ISCA - International Speech
Communication Association


ISCApad Archive  »  2020  »  ISCApad #264  »  Recent Theses

ISCApad #264

Wednesday, June 10, 2020 by Chris Wellekens

8 Recent Theses
8-1Bart Penning de Vries, 'Computerised Speaking Practice: The Role of Automatic Corrective Feedback in Learning L2 Grammar´

Bart Penning de Vries has completed his PhD thesis entitled
'Computerised Speaking Practice: The Role of Automatic Corrective
Feedback in Learning L2 Grammar´ at the Centre of Language and Speech
Technology (CLST) of the Radboud University Nijmegen, the Netherlands.
Promotor - Prof. Dr. R.W.N.M. van Hout
Copromotors - Dr. C. Cucchiarini, Dr. H. Strik

http://www.ru.nl/english/education/masters/historical-studies/?ActLbl=pagina&ActItmIdt=996617

Link to the document:
http://repository.ubn.ru.nl/bitstream/handle/2066/141390/141390.pdf

Email address: bardtpdv@gmail.com

Back  Top

8-2Prasanna Kumar Muthukumar, 'Towards Integrated Acoustic Models for Speech Synthesis'

Prasanna Kumar Muthukumar, 'Towards Integrated Acoustic Models for Speech Synthesis',

CMU-LTI-15-019

Advisor; Alan W. Black

Language Technologies Institute,

School of Computer Science,

Carnegie Mellon University,

5000 Forbes Avenue,

Pittsburgh, PA 15213.

 Thesis text can be found at http://www.cs.cmu.edu/~pmuthuku/publications/thesis/Prasanna_thesis.pdf

Back  Top

8-3Stephen Bodnar,'Affective L2 learning experiences and ideal L2 selves in spoken CALL practice'

Stephen Bodnar has completed his PhD thesis entitled
'Affective L2 learning experiences and ideal L2 selves in spoken CALL practice'
at the Centre of Language and Speech Technology (CLST) of the Radboud University
Nijmegen, the Netherlands.
Promotor - Prof. Dr. R.W.N.M. van Hout
Copromotors - Dr. C. Cucchiarini, Dr. H. Strik

http://repository.ubn.ru.nl/handle/2066/157803

Links to the document:
http://repository.ubn.ru.nl/bitstream/handle/2066/157803/157803.pdf?sequence=1
http://www.lotpublications.nl/Documents/428_fulltext.pdf

Back  Top

8-4Andreas Windmann, 'Optimization-based modeling of suprasegmental speech timing'

Name: Andreas Windmann
Email: windmannandreas@gmail.com
Gender: male
Thesis title: Optimization-based modeling of suprasegmental speech timing
Supervisor: Petra Wagner
University: Bielefeld University
URL: https://pub.uni-bielefeld.de/publication/2906598


 

Back  Top

8-5Catharine Oertel Genannt Bierbach, 'Modeling Engagement in Multi-Party Conversations'
Name: Catharine Oertel Genannt Bierbach
Title: Modeling Engagement in Multi-Party Conversations (2017)
Gender: female
University: KTH
Supervizor: Prof. Joakim Gustafson
Back  Top

8-6Hardik B. Sailor,'Auditory Representation Learning'

Name: Hardik B. Sailor
Email: sailor_hardik@daiict.ac.in
Sex: male
Thesis title: Auditory Representation Learning
Supervisor: Prof. Hemant A. Patil
University: Dhirubhai Ambani Institute of Information and Communication
Technology (DA-IICT), Gandhinagar-382007, Gujarat, India
URL: https://drive.google.com/open?id=11l3d0imk2LuvYUoNgubdDzV3gocIw-pV

Back  Top

8-7Philipp Aichinger, 'Diplophonic Voice - Definitions, models, and detection'

 

Subject: Philipp Aichinger, 'Diplophonic Voice - Definitions, models, and detection'

Body of the announcement:

Email address: philipp.aichinger@meduniwien.ac.at
Philipp Aichinger has completed his PhD thesis entitled 'Diplophonic Voice - Definitions, Models, and Detection', which was conducted at the Medical University of Vienna and the Graz University of Technology, Austria. Examiners were Gernot Kubin, Jean Schoentgen, and Berit Schneider-Stickler. Link to the document: https://www.researchgate.net/publication/271441725_Diplophonic_Voice_-_Definitions_models_and_detection

Back  Top

8-8Yun Wang, 'Polyphonic Sound Event Detection with Weak Labeling'

Yun Wang graduated as a PhD from CMU October 2018

Thesis advisor: Prof. Florian Metze (CMU)

Back  Top

8-9Omid Ghahabi, 'Deep Learning for i-Vector Speaker and Language Recognition'

Omid Ghahabi, 'Deep Learning for i-Vector Speaker and Language Recognition'

email address: omid.ghahabi@eml.org

 

Omid Ghahabi completed his PhD thesis entitled 'Deep Learning for i-Vector Speaker and Language Recognition' at Universitat Politecnica de Catalunya (UPC), Barcelona, Spain. It was supervised by Prof. Javier Hernando at TALP Research Center, Department of Signal Theory and Communications.
Back  Top

8-10Neeraj Kumar Sharma, 'Information-rich Sampling of Time-varying Signals'

 

 

 

Thesis Author: Neeraj Kumar Sharma

 

Current Affiliation:

 

Post-Doctoral Fellow

 

Carnegie Mellon University

 

Pittsburgh 15213, USA

 

E-mail: neerajww@gmail.com

 

URL: neerajww.github.in

 

PhD Granting Institution:

 

Dept. of Electrical Communication

 

Engineering (ECE)

 

Indian Institute of Science

 

Bangalore 560012, India

 

Thesis Advisor:

 

Dr. Thippur V. Sreenivas

 

Professor, Dept. ECE

 

Indian Institute of Science

 

Bangalore 560012, India

 

E-mail: tvsree@iisc.ac.in

 

 

 

Thesis title: Information-rich Sampling of Time-varying Signals

 

Abstract: This thesis investigates three fundamental concepts of interest to time-varying signal analysis: sampling, modulations and modelling.

 

 

 

The underlying goal is speech/audio signal processing and the motivation is drawn by exploring how these information rich signals are represented in the human auditory system. The rich information content of speech naturally requires the signals to be highly time-varying, as is

 

evident in the joint time-frequency representation, such as the short-time Fourier transform. Although the theoretical bandwidth of such time-varying signals is infinite, the auditory nerves are known to carry only low rate sampled information of these signals to the cortex, and yet obtain a rich information content of these signals. Thus, it may be unnecessary to sample the signals at a uniform Nyquist rate, as is done in all current day technology applications. Further, the present day quasi-stationary models of speech/audio, based on a linear time-invariant system may be inadequate. Instead of these models, the thesis explores signal decomposition using time-varying

 

signal components, namely, the amplitude and frequency modulations (AM-FM). The contributions are presented in three parts, and combined these suggest an alternative analysis techniques for fine spectro-temporal analysis of time-varying signals.

 

In part 1, the thesis analyzes non-uniform event-triggered samples, namely zero-crossings (ZCs) and extrema of the signal. The extrema are the ZCs of the signal first derivative and similarly the ZCs of higher derivative of the signal, denoted HoZC-d; using the sparse signal reconstruction approach, the different 'd' HoZCs are compared for their efficiency to reconstruct the signal based on different signal models. It is found that HoZC-1 outperform others, and a combination of HoZC-1 and HoZC-2 provides acceptable reconstruction.

 

In part 2, analyzing an AM-FM signal, it is shown that extrema samples (HoZC-1) are better than ZCs or LCs, in estimating the AM and FM components through local polynomial regression. Similarly, HoZC-1 can provide better AM-FM estimation of sub-band speech, moving source Doppler signal, etc., compared to DESA-1 and analytic signal approach, with additional benefit of sub-sampling. Extending the analysis to arbitrary multi-component AM-FM signals, it is shown that the successive derivative operation aids in separating the highest FM component as the dominant AM-FM component out of the multiple components. This is referred to as the 'dominant

 

instantaneous frequency principle' and is used for sequential estimation of individual mono-component AM-FM signals in the multi-component mixture.

 

The part 3, focusing on speech signals, visits time-varying sinusoidal modeling of speech, and proposes an alternate model estimation approach. The estimation operates on the whole signal without any short-time analysis. The approach proceeds by extracting the fundamental frequency sinusoid (FFS) from speech signal. The instantaneous amplitude (IA) of the FFS is used for voiced/unvoiced stream segregation. The voiced stream is then demodulated

 

using a variant of in-phase and quadrature-phase demodulation carried at harmonics of the FFS. The result is a non-parametric time-varying sinusoidal representation, specifically, an additive mixture of quasi-harmonic sinusoids for voiced stream and a wideband mono-component sinusoid for unvoiced stream. The representation is evaluated for analysis-synthesis, and the bandwidth of IA and IF signals are found to be crucial in preserving the quality. Also, the obtained IA and IF signals are found to be carriers of perceived speech attributes, such as speaker characteristics and intelligibility. On comparing the proposed modeling framework with the existing approaches,

 

which operate on short-time segments, improvement is found in simplicity of implementation, objective-scores, and computation time. The listening test scores suggest that the quality preserves naturalness but does not yet beat the state-of-the-art short-time analysis methods. In summary, the proposed representation lends itself for high resolution temporal analysis of non-stationary speech signals, and also allows quality preserving modification and synthesis.

 

 

 

URL:

https://drive.google.com/open?id=17Olne0RBkVHRd2HcmJc0f44e17m47NZB





 

 

Back  Top



 Organisation  Events   Membership   Help 
 > Board  > Interspeech  > Join - renew  > Sitemap
 > Legal documents  > Workshops  > Membership directory  > Contact
 > Logos      > FAQ
       > Privacy policy

© Copyright 2024 - ISCA International Speech Communication Association - All right reserved.

Powered by ISCA