KR19990043757A - Structure of Microphone Array Beam Former Using Bandwidth Splitter - Google Patents

Abstract

The present invention relates to a structure of a microphone array beam former using a band division processor.

Array microphones are used to sense high sensitivity sound signals from moving sources in a noisy environment. However, in order to maintain the sensitivity, the microphone should be installed over the moving area of the sound source, and it is impossible to extract the signal of only the sound source moving or stopping in a noisy environment.

Since the speech signal is a wide band signal, theoretically, it is impossible to construct a beamformer. However, in the present invention, after narrowing the wideband signal by a band splitting processor, a simple beamformer is implemented using a simple time delay compensator. A structure of a microphone array beam former using a band splitting processor capable of optimally maintaining a signal-to-noise ratio of a sensed acoustic signal is presented.

Description

Structure of Microphone Array Beam Former Using Bandwidth Splitter

The present invention relates to a structure of a microphone array beam former using a band splitting processor, and in particular, in a noisy environment, a microphone beam can be formed in a specific sound source or automatically tracked when a sound source is moved to maintain a signal-to-noise ratio above a specified value. The present invention relates to a structure of a microphone array beam former using a band splitting processor.

Array microphones can be used to detect acoustic signals with high sensitivity from moving sources in a noisy environment. However, in order to maintain the sensitivity from the moving sound source, the microphone must be installed over the moving area of the sound source, and it is impossible to extract a signal of only the sound source moving or stopping.

The adaptive microphone array beamformer (hereinafter referred to as AMAB) is a device that can remove a noise or automatically track a sound source when the sound source is moved by forming a beam of a microphone in a specific sound source in a noise environment. This adaptive beamformer can be operated in either an automatic mode that automatically directs the microphone's beam in the sound source direction or in a manual mode that directs the beam in a specific direction by PC control. In the manual mode, single or multiple beams can be operated. Can be formed.

Acquiring only signals from a particular sound source in a noisy environment can degrade the signal-to-noise ratio due to background noise, and can be susceptible to ambient noise even when the sound source is far from the microphone. In addition, when configuring a conference environment, it is troublesome to install a microphone for each panelist on the table. Steering the microphone according to the direction of the camera in the sports field can collect lively sound signals. In these acoustic environments, the use of an adaptive microphone beamformer can improve the noise-to-interference ratio and collect the desired acoustic signal with the remote microphone function even under low noise environments. However, since the voice signal is a very wideband signal, there is a problem in that it is impossible to apply a beam former for voice signal detection.

Therefore, the present invention improves the noise-to-interference ratio by controlling the beam formation by the time delay compensator after narrowband processing the wideband sound signal sensed by the array microphones using a band splitter, and the remote microphone function is desired even in a low noise environment. It is an object of the present invention to provide a structure of a microphone array beam former using a band splitting processor capable of collecting sound signals.

A structure of a microphone array beam former using a band splitting processor according to the present invention for achieving the above object includes a first low noise amplifier for reducing noise of a received signal input to a microphone, and a microphone amplified by the first low noise amplifier. An analog-to-digital converter for digitizing an input signal, a finite impulse response filter for narrowbanding the digitized wideband microphone signal, and for storing an input signal of a microphone transformed into a narrowband signal by the finite impulse response filter A time delay compensator for reading the data stored in the memory according to the delay difference using a memory, delay difference information according to the distance between each microphone and the sound source, and data stored in the memory read out using the time delay compensator. To form a beam A first adder, a gain adjuster for equalizing the entire composite signal between the auxiliary arrays with respect to the output of the beamed auxiliary array microphones and preventing distortion of the synthesized signal, and a second adder for synthesizing the outputs of the auxiliary array microphones; And a digital to analog converter for analogizing the synthesized signal and removing noise, and a second low noise amplifier.

1 is a structural diagram of an active microphone array beam former according to the present invention;

2 is a graph for explaining a wideband microphone signal divided into a narrowband signal by a digital filter combination;

3 is a schematic diagram of a microphone array;

<Description of the symbols for the main parts of the drawings>

101, 102, 115: low noise amplifier

103, 104: Analog-to-digital converter

114: Digital to Analog Converters

105, 106: FIR filter

107, 108: memory

109: Gain Adjustment and Time Delay Compensator

Hereinafter, with reference to the accompanying drawings will be described in detail the present invention.

1 is a structural diagram of an active microphone array beam former according to the present invention.

The input of the microphone is amplified by first low noise amplifiers (hereinafter referred to as LNAs) 101 and 102 and then digitized by analog-to-digital converters (ADCs) 103 and 104. The digitized wideband microphone signal is narrowbanded by finite impulse response (FIR) filters 105 and 106.

FIG. 2 is a graph for explaining a wideband microphone signal divided into a narrowband signal by the combination of the FIR filters 105 and 106, in which a digitized wideband microphone signal having a frequency of several tens of Hz to several KHz is used for the FIR filter 105 and 106 shows that the signal has been divided into narrowband signals by combination. For example, to perform band split processing on an input signal with a spectrum from 90 Hz to 4800 Hz, the first FIR filter has 120 Hz as the center frequency, 90 Hz and 150 Hz as the cutoff frequency, and the eighth digital filter It has a center frequency of 3840 Hz and a cutoff frequency of 2880 Hz and 4800 Hz. The bandwidth and processing bandwidth of each filter for band division processing are design variables, and the number of filters N is determined accordingly. The microphone array consists of N auxiliary microphone arrays, and the distance between single microphones in each auxiliary microphone array is shown in FIG. 3.

3 is a configuration diagram of the microphone array, and it can be seen that the microphones are arranged at half of the wavelength corresponding to the center frequency of the FIR filter to prevent aliasing.

The input signal of the microphone, which has been converted into a narrowband signal by the FIR filters 105 and 106, is stored in the memories 107 and 108. The time delay compensator 109 having delay difference information according to the distance between the sound source and each microphone synthesizes the data according to the delay difference from the memories 107 and 108 through the first adder 110. Therefore, the beam can be formed in a desired area. Here, a tracking algorithm is required to automatically track the beam formation as the sound source moves.

The outputs of the array microphones that form the beam in this way are combined in the gain adjuster 109 and the mixers 111 and 112 to prevent the distortion of the synthesized signal, so as to equalize the composite signal power between the auxiliary arrays, and then the second It is synthesized in the adder 113. The synthesized signal then passes through a digital to analog converter 114 and a second low noise amplifier 115.

As described in order to implement an active beam former having such a structure, an analog / digital converter, a digital / analog converter, and a FIR filter are required. Since a voice signal is usually limited to 4 KHz or less, high-speed digital signal processing (Digital Signal Processing) ; DSP or Application Specific Integrated Circuit (ASIC) can be easily implemented. In addition, the microphone beam can be manually or automatically formed and a beam tracking algorithm is required when in automatic mode, and all the controls required for operation can be implemented with a PC or a dedicated processor.

As described above, according to the present invention, the noise-to-interference ratio can be improved by collecting the signal from a particular sound source in a background noise environment or by using an adaptive microphone beam former when the sound source is far from the microphone, even in a low noise environment. The remote microphone function allows you to collect the desired acoustic signal. In addition, the adaptive micro beamformer forms the microphone beam by the delay compensator and the band division processor, so its structure is simple, so that the filter, time delay compensation and gain adjuster can be easily implemented by the general-purpose digital processor, There is an excellent effect that the expansion and change can be made flexibly.

Claims (1)

A first low noise amplifier for reducing noise of a received signal input to the microphone, An analog / digital converter for digitizing a microphone input signal amplified by the low noise amplifier, A finite impulse response filter for narrowbanding the digitized wideband microphone signal; A memory for storing an input signal of a microphone converted into a narrowband signal by the finite impulse response filter; A time delay compensator for reading data stored in the memory according to a delay difference by using delay difference information according to a distance between each microphone and a sound source; A first adder for synthesizing the data stored in the memory read using the time delay compensator to form a beam; A gain adjuster for equalizing the entire synthesized signal between the auxiliary arrays with respect to the output of the beamed auxiliary array microphones and preventing distortion of the synthesized signal; A second adder for synthesizing the output of the auxiliary array microphone, And a second low noise amplifier and a digital to analog converter for analogizing and synthesizing the synthesized signal.