Mietta Lennes
University of Helsinki, Department of Modern Languages, Department Member
- Lexical Phonology, Probabilistic Linguistics, Speech Production, Speech Annotation, Phonetics, Speech Acoustics, and 39 moreIntegration of Language Resources, Phonetics and Pronunciation, Finnish phonetics, Emergent Phonology, Speech Corpora, Praat scripting, Acoustic Phonetics, Phonetics-Phonology Interface, Teaching Phonetics, Linguistic research tools, Speech Databases, Lexical Phonetics, Probabilistic Phonetics, Phonetic Research Tools, Integrating the language resources in Finland, Corpus Phonetics, Second Language Acquisition, Computational Linguistics, Phonology, Language Variation and Change, Linguistics, Sociophonetics, Applied Linguistics, Articulatory Phonetics, Digital Humanities, Linguistic databases, Linguistic Annotation, spoken Finnish, Embodied Conversational Agents (ECAs), Laboratory/Experimental Phonology, Corpus Linguistics, Using the R statistical programming environment for linguistic analysis, Speech Processing, Speech Prosody, Language Resources, Speech Sciences, Speech perception, Languages and Linguistics, and FIN-CLARIN - Integrating the language resources in Finlandedit
This course offers a general picture of managing speech corpora and of methods that are available for the acoustic-phonetic study of speech. The course consists of six lectures during which you will use a speech analysis program called... more
This course offers a general picture of managing speech corpora and of methods that are available for the acoustic-phonetic study of speech. The course consists of six lectures during which you will use a speech analysis program called Praat and learn to apply the main features of the program to your own work with speech recordings. In addition, you will learn the basics of another program called ELAN that can be used for transcribing and annotating audio as well as video material.
Tour de CLARIN is an initiative started by CLARIN ERIC in 2016 that has been periodically highlighting prominent user involvement activities of CLARIN national consortia in the form of blog posts published on the CLARIN webpage,... more
Tour de CLARIN is an initiative started by CLARIN ERIC in 2016 that has been periodically highlighting prominent user involvement activities of CLARIN national consortia in the form of blog posts published on the CLARIN webpage, disseminated through the CLARIN news flash and on social media. By focusing a different national consortium every two months and showcasing their outstanding language resources, text processing tools, user involvement events and researchers, we have been aiming to increase the visibility of the various consortia, reveal the richness of the CLARIN landscape, and display the full range of activities throughout the network that can not only inform and inspire other consortia, but also show what CLARIN has to offer<br> to researchers, teachers, students, professionals and the general public interested in using and processing language data in various forms. In the two years we have been running the initiative, and having visited nearly half of all the CLARI...
Research Interests:
Research Interests:
Włodarczak M, Simko J, Wagner P, O'Dell M, Lennes M, Nieminen T. Finnish rhythmic structure and entrainment in overlapped speech. In: Asu E-L, Lippus P, eds. Nordic Prosody. Proceedings of the XIth Conference. Frankfurt a.M.: Peter... more
Włodarczak M, Simko J, Wagner P, O'Dell M, Lennes M, Nieminen T. Finnish rhythmic structure and entrainment in overlapped speech. In: Asu E-L, Lippus P, eds. Nordic Prosody. Proceedings of the XIth Conference. Frankfurt a.M.: Peter Lang; 2013: 421-430
Research Interests:
The Donate Speech campaign has so far succeeded in gathering approximately 3600 h of ordinary, colloquial Finnish speech into the Lahjoita puhetta (Donate Speech) corpus. The corpus includes over twenty thousand speakers from all the... more
The Donate Speech campaign has so far succeeded in gathering approximately 3600 h of ordinary, colloquial Finnish speech into the Lahjoita puhetta (Donate Speech) corpus. The corpus includes over twenty thousand speakers from all the regions of Finland and from all age brackets. The primary goals of the collection were to create a representative, large-scale resource to study spontaneous spoken Finnish and to accelerate the development of language technology and speech-based services. In this paper, we present the collection process and the collected corpus, and showcase its versatility through multiple use cases. The evaluated use cases include: automatic speech recognition of spontaneous speech, detection of age, gender, dialect and topic and metadata analysis. We provide benchmarks for the use cases, as well downloadable, trained baseline systems with open-source code for reproducibility. One further use case is to verify the metadata and transcripts given in this corpus itself, ...
Research Interests:
Tämä PDF-muotoinen versio suomenkielisestä Praat-oppaasta oli vuoteen 2016 asti saatavilla Helsingin yliopiston sivuilla osoitteessa http://www.helsinki.fi/puhetieteet/atk/praat. Jos etsit oppaan HTML-versiota 0.8 vuodelta 2004, katso... more
Tämä PDF-muotoinen versio suomenkielisestä Praat-oppaasta oli vuoteen 2016 asti saatavilla Helsingin yliopiston sivuilla osoitteessa http://www.helsinki.fi/puhetieteet/atk/praat. Jos etsit oppaan HTML-versiota 0.8 vuodelta 2004, katso https://doi.org/10.5281/zenodo.376094 Jos etsit oppaan myöhemmin päivitettyä versiota, katso https://github.com/lennes/praat-opas This is a PDF version of the Praat user's guide written in Finnish that was available at<br> http://www.helsinki.fi/puhetieteet/atk/praat until year 2016. In case you were looking for the HTML version 0.8 of this guide, see https://doi.org/10.5281/zenodo.376094 In case you need an updated version, see https://github.com/lennes/praat-opas
Research Interests:
Oppaan vuonna 2005 julkaistu versio pdf- ja html-muodossa lähdekoodeineen. Tässä zip-paketissa olevaa html-versiota voi lukea avaamalla selaimeen tiedoston index.html. PDF-versio on tiedostossa annotation_guide.pdf. Verkkoversio on... more
Oppaan vuonna 2005 julkaistu versio pdf- ja html-muodossa lähdekoodeineen. Tässä zip-paketissa olevaa html-versiota voi lukea avaamalla selaimeen tiedoston index.html. PDF-versio on tiedostossa annotation_guide.pdf. Verkkoversio on nähtävillä myös suoraan osoitteessa https://lennes.github.io/puheen-annotaatio/. PDF-version voi ladata osoitteesta https://github.com/lennes/puheen-annotaatio/blob/master/annotation_guide.pdf #s3gt_translate_tooltip_mini { display: none !important; }
This is the first release of the <em>Speech Corpus Toolkit for Praat</em> after moving the scripts to GitHub. Identical versions of these scripts were previously provided at http://www.helsinki.fi/~lennes/praat-scripts/. The... more
This is the first release of the <em>Speech Corpus Toolkit for Praat</em> after moving the scripts to GitHub. Identical versions of these scripts were previously provided at http://www.helsinki.fi/~lennes/praat-scripts/. The old site will no longer be maintained. This release is provided as an archive of the original scripts, and some of them may not work in the current version of Praat.
Research Interests:
Pitch analysis tools are used widely in order to measure and to visualize the melodic aspects of speech. The resulting pitch contours can serve various research interests linked with speech prosody, such as intonational phonology,... more
Pitch analysis tools are used widely in order to measure and to visualize the melodic aspects of speech. The resulting pitch contours can serve various research interests linked with speech prosody, such as intonational phonology, interaction in conversation, emotion analysis, language learning and singing. Due to physiological differences and individual habits, speakers tend to differ in their typical pitch ranges. As a consequence, pitch analysis results are not always easy to interpret and to compare among speakers. In this study, we use the Praat program (Boersma & Weenink 2015) for analyzing pitch in samples of conversational Finnish speech and we use the R statistical programming environment (R Core Team, 2014) for further analysis and visualization. We first describe the general shapes of the speaker-specific pitch distributions and see whether and how the distributions vary between individuals. A bootstrapping method is applied to discover the minimal amount of speech that i...
Research Interests:
Research Interests:
Eight vowel qualities are phonologically distinct in Finnish. All of them may occur as either long or short, or they may combine into diphthongs. However, the actual variability of vowel qualities in conversational Finnish speech is... more
Eight vowel qualities are phonologically distinct in Finnish. All of them may occur as either long or short, or they may combine into diphthongs. However, the actual variability of vowel qualities in conversational Finnish speech is unknown. One of the factors affecting this variation is probably the predictability of word forms. The influence of simple word form frequency was addressed in this preliminary study. Five informal Finnish dialogues were recorded and transliterated, and the frequencies of different word forms were obtained from the material. For four speakers, the formant values of F1 and F2 were calculated at the midpoints of vowel segments. Different F1/F2 charts were plotted for vowels according to four word form frequency categories. The results indicate that vowel segments within tokens of common word forms are phonetically more variable than vowels within tokens of rare word forms.
Research Interests:
Research Interests:
Many different research groups in Finland use Finnish speech material as a basis for their studies. Since a general speech database is not available and there exist no common guidelines for collecting and annotating speech material,... more
Many different research groups in Finland use Finnish speech material as a basis for their studies. Since a general speech database is not available and there exist no common guidelines for collecting and annotating speech material, speech corpora end up being compiled in a variety of ways — often for just a single research purpose. This is both inefficient and inhibitory to interdisciplinary cooperation. To improve the situation, a project named "Integrated Resources for Speech Technology and Spoken Language Research in Finland" was initiated. The goal of the project is to design several exemplar and prototypical multimodal speech database systems that can serve as examples for these research groups, build a conforming but extendable infrastructure by selecting, building and refining necessary applications for speech annotation and database access, and to compile guidelines for annotation so that in the future, speech corpora can be shared readily.
Research Interests:
Speech and language researchers need to manage and analyze increasing quantities of material. Various tools are available for various stages of the work, but they often require the researcher to use different interfaces and to convert the... more
Speech and language researchers need to manage and analyze increasing quantities of material. Various tools are available for various stages of the work, but they often require the researcher to use different interfaces and to convert the output from each tool into suitable input for the next one. The Language Bank of Finland (Kielipankki) is developing an on-line platform called Mylly for processing speech and language data in a graphical user interface that integrates different tools into a single workflow. Mylly provides tools and computational resources for processing material and for the inspecting the results. The tools plugged into Mylly include a parser, morphological analyzers, generic finite-state technology, and a speech recognizer. Users can upload data and download any intermediate results in the tool chain. Mylly runs on CSC’s Taito cluster and is an instance of the Chipster platform. Access rights to Mylly are given for academic use. The Language Bank of Finland is a ...
Research Interests:
Tassa raportissa selvitetaan, millaisia suomen kielen aantamisvaikeuksia esiintyy seitsemaa eri lahtokielta edustavilla aikuisilla maahanmuuttajilla. Tiedosta on apua suomen aantamisopetuksessa. Puhujien aidinkielet ovat arabia, kiina,... more
Tassa raportissa selvitetaan, millaisia suomen kielen aantamisvaikeuksia esiintyy seitsemaa eri lahtokielta edustavilla aikuisilla maahanmuuttajilla. Tiedosta on apua suomen aantamisopetuksessa. Puhujien aidinkielet ovat arabia, kiina, somali, tagalog, thai, venaja ja vietnam. Raportissa mukana olevat lahtokielet ovat osa Proof-korpusta, jota varten aanitettiin 72:n paakaupunkiseudulla asuvan 10:ta eri aidinkielta puhuvan maahanmuuttajan lukupuhuntaa ja keskusteluja. Vastaavat puhenaytteet kerattiin myos kontrolliryhmalta, johon kuului 23 syntyperaista suomenkielista. Tassa alustavassa tutkimuksessa kolme tutkijaa kuunteli aaninaytteita ja teki havaintoja ensin itsenaisesti, lopuksi yhdessa. Puhujien lahtokielten todettiin monin tavoin heijastuvan suomen aantamiseen, vaikka yksilollisia eroja oli runsaasti. Yleisesti ottaen suomen konsonantit olivat helppoja lukuunottamattA /ŋ/:aa ja /h/:n allofoneja. Harvinaisimmat vokaalit ja diftongit olivat monille vaikeita. Lahtokielesta riippu...
Research Interests:
Research Interests:
It is often assumed that the participants of a conversation try to avoid simultaneous starts or lengthy silences. For this reason, they may tend to synchronize rhythmically with each other’s speech. A model of conversational turn-taking... more
It is often assumed that the participants of a conversation try to avoid simultaneous starts or lengthy silences. For this reason, they may tend to synchronize rhythmically with each other’s speech. A model of conversational turn-taking based on the idea of coupled oscillators has been suggested by Wilson & Wilson [1]. However, the model has received only weak empirical support from previous studies where distributions of silence durations have been modeled directly. In the present study, we attempt to detect signs of oscillatory behavior during silence utilizing nonparametric hazard regression. In order to understand the shape of the estimated hazard rates, we postulate a latent stochastic process [2] with end of silence occurring when the process crosses a threshold. This finer-grained approach using Bayesian estimation yields a more detailed picture of synchronization between speakers and a more powerful test of oscillatory behavior.
Research Interests:
Research Interests:
Research Interests:
For the Finnish language, the tonal correlates of sen- tence accent or word-internal stress patterns have only been studied in read-aloud laboratory speech. In this preliminary study, the general relationship between pitch patterns and... more
For the Finnish language, the tonal correlates of sen- tence accent or word-internal stress patterns have only been studied in read-aloud laboratory speech. In this preliminary study, the general relationship between pitch patterns and perceived prominence of word-initial syllables is investigated in free con- versational Finnish for one female and one male speaker. The typical pitch levels and distributions are also compared across speakers.
Research Interests:
Tässä raportissa selvitetään, millaisia suomen kielen ääntämisvaikeuksia esiintyy seitsemää eri lähtökieltä edustavilla aikuisilla maahanmuuttajilla. Tiedosta on apua suomen ääntämisopetuksessa. Puhujien äidinkielet ovat arabia, kiina,... more
Tässä raportissa selvitetään, millaisia suomen kielen ääntämisvaikeuksia esiintyy seitsemää eri lähtökieltä edustavilla aikuisilla maahanmuuttajilla. Tiedosta on apua suomen ääntämisopetuksessa. Puhujien äidinkielet ovat arabia, kiina, somali, tagalog, thai, venäjä ja vietnam. Raportissa mukana olevat lähtökielet ovat osa Proof-korpusta, jota varten äänitettiin 72:n pääkaupunkiseudulla asuvan 10:tä eri äidinkieltä puhuvan maahanmuuttajan lukupuhuntaa ja keskusteluja. Vastaavat puhenäytteet kerättiin myös kontrolliryhmältä, johon kuului 23 syntyperäistä suomenkielistä. Tässä alustavassa tutkimuksessa kolme tutkijaa kuunteli ääninäytteitä ja teki havaintoja ensin itsenäisesti, lopuksi yhdessä. Puhujien lähtökielten todettiin monin tavoin heijastuvan suomen ääntämiseen, vaikka yksilöllisiä eroja oli runsaasti. Yleisesti ottaen suomen konsonantit olivat helppoja lukuunottamatta /ŋ/:ää ja /h/:n allofoneja. Harvinaisimmat vokaalit ja diftongit olivat monille vaikeita. Lähtökielestä riippumatta äänteiden kestojen tuottaminen oli monelle vaikeaa, mutta ongelmien luonne vaihteli kielten ja puhujien välillä. Erityisesti /pitkät/ konsonantit lyhenivät. Myös äänteiden fonologisen pituuden ja painotuksen riippumattomuus tuotti ongelmia. Erityisesti painottomien tavujen /pitkät/ vokaalit sekä sananalkuiset painolliset /lyhyet/ vokaalit olivat vaikeita. Poikkeamat eivät aina liity selkeästi yksittäisiin äänteisiin tai tavuihin, vaan ääntämisen kokonaisvaltaisempiin piirteisiin. Esimerkiksi monille aasialaisille puhujille oli tyypillistä kauttaaltaan kireä äänenlaatu.
Research Interests:
Research Interests: Electrophysiology, Phonetics, Cognition, Speech perception, Language, and 21 moreMultidisciplinary, Nature, Information Processing, Finland, Memory, Brain, Humans, Change detection, Estonia, Female, Male, Human Brain, Mismatch Negativity, Magnetics, Native Language, Adult, Auditory evoked Potentials, Auditory Cortex, Left Hemisphere, Part of Speech, and Acoustic Stimulation
Our aim in this paper is to explore ways of modeling the distribution of pause durations in conversation using oscillator models (Wilson, Wilson 2006), and to consider how these models might be integrated into our Coupled Oscillator Model... more
Our aim in this paper is to explore ways of modeling the distribution of pause durations in conversation using oscillator models (Wilson, Wilson 2006), and to consider how these models might be integrated into our Coupled Oscillator Model of speech timing (COM (O’Dell, Lennes, Werner, Nieminen 2007; O’Dell, Lennes, Nieminen 2008; O’Dell, Nieminen 2009)).
Research Interests:
Our exploratory study looks for units of temporal structure in conversational Finnish speech. The relative significance of different hierarchical levels of rhythm was evaluated using Bayesian inference on a linear regression model based... more
Our exploratory study looks for units of temporal structure in conversational Finnish speech. The relative significance of different hierarchical levels of rhythm was evaluated using Bayesian inference on a linear regression model based on coupled oscillators. Results suggest that stress, mora and possibly foot timing as rhythmic factors in Finnish are more relevant than traditionally assumed.
Research Interests:
This study continues our previous investigation into the rhythms contributing to the temporal structure of speech. The relative significance of different hierarchical levels of rhythm was evaluated using Bayesian inference on a linear... more
This study continues our previous investigation into the rhythms contributing to the temporal structure of speech. The relative significance of different hierarchical levels of rhythm was evaluated using Bayesian inference on a linear regression model based on coupled oscillators. Results strengthen our previous conclusions that stress, mora and possibly foot timing are all simultaneously present as rhythmic factors in Finnish conversational speech.
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Most of the phonetic knowledge that is currently available on spoken Finnish is based on clearly pronounced speech: either read-aloud text or other formal speaking styles, such as newscasts or interviews on the radio or on TV. However,... more
Most of the phonetic knowledge that is currently available on spoken Finnish is based on clearly pronounced speech: either read-aloud text or other formal speaking styles, such as newscasts or interviews on the radio or on TV. However, linguistic differences are known to exist ...
Research Interests:
Research Interests:
The aim of the Speech Corpus Toolkit (SpeCT) is to provide an organized collection of well-documented Praat scripts that can be easily downloaded, modified and used in order to perform small tasks during the various stages of building,... more
The aim of the Speech Corpus Toolkit (SpeCT) is to provide an organized collection of well-documented Praat scripts that can be easily downloaded, modified and used in order to perform small tasks during the various stages of building, organizing, annotating, analysing, searching and exporting data from a speech corpus.