Multi microphone Signal Processing Technique to Remove Room Reverberation with Unbiased Coherent-To-Diffuse Power Ratio Estimation

Abstract

In everyday humans encounter different types of speech communication applications, such as hands-free mobile telephony, voice-controlled systems, hearing aids etc, the recorded microphone signals are corrupted by background noise, room reverberation and far-end echo signals. The signal degradation can lead to total unintelligibility of the original speech signal and degrades the performance of automatic speech recognition systems. In this project the estimation of frequency and time-dependent coherent-to-diffuse power ratio (CDR) from the measured spatial coherence between two omni directional microphones were investigated. From traditional methods for unbiased CDR estimators it is shown that knowledge of either direction of arrival (DOA) of the target source or the coherence of noise field is sufficient for unbiased CDR estimation. A CDR-based dereverberation system is presented and evaluated using signal-based quality measures as well as automatic speech recognition accuracy.

Authors and Affiliations

Nikhil S

Keywords

Related Articles

An Attribute Based Secure Data Self-Destructing and Time Server

Recent years all are using the cloud server. The reason of cloud server easy to use and having the security. Once we upload the file store to the cloud server then that files access the world wide and any time. Here acc...

Vapor−Liquid Equilibrium Data for the Binary System of Ethylene Glycol with Butanol At 94, 50, 10 KPA

Vapor−liquid equilibrium (VLE) data for Ethylene Glycol with Butanol are measured at pressure of 94, 50, 10 kPa by using a dynamic recirculation still. The experimental VLE data were tested to be thermodynamically consi...

Measurement of Radiation Abutment Dose for 6MV Photon and 9 MeV Electron Beam Combinations by GAFCHROMIC EBT3 Film at Extended SSD

To measure the radiation abutment dose for 6 mega voltage (MV) photon and 9 mega electron volt (MeV) electron beam combinations by gafchromic EBT3 film at extended source to skin distance (SSD). A point dose measurement...

Multi Core Processor Arrays Cores Optimization in AES Engines

In the present scenario protecting the sensitive information is done by several encryption standards. AES encryption is modern technique which is based on the principle of substitution-permutation network. Advanced encr...

Agent Based QoS Routing using DSR in Mobile Ad hoc Network

This paper modifies the existing DSR routing protocol by using the neuro-fuzzy agent. The proposed system uses the neural as well as the fuzzy agents to improve the QOS of DSR routing protocol. Neural and fuzzy agents a...

Download PDF file
  • EP ID EP22783
  • DOI -
  • Views 223
  • Downloads 3

How To Cite

Nikhil S (2016). Multi microphone Signal Processing Technique to Remove Room Reverberation with Unbiased Coherent-To-Diffuse Power Ratio Estimation. International Journal for Research in Applied Science and Engineering Technology (IJRASET), 4(11), -. https://europub.co.uk/articles/-A-22783