|
|
Effect of background population on Bayesian conclusions in speaker comparison
Štrosová, Veronika ; Skarnitzl, Radek (advisor) ; Bořil, Tomáš (referee)
Salient goal of forensic phonetics is evaluation of evidence. For this purpose, this field widely uses the Bayesian approach to evidence evaluation, specifically the likelihood ratio (LR). In this probabilistic method of evidence evaluation, which works with concepts of similarity and probability, is background population one of its main cornerstones. This thesis focuses on the effect of background population on the performance of LR based forensic speaker comparison. The LR based classification used midpoint formant values F1, F2 and F3 from Czech vowels of male native speakers. Using Lindleys (1977) parametric LR approach, these acoustic parameters were used to observe the effect of background population varying from 1 to 30 speakers on the validity and reliability of the classification. The analysis showed effect of background population on system performance, whereby the LR output had the lowest validity and reliability using small sample sizes of reference speakers. Simultaneously, the analysis showed an onset of stabilization of the system performance from the perspective of validity as well as reliability from a threshold of background population of 10 reference speakers. In particular cases was also observed synonymous classification validity using very low reference samples with...
|
|
|
Acoustic measures of glottalization in Czech
Mazúr, Richard ; Bořil, Tomáš (advisor) ; Houzar, Alžběta (referee)
Word-initial vowel glottalization is used in Czech as a boundary signal. While its contextual occurrence is predictable, it is not always realized. In addition, it is not always realized as a canonical glottal stop. Acoustic parameters which would quantify and identify the various types of glottalizations are selected based on literature research and exploratory data analysis. These acoustic parameters are then used in a machine learning categorization model. Results show that with a small number of parameters, satisfactory results can be obtained, and thus these parameters are deemed suitable in characterizing these glottalizations. acoustics parameters categorization glottalization glottal stop
|
|
|
Production and perception of melody in whisper
Hanžlová, Adléta ; Bořil, Tomáš (advisor) ; Houzar, Alžběta (referee)
Melody is a suprasegmental feature of speech and its perception depends mainly on the speed of vocal fold oscillation reflected in the fundamental frequency (f0). Whisper is defined by the absence of phonation and therefore the lack of fundamental frequency. Intended melody in whisper, however, seems to be discernable regardless of this lack of f0. In my thesis, I consider the topic of melody in whisper from a perceptual and acoustic point of view. I present a perception experiment assessing the discernability of melody in whispered words as well as words sung in whisper. This experiment proved that melody in whisper in certain cases can in fact be discerned. I then further assess the effect of intended melody in whisper on formant frequencies, formant to formant ratios, center of gravity and spectral slope. In whispered speech, the acoustical parameters affected by intended melody turned out to be F2 and center of gravity of stop-band filtered signal with frequencies containing main formanth bandwidths removed. In words sung in whisper, the affected parameters are F2, F3, F2:F1 and F3:F2 ratios, center of gravity and spectral slope.
|
|
|
Perception of affective states in speech in variable musical context
Englmaierová, Martina ; Bořil, Tomáš (advisor) ; Volín, Jan (referee)
This thesis deals with perception of affective states in speech in variable musical context. We assembled a perception test from 15 - 20s excerpts from Czech films, which contained anger or joy in speech. Then we added different types of music to the excerpts according to Russell's model of affect. Respondents were asked to assess the way the items affected them on the scales of arousal and pleasantness. The test consisted of a total of 74 items in four blocks: 24 test items (4 items with joy and 4 with anger in speech in variants without music, with music supporting and with music not supporting the affect in speech) and 50 distractors in pseudo- random order. The test was performed by 56 respondents. We have found that joy is not as dominant as anger in speech, which is very unpleasant and arousing. However, any music has increased the arousal and pleasantness of joy. Despite Russell's model, sadness also caused more arousal in relation both to joyful and angry items. In accordance with the model, relaxed music increased the pleasantness of joy while angry and sad music decreased it. Supportive music emphasized even more pleasantness and arousal in the angry affect in speech. Contrary to expectations, only relaxed music (not joyful) increased the pleasantness. However, arousal was also increased...
|
|
|
Use of glottalization as a factor enabling speaker identification
Skákal, Ladislav ; Skarnitzl, Radek (advisor) ; Bořil, Tomáš (referee)
While handling the task of speaker identification, forensic phoneticians use a combination of various parameters contained in different levels of speech signal. The main aim of the present thesis is to explore whether glottalization in Czech may be considered as a potentially useful parameter in this sense. In our research, we focus on the rate of prevocalic glottalization at word boundaries and we distinguish between different realisations of glottalization: canonical glottal stop and its hypoarticulated form - creaky voice. The studied material consists of repeated recordings of three male and four female speakers and contains both read text and spontaneous speech. The results do not indicate that the same speaker would use glottalization differently in the first and second recording, but a difference in glottalization is found between speakers. From the forensic phonetics point of view, this finding seems to be useful. Marginally, some other factors which are not directly connected with the speaker (height of following vowel, lexical factors and speech rate) were examined, but no influence on glottalization was found. Keywords: glottal stop, glottalization, forensic phonetics, speaker identification
|
|
|
Phonetic attrition and cross-linguistic influence in L1 speech of late Czech-French bilinguals
Hévrová, Marie ; Bořil, Tomáš (advisor) ; Delais-Roussarie, Élisabeth (referee) ; Skarnitzl, Radek (referee)
An increasing number of studies has already investigated the phonetic first language attrition and cross-linguistic influence in the L1 speech of late bilinguals. Neverthe- less, none of these studies has yet to examine the phonetic first language attrition and cross-linguistic influence in the L1 speech of late Czech-French bilinguals, al- though Czech and French languages represent interesting phonetic differences both at segmental and suprasegmental level. This thesis aims to fill this gap. Based on the models of L2 speech production and perception, results of studies on phonetic cross-linguistic influence and phonetic differences between Czech and French lan- guage, the hypothesis of this thesis predicts that phonetic cross-linguistic influence will occur in L1 speech of late Czech-French bilinguals. This hypothesis is tested in two studies by comparing the L1 speech production in a reading aloud task and semi-spontaneous speech of late Czech-French bilinguals with that of Czech mono- linguals. The first study consists of a perception experiment that investigated if the L1 speech of 14 late Czech-French bilinguals may be perceived as less typically Czech sounding compared to that of 11 Czech monolinguals by Czech monolingual listeners. The second study is composed of acoustic analyses comparing the...
|
|
|
Voicing of selected consonants in whispered Czech
Svatošová, Michaela ; Bořil, Tomáš (advisor) ; Šturm, Pavel (referee)
Czech is one of the languages in which the phonological contrast of voicing is primarily realized by the presence or absence of the fundamental frequency. In whisper, however, the vocal folds do not vibrate, therefore voicing is absent. Despite that, in certain situations people do communicate successfully using whispered speech. Research on a range of languages has shown that listeners are able to discriminate voicing pairs in whisper even in isolated words or syllables which have no meaning, that is, on the basis of acoustic cues alone. The aim of this thesis was to verify these findings for Czech and to gain a better understanding of the acoustic features that contribute to the intelligibility of whisper. I have analysed recordings of the voicing pairs of Czech plosives and fricatives embedded in modal and whispered nonsense words. Of the various parameters measured, voicing counterparts differed most prominently in duration, although to a lesser extent than in modal phonation, suggesting weaker pre- servation of an otherwise redundant feature. In addition, the perception test confirmed a very good recognition of phonological voicing in whisper, but with substantial variability between individual voicing pairs. Furthermore, the combi- nation of the acoustic and perceptual perspective provided an...
|
|
|
Imitation of the Russian accent by Czech speakers
Nudga, Natalia ; Skarnitzl, Radek (advisor) ; Bořil, Tomáš (referee)
(in English) Foreign accent imitation is one of the strategies used in intentional voice disguise. This thesis focuses on imitation of the Russian accent in Czech language by Czech speakers and describes the imitated accent based on auditory and acoustic analysis. Both segmental and suprasegmental features of speech have been analysed based on the comparison of audio recordings of regular speech and speech with imitated Russian accent. The most frequent difference implemented by Czech speakers during the imitation task involved the change of duration of vowels in relation to the position of word stress and included both lengthening of stressed vowels and shortening of unstressed ones. Change of vowel quality has been performed mostly on the vowel [ɪ], resulting in a close vowel [i] or resembling [ɨ]. Consonantal deviations usually concerned Czech sounds [ɦ] and [r̝ ], and palatalization of lateral consonant [l]. Audio recordings of Czech imitators as well as Russian speakers were used in the perception test, in which participants had to judge the authenticity of foreign accent. Four out of ten imitators were rated by the majority of listeners as genuine non-native speakers, whereas two out of five Russian speakers were misjudged for imitators. The accent of successful imitators was characterised by...
|
|
|
Acoustic features of speech in multiple sclerosis
Svoboda, Emil ; Bořil, Tomáš (advisor) ; Šturm, Pavel (referee)
This thesis analyzes what acoustically sets apart recordings of healthy people from recordings of people afflicted with multiple sclerosis, and how this distinction can be used to automatically detect multiple sclerosis from fairly simple recordings of a subject's voice, potentially discovering early cases of this disease. Chapter 1 includes the theoretical background of the effect of multiple sclerosis on speech and the descriptions of the data, software, hypotheses and assumptions used here. Two sets recordings of read speech were used, a corpus of afflicted speakers and a control corpus of healthy speakers, totalling 250 individuals. A subset of this corpus was manually annotated, resulting in one dataset. Simultaneously, these entire corpora were also annotated automatically, resulting in another dataset, which was created to explore the possibility of detecting multiple sclerosis automatically. Chapter 2 describes the 13 acoustic parameters used in this thesis, their exact hypothesized relationships with the symptoms of multiple sclerosis and the ways they were calculated. Chapter 3 elaborates on the statistical testing of the aforementioned parameters, their interpretation, the success rate of the two machine learning models used to assess their total predictive power, and a potential way to apply the...
|
|
|
Sophisticated strategies of voice disguise and their phonetic character
Růžičková, Alžběta ; Bořil, Tomáš (advisor) ; Šturm, Pavel (referee)
Speech contains certain attributes characteristic for a speaker, so-called idiosyncratic features. This thesis focuses on the form of these features in intentional voice disguise - whether speakers are able to change them in a substantial way, or if they tend to remain stable in spite of intentional speech modifications. It was also investigated whether any general tendencies to similar changes of such features under voice disguise exist among the speakers. The observed features were statistical f0 indicators, f0 contours, vowel formants, long- term formant distributions, spectral characteristics of sibilants, intensity, intensity contours, speech and articulation rate, %V and local articulation rate contours. In f0 median and standard deviation, vowel formants, LTFDs, intensity, articulation rate, and %V, prominent shifts under voice disguise were observed in general; in the majority of these parameters, the shifts differed among speakers. However it was found that the value of %V generally tends to rise under voice disguise. Also, intensity showed an increase in majority of cases. In f0 contours, similar patterns were observed among speakers in normal speech, however, in disguised speech, greater differences appeared among speakers; speakers tend to employ nonstandard dynamic f0 patterns more...
|
guest ::
