Multiple source direction of arrival estimation using subspace pseudointensity vectors
This work addresses sound source localization for audio processing applications, but it is incremental as it applies an existing method to new data.
The paper tackled direction of arrival estimation for single and multiple sources using the subspace pseudointensity method on Eigenmike recordings from the LOCATA Challenge, achieving an average absolute error angle of about 9 degrees.
The recently proposed subspace pseudointensity method for direction of arrival estimation is applied in the context of Tasks 1 and 2 of the LOCATA Challenge using the Eigenmike recordings. Specific implementation details are described and results reported for the development dataset, for which the ground truth source directions are available. For both single and multiple source scenarios, the average absolute error angle is about 9 degrees.