A probabilistic approach for cross-spectral matrix denoising: Benchmarking with some recent methods.

Area of Science:

• Acoustics
• Signal Processing
• Fluid Dynamics

Background:

• Acoustic array measurements are susceptible to noise, particularly from fluid flow, hindering accurate data analysis and source localization.
• Effective denoising is essential for reliable interpretation of acoustic data.

Purpose of the Study:

• To propose and evaluate a novel denoising method for acoustic array measurements contaminated by flow-induced noise.
• To compare the proposed method with existing techniques for cross-spectral matrix (CSM) denoising.

Main Methods:

• A Probabilistic Factor Analysis (PFA) approach is presented, decomposing the measured CSM.
• The PFA method leverages the correlation structure of the acoustic field and flow noise.
• Performance is evaluated using numerical simulations and wind tunnel experiments.

Main Results:

• The PFA method demonstrates effective denoising of acoustic array measurements.
• Comparison with three existing methods shows the advantages of the proposed PFA approach.
• The method successfully isolates acoustic signals from flow noise without requiring reference measurements.

Conclusions:

• Probabilistic Factor Analysis offers a robust solution for denoising acoustic measurements in flow environments.
• The proposed method enhances the quality of acoustic data for subsequent analysis and source imaging.
• This technique is valuable for applications requiring precise acoustic measurements near fluid flows.