An active noise reduction system and method

Abstract

The invention relates to an active noise reduction system and method. An audio collector can collect a section of ambient sound audio and a section of noise sampling audio corresponding to the ambientsound audio. A first-stage FIR filter obtains a section of first-stage filtering audio data according to the first filtering coefficient sequence and a set window function filtering ambient sound audio digital quantity. A second-stage FIR filter splices multiple sections of first-stage filtering audio data into a group of audios according to the storage sequence of storage units. The multiple sections of audio in one group of audio are filtered according to the second filtering coefficient of each section of first filtering audio data to obtain second-stage filtering audio data. The active noise reduction system and method have the advantages that the work of IIR filter coefficient adjustment is omitted, fitting of characteristics of the frequency with 1 kHz or below is good, the fittingeffect on characteristics of the frequency with 1 kHz or above is good, and the noise reduction frequency range of random noise is widened.

Description

technical field [0001] The invention relates to the field of noise processing, in particular to an active noise reduction system and method. Background technique [0002] Active noise cancellation (Active noise cancellation, ANC) is that the earphone emits a sound wave with the same amplitude and phase opposite to the ambient noise through the speaker inside. The two sound waves are superimposed and cancel each other out, reducing the noise intensity heard by the wearer. [0003] The noise reduction filter is actually a fitting of the frequency characteristics of the acoustic system composed of microphones, speakers, etc. The more accurate the fitting, the better the noise reduction effect. Since the frequency characteristics of the above-mentioned acoustic system are not necessarily smooth curves, the noise reduction It is best to use the FIR structure for the filter. The FIR structure is better for non-smooth curve fitting. In order to obtain a good noise reduction effect,...

AI Technical Summary

Problems solved by technology

[0003] The noise reduction filter is actually a fitting of the frequency characteristics of the acoustic system composed of microphones, speakers, etc. The more accurate the fitting, the better the noise reduction effect. Since the frequency characteristics of the above-mentioned acoustic system are not necessarily smooth curves, the noise reduction It is best to use the FIR structure for the filter. The FIR structure is better for non-smooth curve fitting. In order to obtain a good noise reduction effect, a high order of the FIR filter is generally required, which makes it very difficult to implement in practice. For example, sampling When the frequency is 192kHz, using a multiply-accumulate unit to complete the 4095-order FIR convolution operation requires a digital logic operating clock frequency of up to 786MHz, and needs to store 4096 filter coefficients, which requires high power consumption and complexity for digital logic

[0004] The IIR filter fits the acoustic system with smooth frequency characteristics better, but the actual acoustic system will have singularities, which makes it difficult for the IIR filter to fit this non-smooth frequency characteristic curve well, and the noise reduction effect is not good. On the one hand, the fitting effect of the IIR filter to the frequency characteristics above 1kHz is poor, and repeated tests are required to obtain a good set of coefficients, which makes it difficult to adjust the coefficients of the IIR filter

Images