Audio source localization

Patent metadata

Field	Value
Publication number	US-9025415-B2
Application number	US-201113521244-A
Country	US
Kind code	B2
Filing date	Feb 16, 2011
Priority date	Feb 23, 2010
Publication date	May 5, 2015
Grant date	May 5, 2015

Abstract

An audio source localization apparatus receives signals from a microphone array ( 101 ), and a reference processor ( 105 ) generates at least three reference beams with different directional properties. An estimation processor ( 107 ) which generates simultaneous direction estimates for two sound sources, comprises a circuit ( 401 ) combining signals of the at least three reference beams with a beam shape parameter reflecting a shape of an audio beamform and a beam direction parameter reflecting a direction of an audio beamform for the combined signal. A cost processor ( 403 ) generates a cost measure indicative of an energy of the combined signal and a minimization processor ( 405 ) estimates values of the beam shape parameter and the beam direction parameter which correspond to a local minimum of the cost measure. A direction processor ( 407 ) then determines simultaneous direction estimates for two sound sources from the determined parameter values. Improved direction estimation for two simultaneous sound sources may be achieved.

First claim

The invention claimed is:

1. An audio source localization apparatus comprising:
a receiving circuit for receiving microphone signals from an at least two-dimensional microphone array comprising at least three microphones;
a reference circuit for generating signals of at least three reference beams from the microphone signals, the three reference beams having different directional properties; and
an estimation circuit for generating a simultaneous direction estimate for two sound sources, the estimation circuit comprising:
a circuit for generating a combined signal by combining the signals of the at least three reference beams, the combined signal having a beam shape parameter reflecting a shape of an audio beamform for the combined signal and a beam direction parameter reflecting a direction of an audio beamform for the combined signal,
a circuit for generating a cost measure indicative of an energy measure of the combined signal,
a circuit for estimating a beam shape parameter value for the beam shape parameter and a beam direction parameter value for the beam direction parameter corresponding to a local minimum for the cost measure, and
a circuit for determining a first direction estimate of a first sound source and a second direction estimate for a second sound source as functions of the beam shape parameter value and the beam direction parameter value.

2. The audio source localization apparatus as claimed in claim 1 , wherein the estimation circuit is arranged to iteratively determine an update value for at least a first parameter of the beam shape parameter and the beam direction parameter in response to the combined signal, a current value of the beam shape parameter and a current value of the beam direction parameter, and to generate a new value for the first parameter from the update value and a current value of the first parameter.

3. The audio source localization apparatus as claimed in claim 2 , wherein the update value is dependent on a derivative value of the cost measure with respect to the first parameter.

4. The audio source localization apparatus as claimed in claim 2 , wherein the estimation circuit is arranged to independently determine update values for both the beam shape parameter and the beam direction parameter.

5. The audio source localization apparatus as claimed in claim 1 , wherein the estimation circuit is arranged to estimate the local minimum using a gradient search process for at least one of the beam shape parameter and the beam direction parameter.

6. The audio source localization apparatus as claimed in claim 1 , wherein the estimation circuit is arranged to limit the beam shape parameter to values for which a maximum weighting of a first reference beam of the at least three reference beams being smaller than a combined weighting of a second and third reference beam of the at least three reference beams.

7. The audio source localization apparatus as claimed in claim 1 , wherein the first direction estimate and the second direction estimate are two-dimensional direction estimates in a plane and the estimation circuit is arranged to compensate the first direction estimate and the second direction estimates for an elevation of the source outside the plane.

8. The audio source localization apparatus as claimed in claim 1 , wherein the microphone array is a three-dimensional microphone array comprising at least four microphones;
the reference circuit is arranged to further generate a signal of a fourth reference beam, and in the estimation circuit:
the circuit for generating the combined signal generating a further combined signal by combining signals of the fourth reference beam and at least two other reference beams of the at least three reference beams, the further combined signal having a further beam shape parameter reflecting a shape of an audio beamform for the further combined signal and a further beam direction parameter reflecting a direction of a further audio beamform for the further combined signal,
the circuit for generating the cost measure generating a further cost measure indicative of an energy measure of the further combined signal, and
the circuit for estimating the beam shape parameter value generating a further beam shape parameter value for the further beam shape parameter and a further beam direction parameter value for the beam direction parameter corresponding to a local minimum for the further cost measure; and
the circuit for determining the first direction estimate and the second direction estimate is arranged to further determine the first direction estimate and the second direction estimate in response to the further beam shape parameter value and the further beam direction parameter value.

9. The audio source localization apparatus as claimed in claim 8 , wherein:
the beam shape parameter value and the beam direction parameter value characterize a first set of three-dimensional points for which a derivative of the cost measure with respect to at least one of the beam shape parameter and the beam direction parameter is zero;
the further beam shape parameter value and the further beam direction parameter characterize a second set of three-dimensional points for which a derivative of the further cost measure with respect to at least one of the further beam shape parameter and the further beam direction parameter is zero; and
the estimation circuit is arranged to determine the first direction estimate and the second direction estimate as direction estimates for at least two three-dimensional points comprised in both the first set of three-dimensional points and the second set of three-dimensional points.

10. The audio source localization apparatus as claimed in claim 9 , wherein the estimation circuit is arranged to select the at least two three-dimensional points in response to a requirement that the at least two three-dimensional points correspond to a predetermined gain value.

11. The audio source localization apparatus as claimed in claim 1 , wherein the estimation circuit is arranged to determine the first direction estimate and the second direction estimate as direction estimates which for the beam shape parameter value and the beam direction parameter value for which an analytical derivative of the cost function with respect to at least one of the beam shape parameter and the beam direction parameter is zero.

12. The audio source localization apparatus as claimed in claim 1 , wherein a first reference beam of the at least three reference beams is a monopole and at least a second reference beam and a third reference beam of the at least three reference beams are differently directed dipoles.

13. The audio source localization apparatus as claimed in claim 11 , wherein the beam direction parameter is indicative of a weighting of the second reference beam relative to the third reference beam, and the beam shape parameter is indicative of a weighting of the first reference beam relative to the second and third reference beams.

14. The audio source localization apparatus as claimed in claim 1 , wherein the combination is substantially given by:
y[k]=α m[k ]+(1−α) [cos(φ s ) d x [k ]+sin(φ s ) d y [k ]]

where y[k] represents sample k of the combined signal, m[k] represents sample k of the first reference beam, d x [k] represents sample k of the second reference signal, d y [k] represents sample k of the third reference signal, αis the beam shape parameter and φ s is the beam direction parameter.

15. A method of audio source localization comprising:
receiving microphone signals from an at least two-dimensional microphone array compri

Assignees

• Derkx Rene Martinus Maria
• Koninkl Philips Nv

Inventors

• Derkx Rene Martinus Maria

Classifications

• H04R3/005

  for combining the signals of two or more microphones (specially adapted for hearing aids H04R25/407) · CPC title

• G01S3/8036 Primary

  derived directly from separate directional systems · CPC title

• G01S3/803 Primary

  using amplitude comparison of signals derived from receiving transducers or transducer systems having differently-oriented directivity characteristics · CPC title