% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@PHDTHESIS{Kob:59935,
author = {Kob, Malte},
othercontributors = {Vorländer, Michael},
title = {{P}hysical modeling of the singing voice},
address = {Aachen},
publisher = {Publikationsserver der RWTH Aachen University},
reportid = {RWTH-CONV-121675},
pages = {VII, 154 S. : Ill., graph. Darst.},
year = {2002},
note = {Aachen, Techn. Hochsch., Diss., 2002},
abstract = {This thesis deals with the physical modeling of the parts
of the voice organ relevant for voice generation plus
techniques for the measurement of acoustic voice properties.
An introduction to characteristics of the voice signal is
followed by a literature survey of existing approaches for
the most important functional voice components. Algorithms
that seem to be suitable for modeling of the singing voice
are adopted and extended. The modeling of the vocal fold
movement uses a three-dimensional, symmetric multiple mass
model that is capable of simulating different voice
registers and voice pathologies that are found in singers.
For the wave propagation in the space between glottis and
mouth opening, the vocal tract, two algorithms are
presented, which have been optimized for different
applications. The first model is based on cylinder segments
and requires a fixed sampling rate that yields a high
resolution in space. The second model allows an arbitrary
choice of the sampling rate and makes it possible to reduce
the number of parameters for the description of the vocal
tract by using conical segments. Since the noise component
is required for a natural sounding voice, a model is
implemented that simulates vortex shedding and sound
generation by turbulences. The dependence of the noise
component on the choice of the articulated speech sound is
described by analysis of the voice signal in the domains of
time and frequency. The resonance characteristics of the
vocal tract are evaluated with two measurement approaches: a
direct method that determines the transfer function and a
mobile, non-invasive set-up for the measurement of the
acoustic impedance at the mouth. For comparison of the
characteristic radiation of the human voice with an
artificial singer, a measurement set-up is described that
allows a detailed visualization of the directivity. The
final part of this work investigates the interaction of the
elements of the model. Some examples for the application of
the singing voice model to the simulation of different
singing styles and voice pathologies are presented.
Different voice registers are modelled with special emphasis
on the simulation of overtone singing. The impedance
measurements were the basis for the parameter choice of the
vocal tract model. As a future application of the model, the
investigation of voice pathologies is planned. First
attempts to model edema of the vocal folds and singer's
nodules are presented and the use of the model as a
therapeutic tool for voice therapy is discussed.},
cin = {600000},
ddc = {620},
cid = {$I:(DE-82)600000_20140620$},
typ = {PUB:(DE-HGF)11},
urn = {urn:nbn:de:hbz:82-opus-3930},
url = {https://publications.rwth-aachen.de/record/59935},
}
guest ::