Publication Date

2007-01-01

Availability

Open access

Degree Type

Thesis

Degree Name

Master of Science (MS)

Department

Electrical and Computer Engineering (Engineering)

Date of Defense

2007-11-21

First Committee Member

Michael Scordilis - Committee Chair

Second Committee Member

Xiaodang Cai - Committee Member

Third Committee Member

SubramanianRamakrishnan - Committee Member

Abstract

In automatic speech recognition systems (ASRs), training is a critical phase to the system?s success. Communication media, either analog (such as analog landline phones) or digital (VoIP) distort the speaker?s speech signal often in very complex ways: linear distortion occurs in all channels, either in the magnitude or phase spectrum. Non-linear but time-invariant distortion will always appear in all real systems. In digital systems we also have network effects which will produce packet losses and delays and repeated packets. Finally, one cannot really assert what path a signal will take, and so having error or distortion in between is almost a certainty. The channel introduces an acoustical mismatch between the speaker's signal and the trained data in the ASR, which results in poor recognition performance. The approach so far, has been to try to undo the havoc produced by the channels, i.e. compensate for the channel's behavior. In this thesis, we try to characterize the effects of different transmission media and use that as an inexpensive and repeatable way to train ASR systems.

Keywords

Speech Recognition; System Identification; Channel Modeling; Adaptive Filters; Filter; ARMA; IIR; FIR; ASR; ROC; Optimization; Search; Cyclic Coordinate; Minima; Optimum; Feature; MFCC; Mel Frequency; Cepstral Coefficient; CMS; RASTA; Differential; Breakdown Effect; RLS; RMS; LMS; NLMS; Self-orthogonalizing; Norm; Space; Feature Space; Time Domain; Frequency Domain; Cepstral Domain; Cepstrum; Error Minimization; Error; Measure; Telephone; Landline; Cellphone; Voice Over Ip; Voip; Ip; Jitter; Network; Gilbert-Elliot Model; Impulse Response; Distance Gain Ratio; MOS; PESQ; Perceptual; Audio; Speech; Sound; Voice; Vocal; Model; Channel; System; Training; Simulation; Simulate; IBM; UM-IBM Project; Linear; Nonlinear; Time-invariant; LTI; Dynamic

Share

COinS