Binaural pre-processing for contralateral sound field attenuation and improved speech-in-noise recognition

Understanding speech presented in competition with other sounds can be challenging. Here, we reason that in free-field settings, this task can be facilitated by attenuating the sound field contralateral to each ear and propose to achieve this by linear subtraction of the weighted contralateral stimu...

Descripción completa

Detalles Bibliográficos
Autores: López Poveda, Enrique A., Eustaquio Martín, María Almudena, Martín San Victoriano, Fernando
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2022
País:España
Institución:Universidad de Salamanca (USAL)
Repositorio:GREDOS. Repositorio Institucional de la Universidad de Salamanca
OAI Identifier:oai:gredos.usal.es:10366/154683
Acceso en línea:http://hdl.handle.net/10366/154683
Access Level:acceso abierto
Palabra clave:Beamformer
Binaural hearing
Binaural unmasking
Hearing device
Noise reduction
Hearing
Noise
2406.01 Bioacústica
2490 Neurociencias
Descripción
Sumario:Understanding speech presented in competition with other sounds can be challenging. Here, we reason that in free-field settings, this task can be facilitated by attenuating the sound field contralateral to each ear and propose to achieve this by linear subtraction of the weighted contralateral stimulus. We mathematically justify setting the weight equal to the ratio of ipsilateral to contralateral head-related transfer functions (HRTFs) averaged over an appropriate azimuth range. The algorithm is implemented in the frequency domain and evaluated technically and experimentally for normal-hearing listeners in simulated free-field conditions. Results show that (1) it can substantially improve the signal-to-noise ratio (up to 30 dB) and the short-term objective intelligibility in the ear ipsilateral to the target source, particularly for maskers with speech-like spectra; (2) it can improve speech reception thresholds (SRTs) for sentences in competition with speech-shaped noise by up to 8.5 dB in bilateral listening and 10.0 dB in unilateral listening; (3) for sentences in competition with speech maskers and in bilateral listening, it can improve SRTs by 2 to 5 dB, depending on the number and location of the masker sources; (4) it hardly affects virtual sound-source lateralization; and (5) the improvements, and the algorithm's directivity pattern depend on the azimuth range used to calculate the weights. Contralateral HRTF-weighted subtraction may prove valuable for users of binaural hearing devices.