Reverberation time estimation from speech signals based on blind room impulse response identification

2015 ◽  
Vol 138 (2) ◽  
pp. 731-734 ◽  
Author(s):  
Lifu Wu ◽  
Xiaojun Qiu ◽  
Ian Burnett ◽  
Yecai Guo
Keyword(s):  
Impulse Response ◽  
Time Estimation ◽  
Speech Signals ◽  
Reverberation Time

A blind algorithm for reverberation-time estimation using subband decomposition of speech signals

2012 ◽  
Vol 131 (4) ◽  
pp. 2811-2816 ◽  
Author(s):  
Thiago de M. Prego ◽  
Amaro A. de Lima ◽  
Sergio L. Netto ◽  
Bowon Lee ◽  
Amir Said ◽  
...  
Keyword(s):  
Time Estimation ◽  
Speech Signals ◽  
Subband Decomposition ◽  
Reverberation Time

Toward blind reverberation time estimation for non-speech signals

2013 ◽  
Vol 133 (5) ◽  
pp. 3393-3393
Author(s):  
João F. Santos ◽  
Nils Peters ◽  
Tiago H. Falk
Keyword(s):  
Time Estimation ◽  
Speech Signals ◽  
Reverberation Time

scholarly journals Towards blind reverberation time estimation for non-speech signals

2013 ◽  
Author(s):  
Joao F. Santos ◽  
Nils Peters ◽  
Tiago H. Falk
Keyword(s):  
Time Estimation ◽  
Speech Signals ◽  
Reverberation Time

scholarly journals Effect of phoneme variations on blind reverberation time estimation

Acta Acustica ◽  
2020 ◽  
Vol 4 (1) ◽  
pp. 3
Author(s):  
Andrea Andrijašević
Keyword(s):  
Time Estimation ◽  
Multiple Comparison ◽  
Reverberation Time ◽  
Impulse Responses ◽  
Estimation Errors ◽  
Significance Level ◽  
Frequency Bands ◽  
Interaction Terms ◽  
Multiple Comparison Procedure

This study focuses on an unexplored aspect of the performance of algorithms for blind reverberation time (T) estimation – on the effect that speech signal’s phonetic content has on the value of the estimate of T that is obtained from the reverberant version of that signal. To this end, the performance of three algorithms is assessed on a set of logatome recordings artificially reverberated with room impulse responses from four rooms, with their T20 value in the [0.18, 0.55] s interval. Analyses of variance showed that the null hypotheses of equal means of estimation errors can be rejected at the significance level of 0.05 for the interaction terms between the factors “vowel”, “consonant”, and “room”, while the results of Tukey’s multiple comparison procedure revealed that there are both some similarities in the behaviour of the algorithms and some differences, where the latter are stemming from the differences in the details of algorithms’ implementation such as the number of frequency bands and whether T is estimated continuously or only on the selected, the so-called speech decay, segments of the signal.

Quantifying Non-Linearity in Early Decay Curves of Measured and Computer-Modeled Room Impulse Responses of a Highly Non-Diffuse Room Exhibiting Flutter Echo

2020 ◽  
Author(s):  
Heather L. Lai ◽  
Brian Hamilton
Keyword(s):  
Linear Regression ◽  
Computer Modeling ◽  
Impulse Response ◽  
Decay Curve ◽  
Sound Field ◽  
Energy Decay ◽  
Room Acoustics ◽  
Reverberation Time ◽  
Sound Fields ◽  
Regression Curves

Abstract This paper investigates the use of two room acoustics metrics designed to evaluate the degree to which the linearity assumptions of the energy density curves are valid. The study focuses on measured and computer-modeled energy density curves derived from the room impulse response of a space exhibiting a highly non-diffuse sound field due to flutter echo. In conjunction with acoustical remediation, room impulse response measurements were taken before and after the installation of the acoustical panels. A very dramatic decrease in the reverberation time was experienced due to the addition of the acoustical panels. The two non-linearity metrics used in this study are the non-linearity parameter and the curvature. These metrics are calculated from the energy decay curves computed per octave band, based on the definitions presented in ISO 3382-2. The non-linearity parameter quantifies the deviation of the EDC from a straight line fit used to generated T20 and T30 reverberation times. Where the reverberation times are calculated based on a linear regression of the data relating to either −5 to −25 dB for T20 or −5 to −35 dB for T30 reverberation time calculations. This deviation is quantified using the correlation coefficient between the energy decay curve and the linear regression for the specified data. In order to graphically demonstrate these non-linearity metrics, the energy decay curves are plotted along with the linear regression curves for the T20 and T30 reverberation time for both the measured data and two different room acoustics computer-modeling techniques, geometric acoustics modeling and finite-difference wave-based modeling. The intent of plotting these curves together is to demonstrate the relationship between these metrics and the energy decay curve, and to evaluate their use for quantifying degree of non-linearity in non-diffuse sound fields. Observations of these graphical representations are used to evaluate the accuracy of reverberation time estimations in non-diffuse environments, and to evaluate the use of these non-linearity parameters for comparison of different computer-modeling techniques or room configurations. Using these techniques, the non-linearity parameter based on both T20 and T30 linear regression curves and the curvature parameter were calculated over 250–4000 Hz octave bands for the measured and computer-modeled room impulse response curves at two different locations and two different room configurations. Observations of these calculated results are used to evaluate the consistency of these metrics, and the application of these metrics to quantifying the degree of non-linearity of the energy decay curve derived from a non-diffuse sound field. These calculated values are also used to evaluate the differences in the degree of diffusivity between the measured and computer-modeled room impulse response. Acoustical computer modeling is often based on geometrical acoustics using ray-tracing and image-source algorithms, however, in non-diffuse sound fields, wave based methods are often able to better model the characteristic sound wave patterns that are developed. It is of interest to study whether these improvements in the wave based computer-modeling are also reflected in the non-linearity parameter calculations. The results showed that these metrics provide an effective criteria for identifying non-linearity in the energy decay curve, however for highly non-diffuse sound fields, the resulting values were found to be very sensitive to fluctuations in the energy decay curves and therefore, contain inconsistencies due to these differences.

Speech-Model Based Accurate Blind Reverberation Time Estimation Using an LPC Filter

2012 ◽  
Vol 20 (6) ◽  
pp. 1884-1893 ◽  
Author(s):  
Abbas Keshavarz ◽  
Saeed Mosayyebpour ◽  
Mehrzad Biguesh ◽  
T. Aaron Gulliver ◽  
Morteza Esmaeili
Keyword(s):  
Time Estimation ◽  
Reverberation Time ◽  
Model Based ◽  
Speech Model
Sign in / Sign up

Export Citation Format

Share Document