Motor noise reduction circuit
View Patent ↗A method of reducing noise in an environment where the noise source is in a fixed location relative to a pair of microphones, such as in a camera with a zoom motor, involves receiving signals x 1 (t), x 2 (t) from the respective microphones, and filtering each of the signals x 1 (t), x 2 (t) with respective first and second linear filters having filter coefficients obtained by computing eigenfilters corresponding to data samples from the respective microphones for noise only and signal only conditions.
1. A method of reducing noise in an environment where the noise source is in a fixed location relative to a pair of microphones, comprising:
receiving signals x 1 (t), x 2 (t) from the respective microphones; and
filtering each of the signals x 1 (t), x 2 (t) with respective first and second linear filters having filter coefficients obtained by:
computing the cross-covariance matrices A and B of data sets from the respective microphones for noise only and signal only conditions,
computing the generalized eigenvectors and eigenvalues of the matrices A and B,
finding a specific eigenvector corresponding to the maximum eigenvalue of said generalized eigenvalues, and
calculating the filter coefficients from said specific eigenvector; and
outputting a noise-reduced filtered output signal to an audio processor.
2. A method as claimed in claim 1 , wherein the linear filters are Finite Impulse response filters.
3. A method as claimed in claim 1 , wherein the filtered signals are summed and passed through an inverse linear filter.
4. A method as claimed in claim 1 , wherein the coefficients of the first linear filter are the first half values of said specific eigenvector and the coefficients of the second linear filter are the second half values of said specific eigenvector.
5. A method as claimed in claim 4 , wherein the filtered signals are summed and passed through an inverse linear filter and the inverse linear filter is the inverse of said specific eigenvector.
6. A method as claimed in claim 1 , wherein the noise is zoom motor noise in a video camera.
7. A noise reduction circuit comprising:
first and second linear filters for filtering input signals x1(t), x2(t) from respective microphones subject to a noise source that is in a fixed location relative to a pair of microphones; and
an audio processor configured to receive noise-reduced filtered signals from said first and second linear filters; and
and wherein the first and second linear filters have filter coefficients obtained by:
computing the cross-covariance matrices A and B of data sets from the respective microphones for noise only and signal only conditions,
computing the generalized eigenvectors and eigenvalues of the matrices A and B,
finding a specific eigenvector corresponding to the maximum eigenvalue,
and calculating the filter coefficients from said specific eigenvector.
8. A noise reduction circuit as claimed in claim 7 , wherein the linear filters are Finite Impulse response filters.
9. A noise reduction circuit as claimed in claim 7 , further comprising a summer for summing the filtered signals, an inverse linear filter for inverse filtering the summed filtered signals to provide a noise reduced output signal.
10. A noise reduction circuit as claimed in claim 7 , wherein the coefficients of the first linear filter are the first half values of said specific eigenvector and the coefficients of the second linear filter are the second half values of said specific eigenvector.
11. A noise reduction circuit as claimed in claim 10 , further comprising a summer for summing the filtered signals, an inverse linear filter for inverse filtering the summed filtered signals to provide a noise reduced output signal and wherein the inverse linear filter is the inverse of said specific eigenvector.
12. A camera including noise reduction circuit as claimed in claim 7 to reduce motor noise.