Voice transformation and development of related speech analysis tools for Turkish

Download

index.pdf

Date

2005

Author

Salor, Özgül

Metadata

Show full item record

Item Usage Stats

175
views

103
downloads

In this dissertation, new approaches in the design of a voice transformation (VT) system for Turkish are proposed. Objectives in this thesis are two-fold. The first objective is to develop standard speech corpora and segmentation tools for Turkish speech research. The second objective is to consider new approaches for VT. A triphone-balanced set of 2462 Turkish sentences is prepared for analysis. An audio corpus of 100 speakers, each uttering 40 sentences out of the 2462-sentence set, is used to train a speech recognition system designed for English. This system is ported to Turkish to obtain a phonetic aligner and a phoneme recognizer. The triphone-balanced sentence set and the phonetic aligner are used to develop a speech corpus for VT. A new voice transformation approach based on Mixed Excitation Linear Prediction (MELP) speech coding framework is proposed. Multi-stage vector quantization of MELP is used to obtain speaker-specific line-spectral frequency (LSF) codebooks for source and target speakers. Histograms mapping the LSF spaces of source and target speakers are used for transformation in the baseline system. The baseline system is improved by a dynamic programming approach to estimate the target LSFs. As a second approach to the VT problem, quantizing the LSFs using k-means clustering algorithm is applied with dimension reduction of LSFs using principle component analysis. This approach provides speaker-specific codebooks out of the speech corpus instead of using MELP's pre-trained LSF codebook. Evaluations show that both dimension reduction and dynamic programming improve the transformation performance.

Subject Keywords

Telecommunication.

URI

http://etd.lib.metu.edu.tr/upload/3/12605787/index.pdf
https://hdl.handle.net/11511/14971

Collections

Graduate School of Natural and Applied Sciences, Thesis

Suggestions

OpenMETU
Core

Simulation based investigation of an improvement for faster SIP re-registration Tanrıverdi, Eda; Bilgen, Semih; Department of Electrical and Electronics Engineering (2004) In this thesis, the Session Initiation Protocol (SIP) is studied and an improvement for faster re-registration is proposed. This proposal, namely the أregistration ا activationؤ, is investigated with a simulation prepared using OPNET. The literature about wireless mobile networks and SIP mobility is reviewed. Conditions for an effective mobile SIP network simulation are designed using message sequence charts. The testbed in [1] formed by Dutta et. al. that has been used to observe SIP handover performance i...
Design and construction of reduced size planar spiral antenna in the 0.5-18 ghz frequency range Yıldız, İnanç; Hızal, Altunkan; Department of Electrical and Electronics Engineering (2004) In this thesis, theoretical and practical evaluation of usual spiral antenna is revised. Working principles of both types of planar spiral antennas as Equiangular and Archimedean are introduced. A predesigned microstrip tapered balun used for feeding section of a spiral antenna is simulated on Ansoft HFSS software. Successful simulation results are obtained and measurements of implemented balun structure are made by using an HP 8722 D vector network analyzer. Antenna measurement techniques used in this stud...
Nonlinear estimation techniques applied to econometric problems Aslan, Serdar; Demirbaş, Kerim; Department of Electrical and Electronics Engineering (2004) This thesis considers the filtering and prediction problems of nonlinear noisy econometric systems. As a filter/predictor, the standard tool Extended Kalman Filter and new approaches Discrete Quantization Filter and Sequential Importance Resampling Filter are used. The algorithms are compared by using Monte Carlo Simulation technique. The advantages of the new algorithms over Extended Kalman Filter are shown.
Constructions of resilient boolean functions with maximum nonlinearity Şahin, M. Özgür; Yücel, Melek D; Department of Electrical and Electronics Engineering (2005) In this thesis, we work on the upper bound for nonlinearity of t-resilient Boolean functions given by Sarkar and Maitra, which is based on divisibility properties of spectral weights of resilient functions and study construction methods that achieve the upper bound. One of the construction methods, introduced by Maity and Johansson, starts with a bent function and complements some values of its truth table corresponding to a previously chosen set of inputs, S, which satisfies three criteria. In this thesis,...
An overview of detection in MIMO radar Bilgi Akdemir, Şafak; Candan, Çağatay; Department of Electrical and Electronics Engineering (2010) In this thesis study, an overview of MIMO radar is presented. The differences in radar cross section, channel and received signal models in different MIMO radar configurations are examined. The performance improvements that can be achieved by the use of waveform diversity in coherent MIMO radar and by the use of angular diversity in statistical MIMO radar are investigated. The optimal detector under Neyman-Pearson criterion for Coherent MIMO radar when the interfering signal is white Gaussian noise is devel...

Citation Formats

Ö. Salor, “Voice transformation and development of related speech analysis tools for Turkish,” Ph.D. - Doctoral Program, Middle East Technical University, 2005.