Audio examples for paper "A Compact Noise Covariance Matrix Model for MVDR Beamforming"

List of Acronyms:

• SIR: Signal-to-Interference Ratio
• SBNR: Signal-to-Babble Noise Ratio
• SSNR: Signal-to-Sensor Noise Ratio
• RIR: Room Impulse Response

Room-Array Configurations:

• Room Dimension: 8.3 x 6.5 x 2.9 m
• Array Type: 32-element rigid Spherical Microphone Array (em32 Eigenmike®, 4.2cm radius)
• Array Location: (4.65 , 3.25 , 1.5) m

Datasets:

• Raw Speeches: The Open Speech Repository. [Online]
• Measured RIR Dataset: Orhun Olgun, & Huseyin Hacihabiboglu. (2019). METU SPARG Eigenmike em32 Acoustic Impulse Response Dataset v0.1.0 (Version 0.1.0) [Dataset]. Zenodo. [Online]
• Simulated RIR: D. P. Jarrett, “Spherical Microphone array Impulse Response (SMIR) generator.” [Online]

A. Varying separation & noise (based on measured RIR)

T₆₀:1.12 s, C₅₀:5.4~2.3 dB, source-array range:1.0~1.4 m, babble-array range:1.6 m, SSNR:40 dB

Separation: 27°, SIR: -5 dB, SBNR: 10 dB

Separation: 27°, SIR: 0 dB, SBNR: 20 dB

Separation: 45°, SIR: -10 dB, SBNR: 20 dB

B. Varying reverberation (based on simulated RIR)

Separation: 15°, SIR: -5 dB, SBNR: 15 dB, SSNR: 40 dB, source-array range: 0.5 m, babble-array range: 1.6 m

T₆₀: 0.2 s, C₅₀: 27.5 dB

T₆₀: 0.5 s, C₅₀: 9.0 dB

T₆₀: 1.0 s, C₅₀: 2.0 dB