Objective and subjective evaluation of noise reduction algorithms for hearing aids

Vijay Parsa, Gurjit Singh, Guo Chen, Karthikeyan Umapathy


A study of the subjective and objective evaluation of noise reduction algorithms for hearing aids is presented. The quality of speech processed through six different noise reduction algorithms with normal hearing and hearing impaired listeners is evaluated. The instrumental measures of noise reduction performance that correlate best with the behavioural data is identified. The sound quality ratings are obtained using two sentences from the Hearing In Noise Test (HINT) database which are corrupted by either speech-shaped noise (SSN) or multi-talker babble (MTB) at SNRs ranging between -4dB to +12 dB.


Algorithms; Database systems; Functions; Noise abatement; Signal to noise ratio; Speech processing; Hearing In Noise Test (HINT); Hearing-impairment; Power spectrum estimators; Speech-shaped noise (SSN)

Full Text:



  • There are currently no refbacks.