from the sound source position in the free sound field to the ear position (the binaural tympanic membrane position or the binaural ear canal entrance), and therefore includes a time delay component. ing. If the room impulse response by the second binaural direction is deconvolved as it is, it is usually offset by the time delay component of the room impulse response by the second binaural direction, but due to factors such as the displacement of the microphone and the measurement error, May not meet causality. Therefore, the head impulse response dividing means 70 divides the head impulse response into a minimum phase component that does not include a time delay component and an all-pass component that includes a time delay component. The convolution means 60A can perform deconvolution of the minimum phase component not including the time delay component and the second binaural direction-specific room impulse response.

  For example, as described above, when the head impulse response is input for 16 directions, the head impulse response dividing means 70 divides each into a minimum phase component and an all-pass component, thereby obtaining the minimum phase for 16 directions. A component and an all-pass component for 16 directions are generated. Then, the head impulse response dividing means 70 outputs the minimum phase component for 16 directions to the impulse response deconvolution means 60A and the all-pass component for 16 directions to the delay amount deleting means 80 as shown in FIG. Output.

  The delay amount deleting unit 80 deletes the delay amount from the second direction-specific room impulse response. Specifically, the delay amount deleting unit 80, when the delay amount obtained as an approximate amount of the all-pass component of the head impulse response is smaller than the time delay component (delay amount) of the second direction room impulse response, The delay amount (approximate amount of the all-pass component) is deleted from the room impulse response for each second direction. On the other hand, when the delay amount obtained as an approximate amount of the all-pass component of the head impulse response is larger than the time delay component (delay amount) of the second-direction room impulse response, the delay amount deleting unit 80 The delay amount of the second room-specific room impulse response is deleted from the separate room impulse response. That is, the delay amount deleting unit 80 deletes the sample corresponding to the delay amount from the head portion of the second-direction room impulse response. The reason why the delay amount is deleted from the second-direction room impulse response in this way is that the response is delayed by the above-described delay amount only by deconvolution with the minimum phase component of the head impulse response. .

  The reverberation response generating apparatus 1A having the above configuration divides the head impulse response into the minimum phase component not including the time delay and the all-pass component including the time delay by the head impulse response dividing means 70. Therefore, the impulse response deconvolution means 60A can perform deconvolution with the second binaural direction room impulse response using the minimum phase component of the head impulse response not including the time delay.

  Therefore, according to the reverberation response generation device 1A, even if the head impulse response is deconvolved as it is due to factors such as the displacement of the microphone and the measurement error, even if the causality is not satisfied, the second is performed while the causality is satisfied. Deconvolution with the binaural direction-specific room impulse can be performed, and a second direction-specific room impulse response indicating acoustic transfer characteristics such as reflection in the acoustic space can be generated.

[Processing procedure of reverberation response generating apparatus 1A]
Hereinafter, the processing procedure of the reverberation response generation device 1 </ b> A will be described with reference to FIGS. 10 and 11. In the following description, as shown in FIGS. 10 and 11, measurement processing of each impulse response (head impulse response and binaural room impulse response) will be described. Further, the spectrogram calculation process of the first binaural direction-specific room impulse response is the same as that of the reverberation response generation device 1 described above (see FIG. 6), and thus the description thereof is omitted here.

  First, as shown in FIG. 10, when the head impulse response such as a dummy head is measured by a measuring unit (not shown) or the like, the reverberation response generating apparatus 1A inputs this to the head impulse response dividing unit 70 ( Step S40). Next, as shown in FIG. 10, the reverberation response generating device 1A divides the head impulse response into the minimum phase component and the all-pass component by the head impulse response dividing means 70 (step S41).

  Next, as shown in FIG. 11, when the binaural room impulse response is measured by, for example, a measurement unit (not shown), the reverberation response generation device 1A inputs the binaural room impulse response to the second impulse response conversion unit 30 (step S50). . Next, as shown in FIG. 11, the reverberation response generation device 1 </ b> A calculates a spectrogram by performing a short-time Fourier transform on the binaural room impulse response by the second impulse response conversion unit 30 (step S <b> 51).

  Next, as shown in FIG. 11, the reverberation response generating apparatus 1 </ b> A uses the spectrogram calculating unit 40 to calculate the ratio of the amplitude response of the first binaural direction-specific room impulse response (amplitude of the spectrogram) to the entire binaural room impulse. By multiplying the response amplitude response, the amplitude response of the second binaural direction-specific room impulse response is calculated (step S52). Next, as shown in FIG. 11, the reverberation response generation device 1 </ b> A replicates the phase response of the binaural room impulse response (the phase of the spectrogram) by the spectrogram calculation unit 40, and obtains the spectrogram of the second binaural direction-specific room impulse response. Calculate (step S53).

  Next, as shown in FIG. 11, the reverberation response generation device 1 </ b> A performs a short-time Fourier transform on the spectrogram of the second binaural direction-specific room impulse response by the spectrogram conversion unit 50, thereby generating the second binaural direction-specific room impulse response. Is calculated (step S54). Next, as shown in FIG. 11, the reverberation response generation device 1A performs deconvolution of the second binaural direction room impulse response and the minimum phase component of the head impulse response by the impulse response deconvolution means 60A. Then, the room direction impulse response in the second direction is calculated (step S55). Then, as shown in FIG. 11, the reverberation response generation device 1A deletes the delay amount obtained as the approximate amount of the all-pass component from the room impulse response for each second direction by the delay amount deleting unit 80 (step S56). The process is terminated.

  The reverberation response generation device 1A separates the head impulse response from the binaural recorded binaural room impulse response and performs the reverberation response even when the head impulse response includes a time delay component by the above procedure. (Impulse response according to the second direction) can be generated.

[Reverberation response generation program]
The above-described reverberation response generation apparatuses 1 and 1A can be realized by operating a general computer with a program that functions as each of the above-described units and units. This program can be distributed via a communication line, or can be written on a recording medium such as a CD-ROM for distribution.

[Reverberation generator]
Hereinafter, a reverberation generating device that generates reverberation using the above-described second-direction room impulse response will be described. The reverberation generation device includes, in addition to the components of the above-described reverberation response generation devices 1 and 1A, second-direction room impulse response convolution means, head impulse response convolution means, and signal addition means.

  The second direction-specific room impulse response convolution means convolves the second direction-specific room impulse responses generated by the reverberation response generation apparatuses 1 and 1A and the acoustic signal. As a result, as shown in FIG. 8, N different room impulse responses according to the second direction are convoluted with the acoustic signal, and N signals are generated. The second direction-specific room impulse response convolution means then outputs the generated N signals to the head impulse response convolution means.

  The head impulse response convolution means convolves the signal generated by the second direction room impulse response convolution means with the left and right head impulse responses of the listener. As a result, as shown in FIG. 8, the N signals have different N left-ear head impulse responses and N different right-ear head impulse responses convolved with each other, 2N signals are generated.

  The signal adding means adds 2N signals generated by the head impulse response convolution means for each left ear and each right ear. Thereby, as shown in FIG. 8 described above, a signal for the left ear and a signal for the right ear are generated. Then, as shown in FIG. 8, the reverberation is generated by presenting the signal generated in this way to the left and right ears of the listener L via headphones or the like.

  The reverberation response generation apparatus and the program thereof according to the present invention have been specifically described above by the mode for carrying out the invention, but the gist of the present invention is not limited to these descriptions, and the scope of the claims Should be interpreted broadly based on the description. Needless to say, various changes and modifications based on these descriptions are also included in the spirit of the present invention.

  Here, as a specific application of the reverberation response generating apparatus according to the present invention, for example, when a program production engineer performs editing of a recorded program while wearing headphones, the program is actually transmitted to the engineer. It is assumed that the reverberation of an acoustic space such as a studio that records is presented. In this case, head impulse responses for a plurality of engineers are measured in advance in an acoustic anechoic chamber. And reverberation is produced | generated by convolution with the reverberation response (room impulse response according to 2nd direction) produced | generated by the reverberation response production | generation apparatus which concerns on this invention, switching a head impulse response for every engineer who performs editing work. This makes it possible to present reverberation that reproduces the reverberation as if it were actually in the studio to an engineer who edits the recorded program.

1, 1A Reverberation response generating apparatus 10 Impulse response convolution means 20 First impulse response conversion means 30 Second impulse response conversion means 40 Spectrogram calculation means 50 Spectrogram conversion means 60, 60A Impulse response deconvolution means 70 Head impulse response division means 80 Delay amount deletion means DH Dummy head L

Claims (3)

A reverberation response generator for generating a reverberation response for generating reverberation in an acoustic space,
A first room-specific room impulse response indicating a sound transfer characteristic for each of a plurality of directions from the sound source position in the acoustic space to an arbitrary sound receiving point as a reference, and the sound source position in the anechoic space to the head. The first binaural of each direction is convolved with the head impulse response indicating the acoustic transfer characteristics in each of the plurality of directions with respect to the position of the ear up to the position of the ear of each part for each of the matching directions. Impulse response convolution means for calculating a direction-specific room impulse response;
A first impulse response conversion means for calculating a spectrogram of the first binaural direction-specific room impulse response by performing a short-time Fourier transform on the first binaural direction-specific room impulse response;
The binaural room impulse response is obtained by performing a short-time Fourier transform on the binaural room impulse response indicating the acoustic transfer characteristic from the sound source position in the acoustic space to the position of the ear of the head arranged at the same position as the sound receiving point. Second impulse response conversion means for calculating a spectrogram of
The ratio of the spectrogram amplitude of the first binaural direction room impulse response in each direction to the sum total of the spectrogram amplitude of the first binaural direction room impulse response in all directions is set to the spectrogram amplitude of the binaural room impulse response. A spectrogram calculating means for calculating a spectrogram of the second binaural direction-specific room impulse response in each direction by multiplying, by copying the phase of the spectrogram of the binaural room impulse response,
A spectrogram converting means for calculating a second impulse response according to the binaural direction by performing an inverse short-time Fourier transform on the spectrogram of the second impulse response according to the binaural direction;
Impulse response for calculating the second room-specific room impulse response in each direction as the reverberation response by deconvolution of the second binaural direction-specific room impulse response and the head impulse response for each of the matching directions. Deconvolution means,
A reverberation response generating apparatus comprising: A head impulse response splitting means for splitting the head impulse response into a minimum phase component and an all-pass component;
A delay amount deleting means for deleting a delay amount from the second direction-specific room impulse response,
The impulse response deconvolution means deconvolutes the second binaural direction-specific room impulse response and the minimum phase component of the head impulse response for each of the matching directions, so that the second in each direction. Calculate the room impulse response by direction,
2. The reverberation according to claim 1, wherein the delay amount deleting unit deletes the delay amount obtained as an approximate amount of the all-pass component of the head impulse response from the second direction-specific room impulse response. Response generator. To generate a reverberation response to generate reverberation in acoustic space,
A first room-specific room impulse response indicating a sound transfer characteristic for each of a plurality of directions from the sound source position in the acoustic space to an arbitrary sound receiving point as a reference, and the sound source position in the anechoic space to the head. The first binaural of each direction is convolved with the head impulse response indicating the acoustic transfer characteristics in each of the plurality of directions with respect to the position of the ear up to the position of the ear of each part for each of the matching directions. Impulse response convolution means for calculating the direction-specific room impulse response,
First impulse response conversion means for calculating a spectrogram of the first binaural direction-specific room impulse response by performing a short-time Fourier transform on the first binaural direction-specific room impulse response;
The binaural room impulse response is obtained by performing a short-time Fourier transform on the binaural room impulse response indicating the acoustic transfer characteristic from the sound source position in the acoustic space to the position of the ear of the head arranged at the same position as the sound receiving point. Second impulse response conversion means for calculating a spectrogram of
The ratio of the amplitude of the spectrogram of the room impulse response by the first binaural direction in each direction to the sum of the amplitudes of the spectrogram of the room impulse response by the first binaural direction in all directions is set to the amplitude of the spectrogram of the binaural room impulse response. A spectrogram calculating means for calculating the spectrogram of the second binaural direction-specific room impulse response in each direction by multiplying the phase of the spectrogram of the binaural room impulse response,
A spectrogram converting means for calculating a room impulse response according to the second binaural direction by performing inverse short-time Fourier transform on the spectrogram of the room impulse response according to the second binaural direction;
Impulse response for calculating the second room-specific room impulse response in each direction as the reverberation response by deconvolution of the second binaural direction-specific room impulse response and the head impulse response for each of the matching directions. Deconvolution means,
Reverberation response generation program to function as.

Images