Doctoral theses currently supervised by Prof. Sendlmeier
How voices and faces affect our perception of attractiveness
An empirical analysis of the effect of a German original film compared to the Italian dubbing voice using the film “The Baader Meinhof Complex” as an example
Optimal parameterization of 3D motion capture data to animate talking heads
Shared communication strategies of language and music - musical intervals and rhythms in emotional speech
Short description:
The dissertation focuses on the analysis of the tonal and rhythmic structuring of emotional speech with regard to acoustic perception categories, whose impression effect has already been studied in detail in music psychology.
Author:
Hans-Henning Raven, M.A.
Reviewer:
Prof. Dr. Walter F. Sendlmeier
Date:
29.01.2006
Summary:
This research project assumes that the affect content of emotional speech is conveyed similarly to emotions in music. The primary aim is to identify acoustic perceptual categories that are responsible for the emergence of impressions of emotion in emotional speech.
Focus is placed on those categories whose impression effect has already been studied in detail in music psychology. Emotion in music is largely conveyed based on tonal tensions between successive tones and chords as well as rhythmic structuring of the acoustic signal. Musical categories (tonality, time signature, etc.) and research results from music psychology have received little attention in previous prosodic studies of emotional speech. With the help of perception tests and acoustic analyses, expressions of the emotions joy, sadness, anger, fear, disgust, and boredom will be examined more closely with regard to tonal and rhythmic structuring. More than 500 emotionally linguistic utterances provided by actors are available as a data basis.
Doctoral theses previously supervised by Prof. Sendlmeier
How the central Austro-Bavarian, Upper Saxonian, and Berlin dialects impact perception of the personality traits conscientiousness and agreeableness
Author:
Kerstin Trillhaase
Reviewer:
Prof. Dr. Walter F. Sendlmeier
Date:
Completed 2021
Summary:
How we perceive others is significantly impacted by their voices and way of speaking. This means that in addition to age, gender, and emotional well-being, we can also infer enduring character traits. Moreover, a person’s dialect provides clues to their regional origin and at the same time evokes judgments linked to the dialect. But how does this impact our perception of personality? Are people perceived or judged differently if they speak the standard language versus a dialect?
This question is answered using listening experiments and a standardized procedure of personality diagnostics. Focus is placed on Upper-Saxonian and central Austro-Bavarian as these are two of the most prominent dialects in German. In addition, linguistic-historical, political as well as socio-demographic contributing factors are considered as possible triggers for dialect-related stereotypes.
Publication of the dissertation:
How a dubbed voice influences character representation
Hearing aid vs. inner ear prosthesis audiometric indication for cochlear implantation in prelingually severely hearing impaired children.
Author:
Silvia Zichner
Reviewer:
Prof. Dr. Walter F. Sendlmeier
Date:
Completed 2012
Summary:
Cochlear implantation has been a successful therapy for postlingually deaf adults and prelingually deaf children as well as, under certain conditions, adolescents in Germany for over 20 years.
While the audiometric criterion was complete deafness in the first years of CI treatment, it has been shown in recent years that the CI is superior to hearing aids even in cases of residual hearing loss or hearing loss bordering on deafness. However, it is still not sufficiently clear where exactly the audiometric threshold for the indication of cochlear implantation is and whether it can be defined precisely at all.
Current practice assumes a mean hearing loss of about 80dB in the better hearing ear for CI indication. However, it is possible that children with hearing loss below 80 dB would also benefit from CI. To determine this, the children's hearing development (sound/speech audiometry, etc.) as well as speech development (speech comprehension, expressive language communicative competences, etc.) will be used as assessment criteria.
The results could be key to decisions about optimal care for very young hearing-impaired children who do not yet have significant language development.
Publication of the dissertation:
Age-related changes in the voice and speech of women
Short description:
Acoustic, phonetic, and perceptive analyses of the voices and speech of women of different ages will be conducted.
Author:
Markus Brückl, M.A.
Reviewer:
Prof. Dr. Walter F. Sendlmeier
Prof. Dr.-Ing. Klaus Hobohm
Date:
Completed 2011
Summary:
Every oral utterance contains information about a speaker’s gender, emotion, personality and age.
This dissertation will examine the theoretical and physiological principles of age-related changes in the voice and speech of women. It will investigate how precisely such changes can be attributed to the speakers’ age and which acoustic traits can convey the age information.
For this purpose, several utterance types of 88 female speakers of all adult ages are collected and analyzed in detail in cross-sectional and longitudinal studies.
The analyses integrate phonetic, acoustic, and perceptual psychological methods. Extensions of standard procedures for determining perception reliability will be developed to determine perceptual judgments.
Publication of the dissertation:
Acoustic correlates of vocal emotional expression in spoken language
Voice quality characteristics of young and old people
Short description:
This dissertation aims to systematically investigate the acoustic characteristics of adult speaker voices resulting primarily from age-related changes in the larynx.
Author:
Ralf Winkler M.A.
Reviewer:
Prof. Dr. Walter F. Sendlmeier
Date:
Completed 2009
Summary:
This dissertation aims to systematically investigate the acoustic characteristics of adult speaker voices resulting primarily from age-related changes in the larynx. In addition to the age-typical expression of voice-qualitative parameters, their perceptual relevance will also examined. The analysis of electroglottographic data (EGG) of sustained vowels is suitable for investigating the vibrational behavior of the vocal folds as well as the analysis of longer, articulatory more complex signal segments such as words and sentences.
From these analyses, hypotheses about the source signal can be derived, which are then supported or validated by inverse filtering of the acoustic signal. Resynthesis will be used to try to systematically generate an impression of age (perceptual voice age) in synthetic speech samples. Knowledge about the distribution of voice qualitative parameters with respect to the true (chronological) age of speakers will be applied in the machine estimation of speaker age based on speech samples.
Publication of the dissertation:
German word stress - symbols, signals
Approaches to speaker-independent recognition of foreign-accented speech
3D articulatory speech synthesis
A phonetic-acoustic motivated inventory for the synthesis of German utterances
Digital archiving and network-based distribution of audio material
Author:
Dr. Florian Krause
Reviewer:
Prof. Dr. Walter F. Sendlmeier
Date:
Completed 2006
Summary:
This dissertation focuses on the archiving of audio files in digital format as well as the distribution of these data via networked computers. It applies an interdisciplinary approach aiming to combining perspectives from studio automation and information technology. One of its primary aims is to develop a software system that allows users to affordably send music and voice and written messages, as well as other content to other users via an IP-based network. The system will be managed via a user-friendly website, while the actual distribution process will run transparently in the background for the user.
Unlike audio streaming, there is no continuous data stream sent to the participants; instead only the required elements as well as the meta data needed to create the program are sent.
Phonetic changes in emotional speech
Short description:
The aim of this study is to find evidence for the extent to which these emotions differ in terms of articulation accuracy and resulting acoustic parameters by measuring acoustic and articulatory parameters.
Author:
Dr. Miriam Kienast (now Rolfes)
Reviewer:
Prof. Dr. Walter F. Sendlmeier
Date:
Completed 2002
Summary:
This investigation focuses on the speech expression of emotions. Focus will be placed on changes to articulatory precision when expressing the emotions joy, sadness, anger, fear, disgust, and boredom compare to neutral speech. The aim of this study is to find evidence for the extent to which these emotions differ in terms of articulation accuracy and resulting acoustic parameters by measuring acoustic and articulatory parameters.
Prosodic analysis of emotional speech
Short description:
This dissertation analyzes prosodic and spectral features of emotional speech. Particular focus is placed on the characteristic traits of the emotions joy, sadness, anger, fear, disgust, and boredom compared to neutral speech.
Author:
Dr. Astrid Paeschke (now Bartels)
Reviewer:
Prof. Walter F. Sendlmeier
Date:
Completed 2003
Summary:
This dissertation analyzes prosodic and spectral features of emotional speech. Particular focus is placed on the characteristic traits of the emotions joy, sadness, anger, fear, disgust, and boredom compared to neutral speech. The acoustic parameters studied are predominantly features of fundamental frequency (F0) and concern range, declension, and the shape and course of fundamental frequency in syllables, accents, phrases, and whole sentences. The speech material studied is from the database created in the research project "Phonetic Elaboration and Reduction in Emotional Speech," which contains both individually spoken sentences and a short text.
Publication of the dissertation:
Simulation of emotional speech using speech synthesis methods
Phonetic aspects of the acquisition of German as a foreign language by Italian native speakers
Short description:
This dissertation examines the phonetic difficulties adult Italian native speakers have when learning German. It aims to investigate the reasons for the characteristic “foreign accent” of Italian German language learners.
Author:
Prof. Dr. Federica Missaglia
Reviewer:
Prof. Dr. Walter F. Sendlmeier
Date:
Completed 1999
Audiovisual speech synthesis - system development and assessment
Short description:
Visual information can be extremely helpful for understanding natural language, particularly under poor acoustic conditions. Human-machine interfaces can also be improved in many cases by speech output using visual and audio signals, as coherent audiovisual speech synthesis increases clarity compared to audio-only presentation.
Author:
Dr. Sascha Fagel
Reviewer:
Prof. Dr. Walter F. Sendlmeier
Date:
Completed 2004
Summary:
Visual information can be extremely helpful for understanding natural language, particularly under poor acoustic conditions. Human-machine interfaces can also be improved in many cases by speech output using visual and audio signals, as coherent audiovisual speech synthesis increases clarity compared to audio-only presentation.
An audiovisual language synthesis system is developed, implemented, and evaluated.
Publication of the dissertation:
Processing of emotional speech by CI patients
A systematic approach to evaluating loudness models
Author:
Cornelius Bradter
Reviewer:
Prof. Dr. Sendlmeier
Summary:
This dissertation discusses possibilities and concepts to determine the sensory quantity loudness by analyzing a signal. For this purpose, factors determining loudness are identified and their quantitative recording discussed. By analyzing established loudness determination methods, this research will show the extent these factors have found their way into practical application. Improvements for the implementation of the factors and new ideas for loudness determination are proposed. To achieve this, a modular software for the construction of loudness determination procedures is also developed and presented.