An improved eigenvalue background noise reduction method for acoustic beamforming

•A modification of an existing eigenvalue background noise reduction method is proposed.•The new method is compared to conventional techniques on several numerical and experimental datasets.•The new method drastically improves the resolution on beamforming maps with an accurate level estimation. This paper presents a modification to an existing eigenvalue identification and subtraction method [Bahr. &Horne, Advanced background subtraction applied to aeroacoustic wind tunnel testing, 21st AIAA/CEAS Aeroacoustics Conference 2015] to locate sound sources in noisy environments. That novel method is compared with two others: classical background noise removal and the original eigenvalue identification and subtraction method. In addition, an improved version of the proposed method is proposed to obtain better level estimate. Four test cases are proposed, involving numerical and experimental studies. Overall, it is found that the new method provides a very accurate location when compared with other techniques, and levels are comparable with those obtained with classical methods. This was especially found in cases where the background noise is very high compared with the source level, with signal-to-noise ratio down to −60 dB.