JPH01144814A - Sound field reproducing device - Google Patents

Abstract

PURPOSE:To realize the reproduction of a sound field having good attendant feeling by executing the convolutional arithmetic operation of a part where the delay time of impulse response is big with lower sampling frequency than that in an A/D part. CONSTITUTION:A first convolution means 3 and a second convolution means 9 for convolution-processing a digital signal outputted from a low-pass filter down-sampling means 8 with the lower sampling frequency than that of the means 3 are provided. Then, the impulse response in an audition space such as a hall where the reproduction is desired to be realized is divided to plural time areas and the impulse row included in the late time area and the convoluted digital signal of a music source are made to be a signal low-pass-filtered with low cutoff frequency. Then, the multiplication and addition of plural times is executed with the low sampling frequency. Thus, the live sound field having the long impulse response or the sound field of the audition space of big capacity is faithfully realized.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a sound field reproducing device for reproducing realistic sound.

Conventional technology In recent years, realistic sound reproduction has been desired, and the impulse response of the hall has been determined by measurement or simulation.
Equipment has been developed that convolutes this signal with a music signal and reproduces the result using speakers placed around the listener.

FIG. 4 shows a block diagram of the conventional sound field reproducing device described above, and FIG. 6 a shows the impulse response of a space such as a hall to be realized. FIG. 6 shows an example of the use of a data memory used in a conventional sound field reproducing apparatus.

In FIG. 4, 1 is an input terminal into which a music signal is input;
2 is an A/D unit for converting an input music signal into a digital signal; 26 is a data memory for storing the digital signal; 26 is a coefficient for storing tap coefficients based on impulse responses of halls, etc. 24 is a convolution unit for convolving the digital signal and the tap coefficient; 22 is a D/A unit for converting the output signal from the convolution unit 24 into an analog signal; 23 is an output terminal; It is.

In the conventional sound field reproduction device configured as described above,
The digital signal output from the A/D section 2 is written into the data memory 26. As shown in FIG. 6, the data memory 25 forms a ring buffer of addresses 0 to FFFF, for example, and the digital signal is written to the next younger address at each sampling period. Then, every sampling period, the digital signal of the address corresponding to each delay time of the tap coefficient sequence shown in FIG. Involving operations will be performed.

Problems to be Solved by the Invention However, in the above configuration, the maximum delay time that can be processed in the convolution operation is determined by the capacity of the data memory 26 and the sampling frequency, and the number of multiplications within one sampling time depends on the sampling period. The number divided by the machine cycle of the convolution unit 24 becomes the maximum value. Therefore, if the sampling frequency is set high for wideband reproduction, only a part of the impulse response (a time width of 3T in the figure) can be processed, as shown in Figure 5a, and even in that time width, there will be a sufficient number of taps. There was a problem that convolution was not possible.

In view of the above-mentioned problems, the present invention focuses on the fact that the higher the order of sound reflected from a space such as a hall, the more the high-frequency components are attenuated.
By executing at a lower sampling frequency than the /D section, it is possible to convolve a large number of impulse trains with a long response time using small-scale hardware.
It is an object of the present invention to provide a sound field reproducing device capable of reproducing a sound field with a highly realistic feeling.

Means for Solving the Problems In order to achieve the above object, the sound field reproducing apparatus of the present invention includes a first convolution means for convolving and processing a digital signal from an analog-to-digital converted music source; a low-pass filter downsampling means for low-pass filtering and lowering the sampling frequency after delaying the digital signal for a certain period of time; The structure includes a second convolution means for performing convolution processing.

Effect of the Invention The present invention has the above-mentioned configuration, and divides the impulse response of a listening space such as a hall to be realized into a plurality of time domains, and divides the impulse response included in the slow time domain and the digital signal of the music source convoluted with the impulse response included in the slow time domain. The signal is a low-pass filtered signal with a low cutoff frequency, and the sound field of a live or large-volume listening space with a long impulse response is created by performing multiple multiplications and additions at a low frequency/bring frequency. It becomes possible to reproduce the image faithfully.

EXAMPLES Hereinafter, examples of the present invention will be described with reference to the drawings.

FIG. 1 shows a block diagram of a sound field reproduction device in an embodiment of the present invention.

In FIG. 1, input terminal 1. A/D section 2. D/A section 2
2. The output terminal 23 is similar to that shown in FIG. 4 is a first circuit for storing the digital signal output from the A/D section 2;
a data memory 6; a first coefficient memory 3 storing tap coefficients to be convolved with the digital signal;
is the first convolution unit that performs these convolution processes.

6 is a first delay means for delaying the output signal of the A/D section 2; 7 is a first low-pass filter (I, PF) for low-pass filtering the output signal from the first delay means 6; , 8 is a first downsampling section for lowering the sampling frequency by thinning out the digital signal output from the first low-pass filter 7, and 10 stores the signal from the first downsampling section. 11 is a second coefficient memory in which tap coefficients to be convolved with the above signal are stored; 9 is a second convolution unit that performs these convolution processes; 14 is a first 16 is a second low-pass filter for low-pass filtering the output signal from the second delay means 14; 18 is a second downsampling section for lowering the sampling frequency by processing such as thinning out the digital signal output from the second low-pass filter 16; 3 data memory, 19 a third data memory in which tap coefficients to be convolved with the above signal are stored;
17 is a third convolution unit that performs these convolution processes; 20 is a coefficient memory for reducing the sampling frequency by processing such as zero value interpolation on the output signal from the third convolution unit 17; 21 is a second low-pass filter for smoothing the output signal from the second up-sampling unit 20; A filter with the same characteristics as the filter 16 is used. 12
is the second convolution unit 9 and the second low-pass filter 21
A second up-sampling section 13 increases the sampling frequency to the original value by performing processing such as zero value interpolation on the signal obtained as a result of addition of the zero rainfall force signal; 13 is the same as the first up-sampling section 12; A third low-pass filter for smoothing the output of is used, which has the same characteristics as the first low-pass filter 6. 22 is D/A
, converts the digital signal obtained as a result of the addition of the rainfall force signals of the first convolution section 3 and the third low-pass filter 13 into an analog signal, and outputs it from the output terminal 23 .

The operation of the sound field reproduction device of this embodiment configured as described above will be explained. The music signal input from the input terminal 1 is converted into a digital signal by the A/D section 2, and then
The data is stored in the first data memory 4 and subjected to convolution processing with the tap coefficients stored in the first coefficient memory 6 in the first convolution section 3 . FIG. 2 is a diagram showing the impulse response of a space such as a hall realized in this device, and the tap coefficients calculated in the first convolution unit 3 are in the time domain (T in the figure) (
Region 1), that is, the impulse train belonging to the earliest time region. Therefore, the capacity of the first data memory 4 is a size corresponding to the delay time T.

On the other hand, the digital signal output from the A/D section 2 is sent to a first delay means 6 having a delay time approximately equal to the delay time T.
After being delayed by 20 seconds, it is low-pass filtered by a first low-pass filter 7 so that the band is half that of the original signal, and then thinned out by a first down-sampling section 8 to reduce the original value. signal sampling frequency. This signal is transmitted to the second data memory 10
The second convolution unit 9 executes convolution processing with the tap coefficients stored in the second coefficient memory 11 at a sampling frequency of %. The tap coefficient calculated in the second convolution section 9 is calculated in the time domain (region 2
) is an impulse train belonging to Therefore, the capacity of the second data memory 10 is a size corresponding to the delay time 2T, but since the sampling frequency is %, the capacity of the second data memory 10 may be approximately the same as that of the first data memory 4.

The signal output from the first downsampling section 8 is
On the other hand, after being delayed by the second delay means 6 having a delay time approximately equal to the delay time 2T, the signal is low-pass filtered by the second low-pass filter 16 to have a band of % of the original signal. , and is further reduced to the original value Z sampling frequency through a thinning process by the second downsampling unit 16. This signal is stored in the third data memory 18, and convolved with the tap coefficients stored in the third coefficient memory 19 in the third convolution unit 17 at a sampling frequency of K. Ru. The tap coefficients calculated in the third convolution section 17 are impulse trains belonging to the time domain (region 3) indicated by 4T in FIG. therefore,
The capacity of the third data memory 18 is a size corresponding to the delay time 4T, but since the sampling frequency is K, the capacity may be approximately the same as that of the first data memory 4.

The output signal from the third convolution section 17 is passed to the second upsampling section 20 by interpolating zero values between the signals, thereby doubling the sampling frequency and converting the signal to % of the original value. It will be smoothed out. This smoothed signal is passed to the second convolution section 9
After being added to the signal output from the first up-sampling unit 12, the sampling frequency is doubled and returned to the original sampling frequency by interpolating zero values between the signals again. After the above signal is added with the output signal from the first convolution section 3, the D/A
The signal is converted into an analog signal by the section 22 and output from the output terminal 23.

Due to the above operation, the sampling period in the second convolution unit 9 and the third convolution unit 17 is expanded, so even if the hardware configuration is the same as that of the first convolution unit 3, the sampling period is doubled. .

It becomes possible to perform four times the number of multiplications and additions, and it becomes possible to easily realize the sound field, that is, the impulse response of a hall or the like shown in FIG. 2, without increasing the scale of the hardware. - Although there are three types of sampling frequencies in FIG. 1, there may be two types by omitting the path from the second delay means 14 to the second low-pass filter 21, or there may be four or more types. Also, the decrease in sampling frequency is %n (n=1. 2...
), but it is not limited to this.

FIG. 3 shows the first data memory 4 shown in FIG.
．． Second data memo! J10. Third data memory 1
8 is a diagram showing an example of a case where 8 is configured with one memory. In FIG. 3, the memory is divided into three parts, each with a data memory 1 (used as the first data memory 4) in ascending order of address.

Data memory 2 (used as second data memory 1o)
, data memory 3 (used as the third data memo January 8). Then, the data in data memory 1 is changed to the original sampling frequency, and the data in data memory 2. Data memory 3 is % each.

It supports 2 sampling frequencies. Further, the first delay means 6 can be easily realized using the data memory 1, and after reading out data near the final address of the data memory 1 and thinning out the data through an FIR or IIR type filter, the first delay means 6 can be easily realized using the data memory 1. By writing to the youngest address of , the processing from the first delay means 6 to the first downsampling section 8 in FIG. 1 can be easily performed. The processing from the second delay means 14 to the second downsampling section 16 can also be realized between the data memory 2 and the data memory 3 in the same manner as described above. Therefore, a delay time of T+2T+4T can be realized for the memory as a whole. In addition, if the hardware scale of the three convolution units 3 and 9.17 is the same, the data memory 1
It is possible to multiply and add N times from the data memory 2, 2N times from the data memory 3, and 4N times from the data memory 3, and a live spatial sound field with a long impulse response can be easily realized.

Effects of the Invention The present invention is realized by repeating the operation of lowering the sampling frequency after delaying and low-pass filtering the digital signal of a music source, and performing a convolution operation with a tap coefficient at each sampling frequency. The response time can be reduced by dividing the impulse response of a space, such as a hall, into a number of regions equal to the type of sampling frequency, and by processing the impulse train (tap coefficient) in the region with a long delay time at a lower sampling frequency. The objective is to realize an excellent sound field reproduction device that can easily perform long live performances or large-volume spatial sound fields with small-scale hardware.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a sound field reproducing device according to an embodiment of the present invention, FIG. 2 is an impulse response diagram of a space such as a hall to be realized, FIG. 3 is a state diagram showing the memory configuration, and FIG.
Figure 6 is a block diagram showing a conventional sound field reproduction device, Figure 6a is an impulse response diagram of a space such as a hall, Figure 5 is an impulse response diagram showing the compression result of the impulse response, and Figure 6 is a conventional memory. FIG. 3 is a state diagram showing the configuration. 1...Input terminal, 2...A/D section, 3
・・・・・・First convolution part, 4・・・・・・First
data memory, 6...first coefficient memory, 6
....First delay means, 7..First low-pass filter, 8..First downsampling unit, 9.. Second convolution. Including part, 10...
. . . second data memory, 11 . . . second coefficient memory, 12 . . . first upsampling unit, 13 . 14
...Second delay means, 16...Second low-pass filter, 16...Second downsampling section, 17...Third convolution Inclusion part, 1
8...Third data memory, 19...
Third coefficient memory, 2o...second upsampling section, 21...second low-pass filter, 22...D/A section, 23... ...Output terminal 1.24...Convolution part, 26...
・Data memory, 26...Coefficient memory. Name of agent: Patent attorney Toshio Nakao and 1 other person
2! I! J 4 Figure 2m Figure 5

Claims (3)

[Claims] (1) a first convolution means for convolving a digital signal from an analog-to-digital converted music source; and a second convolution means for convolving the output signal from the low pass filter downsampling means at a sampling frequency lower than that of the first convolution means. A distinctive sound field reproduction device. (2) By configuring multiple stages of low-pass filtering down-sampling means in series, the sampling frequency of the digital signal of the music source is sequentially lowered, and the output signal from each low-pass filtering down-sampling means is adjusted to the sampling frequency of each signal. 2. The sound field reproducing device according to claim 1, further comprising a plurality of convolution means for performing convolution processing. (3) The time axis of the impulse response of the sound field space to be realized is divided into multiple regions, and the first convolution unit executes convolution processing using the impulse group of the earliest region, and the delay time is 3. The sound field reproducing apparatus according to claim 1, wherein the convolution processing using a large time domain impulse group is performed by a convolution unit that processes at a lower sampling frequency.