JP4677587B2 - Apparatus and method for controlling sense of distance in auditory reproduction - Google Patents

Description

  The present invention relates to a technology for controlling a sense of distance in the sense of hearing of a reproduced sound when reproducing an audio signal as a sound.

Various techniques have been proposed for controlling the sense of distance given to the listener in sound reproduction, that is, the distance from the reproduced sound to the sound source that the listener feels. For example, in Patent Document 1, an audio signal in which a reflected sound component is added to a direct sound component is generated by a reflected sound adding circuit, and the level ratio and time interval between the direct sound component and the reflected sound component at that time are adjusted. We have proposed a technology that controls the sense of distance given to the listener. Further, Patent Document 2 provides two sound reproduction systems for direct sound and indirect sound, and a sense of distance given to the listener by adjusting the level ratio of direct sound and indirect sound reproduced from each sound reproduction system. The technology to control
JP-A-6-315200 JP-A-9-121400

  By the way, as described above, in order to control the sense of distance by adjusting the level ratio of the direct sound and the reflected sound or the indirect sound, the direct sound and the reflected sound or the indirect sound signal must be obtained in a separated state. I must. Since the techniques of Patent Documents 1 and 2 described above both generate a reflected sound signal or an indirect sound signal by performing predetermined signal processing on a direct sound signal obtained from a sound source. The level ratio between sound and reflected sound or indirect sound can be adjusted. However, with regard to sense of distance control, it is possible to freely control the sense of distance given to the listener when the audio signal is mixed with direct sound components and reverberation sound components collected in a reverberant sound field. There is also a demand to do it. A technology that meets such a demand has not been provided.

  The present invention has been made in view of the circumstances described above, and is a technique that can process an already-obtained audio signal including a direct sound component and a reverberation sound component and control a sense of distance given to a listener at the time of reproduction. The purpose is to provide an effective means.

The present invention extracts an extraction means for extracting a common vector of an audio signal of each channel from a plurality of channels of an audio signal, and extracting a difference vector between each vector of the audio signal of each channel and the common vector, and the extraction means And a synthesizing unit that generates a multi-channel audio signal by synthesizing the common vector and the difference vector obtained by the above in a specified ratio.
According to such a sense of distance control device, a common vector close to the direct sound component and a difference vector close to the reverberant sound component are extracted from the obtained audio signal including the direct sound component and the reverberant sound component, and these are expressed at a specified ratio. A multi-channel audio signal is obtained by synthesizing. At this time, the ratio of the common vector and the difference vector can be appropriately adjusted to control the sense of distance given to the listener by reproducing the audio signals of a plurality of channels.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram showing the configuration of a sense of distance control apparatus according to an embodiment of the present invention. In this sense of distance control device, the input unit 1 is a device that takes in an L-channel audio signal a _L and an R-channel audio signal a _R as processing targets from the outside. The input unit 1 may be a device that receives an audio signal via a network, or may be a device that reproduces an audio signal stored in a recording medium. The first waveform memory 2 is a memory that stores the audio signals a _L and a _R captured by the input unit 1. The signal processing unit 3 is a device that reads the audio signals a _L and a _R from the first waveform memory 2 and performs signal processing for distance sense control. The contents of this signal processing will be described later. The operation unit 4 is a device that receives various commands and information from the user. Information input via the operation unit 4 includes a parameter k for distance sense control. In the signal processing performed by the signal processing unit 3, the parameter k is referred to. Memory 5 and the second waveform is a memory for storing the audio signal b _L and b _R obtained by the signal processing for the sense of distance control. The output unit 6 is a device that reads out the audio signals b _L and b _R from the second waveform memory 5 and outputs them as sound from left and right speakers (not shown).

  Next, the principle of distance control in this embodiment will be described. First, in the reverberation room 100 illustrated in FIG. 2, when sound is output from the sound source 101, the microphones 102 </ b> L and 102 </ b> R at the sound receiving point separated from the sound source 101 by the distance d receive direct sound from the sound source 101. Reverberant sound. The energy ratio between the direct sound and the reverberant sound at the sound receiving point depends on the distance d between the sound source and the sound receiving point. The critical distance is a distance d between the sound source and the sound receiving point at which the energy ratio between the direct sound and the reverberant sound becomes equal.

  When the distance d between the sound source and the sound receiving point is longer than the critical distance, there is almost no direct sound at the sound receiving point, and the reverberant sound is dominant. For this reason, the signals that reach the microphones 102L and 102R at the sound receiving point are nearly uncorrelated, and it is impossible to obtain a sense of distance even when listening to the sound collected by these microphones.

  On the other hand, when the distance d between the sound source and the sound receiving point is shorter than the critical distance, an in-phase direct sound reaches the sound receiving point with sufficient intensity, and the energy ratio of the direct sound to the reverberant sound is the sound source. It becomes small according to the distance from 101. For this reason, a sense of distance can be obtained from each sound picked up by the microphones 102L and 102R at the sound receiving point. Therefore, if the direct sound component and the reverberation sound component are separated and extracted from the audio signals obtained by the microphones 102L and 102R and the energy ratio of the direct sound component and the reverberation sound component can be adjusted, the audio signal can be received by the listener. It is thought that the sense of distance given to the can be changed. However, there is currently no method for strictly separating and extracting a direct sound component and a reverberant sound component from an audio signal.

Therefore, the signal processing unit 3 in the present embodiment executes an extraction process 31 and a synthesis process 32 described below, thereby controlling the sense of distance that the audio signal gives to the listener. In FIGS. 3A to 3E referred to in the following description, the horizontal axis is time, and the vertical axis is a value indicated by the sample in the sample sequence of the audio signal. First, in the case where audio signals a _L and a _R composed of a plurality of samples as illustrated in FIGS. 3A and 3B are given as processing targets, in the present embodiment, each audio signal a _L and a _R is considered as a vector (hereinafter referred to as vectors a _L and a _R. Vectors corresponding to other signals also use the same vector name as the signal name), and each audio signal sample is considered as a vector component. Then, the extraction process 31, according to the following equation, the vector _{a L} and _a common vector _{a com} spatially common component of _R, the difference vector between the common vector _{a com} vector _{a L} and _{a R a L- com} and a _R-com .
Δ = a _L −a _R (1)
z = Δ / ‖Δ‖ (2)
_{a com = a L - <a L} ,z> z = a R - <a R ,z> z ...... (3)
_{a L-com = a L -a} com = <a L ,z> z ...... (4)
_{a R-com = a R -a} com = <a R ,z> z ...... (5)
However, in the above formula (2), z is the inner product of vectors _{a L} and _a unit vector of the difference vector Δ same direction _R, the formula (3) ~ _<a L ,z> like two in (5) Vector Represents.

FIGS. 3C to 3E show a sample sequence that is a component of the common vector a _com obtained by the extraction process 31, a sample sequence that is a component of the difference vector a _L-com , and the difference vector a _R-com . Each sample row that is a component is illustrated. FIG. 4 shows the relationship in the vector space between the vectors a _L and a _R , the common vector a _com, and the difference vectors a _L-com and a _R-com corresponding to the original L and R2 channel audio signals. Show. As shown in FIG. 4, the common vector a _com becomes a vector orthogonal to the difference vector between the vectors a _L and a _R , and both the difference vectors a _L-com and a _R-com are orthogonal to the common vector a _com . Vector. Thus, in the extraction process 31, the original audio signal a _L (a _R ) is decomposed into a common vector a _com and a difference vector a _L-com (a _R-com ), which are vectors orthogonal to each other in the vector space. To do. Here, since the common vector a _com indicates an acoustic characteristic common to both the L and R channels, it is considered that it corresponds to the direct sound component collected at the sound receiving point. The difference vector a _L-com (a _R-com ) exhibits an acoustic characteristic different from that of the common vector a _com. Therefore, the sound other than the direct sound component collected at the sound receiving point, that is, reverberation. It is considered that it corresponds to a reverberant sound component that is created under the participation of the room 100 and has an acoustic characteristic different from that of the direct sound component.

In the synthesis process 32, the vectors b _L and b _R are obtained according to the following equations (6) and (7) using each vector obtained by the extraction process 31 described above and the parameter k given through the operation unit 4. Ask.
b _L = a _L-com + k · a _com (6)
b _R = a _R-com + k · a _com (7)
Then, the vector b _L component is written in the second waveform memory 5 as a sample of the L channel audio signal b _L and the vector b _R component is written as a sample of the R channel audio signal b _R. Samples of the audio signals of the L channel and the R channel are sequentially read from the second waveform memory 5 at a predetermined rate and output as sound by the output unit 6.

The extraction process 31 and the synthesis process 32 are preferably executed in an appropriate manner in accordance with the time lengths of the audio signals a _L and a _{R to be} processed. When the audio signals a _L and a _{R to} be processed are single-shot and have a relatively short duration, all samples of each audio signal are regarded as vectors, and the above-described extraction processing 31 and The synthesis process 32 may be executed to control the sense of distance. When the audio signals a _L and a _{R to} be processed are those having a long duration or a sequence of a plurality of sounds, both audio signals are divided into frames having a certain length of time, and the above-mentioned units of frames are described. The extracted processing 31 and the combining processing 32 may be executed to control the sense of distance.

In the configuration described above, if the parameter k given from the operation unit 4 is 1, the same audio signal a _L (a _R ) as the original audio signal a _L (b _R ) is output by the output unit 6. On the other hand, when the parameter k is larger than 1, the energy ratio of the common vector k · a _{com to} the difference vector a _L-com (a _R-com ) in the audio signal b _L (b _R ) output by the output unit 6 is , Larger than that of the original audio signal a _L (a _R ). Therefore, the listener feels the position of the sound source closer compared to the case where the same audio signal a _L (a _R ) as the original audio signal a _L (a _R ) is output by the output unit 6 (that is, when k = 1). Conversely, when the parameter k is smaller than 1, the energy ratio of the common vector k · a _{com to} the difference vector a _L-com (a _R-com ) in the audio signal b _L (b _R ) is equal to the original audio signal aL ( It is smaller than that of aR). Therefore, the listener feels the position of the sound source farther than when k = 1.

In order to confirm the effect of this embodiment, the inventors of the present application output impulse sound from the sound source 101 while changing the distance d between the sound source and the receiving point in the rectangular parallelepiped reverberation chamber 100 as shown in FIG. The impulse response waveform was collected by the microphones 102L and 102R. The size of the reverberation chamber 100, 5.7 m × 5.9 m × 4.7 m, reverberation time _{T R} of the reverberation chamber 100 was about 6.5 seconds. The distance between the microphones 102L and 102R was 0.2 m, and the distance d between the sound source and the sound receiving point was changed from 0.1 m to 1.5 m in increments of 0.1 m. Then, with respect to various distances d between sound sources and receiving points, common vectors and difference vectors are obtained from the collected impulse response waveforms of the L and R2 channels by the same process as the extraction process 31 described above, and the energy of the common vector and the difference vector is obtained. The ratio E _com / E _def was calculated.

FIG. 5 shows the energy ratio E _com / E _def obtained in this way and the energy ratio E _d / E _rev between the direct sound and the reverberant sound calculated by the following theoretical formula (8). .
E _d / E _rev = 0.01 V / (πT _R d ² ) (8)
In the above formula (8), E _d is the energy of the direct sound at the sound receiving point, E _rev is the energy of the reverberant sound at the sound receiving point, and V is the volume of the reverberation chamber 100 (in this example, 5.7 m × 5 .9m × 4.7m), _{T R} is the reverberation time of the reverberation chamber 100 (6.5 seconds in this example), d is the distance between the sound source to receiving point.

As shown in FIG. 5, the energy ratio E _com / E _def between the common vector and the difference vector in a region within a critical distance where the energy ratio E _d / E _rev between the direct sound and the reverberant sound is 1 (= 0 dB). The energy ratio E _d / E _rev between the direct sound and the reverberant sound is a substantially parallel curve. Therefore, by adjusting the energy ratio between the common vector and the difference vector as described above, the sense of distance that the audio signal gives to the listener can be controlled in the same manner as when the energy ratio between the direct sound and the reverberant sound is adjusted. it can.

  There is a reverberation effect device as an application of the sense of distance control device according to the present embodiment. As a general form of this reverberation effect device, a filter coefficient sequence obtained by sampling an impulse response waveform obtained in various reverberation sound fields is stored, and this filter coefficient sequence is convolved with a sample sequence of an input audio signal to output an audio signal. There is something that generates When such a reverberation effect device is used, the distance between the sound source and the sound receiving point obtained by applying the reverberation effect may not meet the user's request. In other words, the filter coefficient sequence corresponding to the user's desired reverberation space is collected under the condition that the distance d between the sound sources and the sound receiving points is d1, but the user has the distance d between the sound sources and the sound receiving points. Is desired to be set to d2. In such a case, in order to meet the user's request, the impulse response waveform is collected under the condition that the distance d between the sound source and the sound receiving point is d2 in the user's desired reverberation sound field, and a filter is obtained from the impulse response waveform. It is necessary to obtain a coefficient sequence. However, such a process requires a lot of labor and is difficult. The distance sensation control apparatus according to the present embodiment is useful for solving such problems.

FIG. 6 is a block diagram showing the configuration of a reverberation effect device configured using the sense of distance control device of FIG. In this reverberation effect device, a sample train of impulse response waveforms of L and R2 channels collected in various reverberation sound fields is stored in the first waveform memory 2 via the input unit 1. When the user instructs the selection of the impulse response waveform of the desired reverberation field by operating the operation section 4, designated The L, sample sequence a _L and a _R of R2 channel impulse response waveform is read from the first waveform memory 2 Is output to the signal processing unit 3. In the signal processing unit 3, the extraction process 31 described above is performed using the sample sequences a _L and a _R , and then the synthesis process 32 is performed using the parameter k given from the operation unit 4. Then, the sample trains b _L and b _R of the impulse response waveform obtained by the synthesis process 32 are stored in the second waveform memory 5. The convolution unit 7 receives a sample of a monaural input audio signal from the outside at every fixed sampling period, and applies the L-channel sample string b _L stored in the second waveform memory 5 to the received predetermined number of sample strings. generating an audio signal C _L of L channel convoluting, also convolving the sample sequence b _R a R channel to sample sequences of the same past predetermined number to generate an audio signal C _R of the R channel. The output unit 6 is to output such audio signal is generated in the C _L and C _R as sound.

Here, if the parameter k is larger than 1, the energy ratio between the common vector and the difference vector in the sample sequence b _L and b _R of the impulse response waveform used for the convolution operation is the sample sequence a _L and a _{R of the} original impulse response waveform. Larger than For this reason, the distance between the sound source and the sound receiving point that the listener feels is made shorter than when the reverberation effect is given by using the sample sequence a _L and a _R of the original impulse response waveform for the convolution calculation. Can do. In contrast, when less than 1 parameter k, the energy ratio of the common vector and difference vector of the sample sequence b _L and b _R of the impulse response waveform is smaller than that of the original impulse response waveform. For this reason, the distance between the sound source and the sound receiving point that the listener feels is made longer than when the reverberation effect is given by using the sample sequence a _L and a _R of the original impulse response waveform for the convolution calculation. Can do.

  Although one embodiment of the present invention has been described above, other embodiments are possible for the present invention. For example, in the above embodiment, the sense of distance is controlled by processing 2-channel audio signals obtained by two microphones, but the audio signals obtained by a larger number of microphones are processed and the distance is controlled. You may make it control feeling.

It is a block diagram which shows the structure of the distance feeling control apparatus which is one Embodiment of this invention. It is a figure which shows the example of the reverberation room where the audio signal used as the process target is extract | collected in the same embodiment. It is a wave form diagram which shows a process in the signal processing in the embodiment. It is a figure which shows the relationship of the various vectors handled in the embodiment. It is a figure which shows the relationship between the energy ratio of the common vector and difference vector which are handled in the embodiment, and the energy ratio of a direct sound and a reverberation sound. It is a block diagram which shows the structure of the reverberation effect apparatus which is other Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Input part, 2 ... 1st waveform memory, 3 ... Signal processing part, 31 ... Extraction process, 32 ... Synthesis process, 4 ... Operation part, 5 ... 2nd waveform memory, 6 ... Output unit, 7 ... convolution operation unit.

Claims (3)

For each of the audio signals of a plurality of channels, a vector having each sample of the sample sequence of the audio signal as a component is generated, a common vector is extracted from each vector of the audio signal of each channel, and each audio signal of each channel is Extraction means for extracting each difference vector between the vector and the common vector;
Sense of distance control device, characterized in that by combining in a ratio to the specified common vector and difference vector obtained comprising a synthesizing means for generating audio signals of a plurality of channels by the extraction unit. The sense of distance control device according to claim 1;
The reverberation effect device comprising: an effect applying unit that applies a reverberation effect to the input audio signal using a plurality of channels of audio signals obtained by the synthesizing unit of the distance control device. For each of the audio signals of a plurality of channels, a vector having each sample of the sample sequence of the audio signal as a component is generated, a common vector is extracted from each vector of the audio signal of each channel, and each audio signal of each channel is An extraction process for extracting a difference vector between the vector and the common vector,
Combining a common vector and a difference vector obtained by the extraction process at a specified ratio to generate a multi-channel audio signal;
A sense of distance control method comprising:

Images