Abstract
The generalized spectral subtraction algorithm (GBSS), which has extraordinary ability in background noise reduction, is historically one of the first approaches used for speech enhancement and dereverberation. However, the algorithm has not been applied to de-noise the room impulse response (RIR) to extend the reverberation decay range. The application of the GBSS algorithm in this study is stated as an optimization problem, that is, subtracting the noise level from the RIR while maintaining the signal quality. The optimization process conducted in the measurements of the RIRs with artificial noise and natural ambient noise aims to determine the optimal sets of factors to achieve the best noise reduction results regarding the largest dynamic range improvement. The optimal factors are set variables determined by the estimated SNRs of the RIRs filtered in the octave band. The acoustic parameters, the reverberation time (RT), and early decay time (EDT), and the dynamic range improvement of the energy decay curve were used as control measures and evaluation criteria to ensure the reliability of the algorithm. The de-noising results were compared with noise compensation methods. With the achieved optimal factors, the GBSS contributes to a significant effect in terms of dynamic range improvement and decreases the estimation errors in the RTs caused by noise levels.
Highlights
A well-known and widely used method to calculate Reverberation time (RT) is determined by the energy decay curve (EDC) generated by Schroeder’s method [6]; the measured room impulse response (RIR) presents ambient noise, and equipment noise may deteriorate the EDC, leading to errors in predicting room acoustic parameters [7,8]
The values of 0.1 dB and 0.05 dB were set to calculate the dynamic ranges of the reference EDCs and the de-noised EDCs with different SNRs to ensure the computed errors of the parameters early decay time (EDT) and the RTs were below 1%
When the noise levels were higher than −40 dB, leading to similar poor results in the mean dynamic range improvement, approximately 5 dB, most of the SNRs estimated in the octave band were lower than 10
Summary
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Many algorithms have been developed to minimize the distortion caused by noise subtraction processing between an enhanced signal and a clean signal [27] Among these methods, Berouti proposed a generalized spectral subtraction based on short-time spectral amplitudes, called Berouti’s GSS (GBSS), which was implemented to de-noise the RIRs because of its computational simplicity and the flexibility of parametric adjustments in the technique for producing significant de-noising improvements [28,29,30,31,32,33,34]. Subtracting a constant ratio of noise levels over the entire frequency of the RIR may remove parts of the clean signal and deteriorate the estimation of the RTs. it is essential to have prior knowledge of the motivation of the parameters of the GBSS algorithm to determine the optimal sets of factors to mitigate unwanted noise effects and improve the dynamic range, without or with an acceptable minimal degradation of the reverberation decay. The GBSS algorithm with optimal factors significantly improved the dynamic range and decreased the estimation errors in RTs caused by noise levels
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
