I am interested in various kinds of phonetic variation, including dialectal variation in Swedish and other languages, paralinguistic variation like age, emotion, attitude and health condition, and also phonetic variation in animal vocalisations. Specifically, I take interest in the following areas:
CAT VOCALISATIONS (meows, moans, purrs, chirrups, growls, hisses, etc.)
• Phonetic analysis of vocalisations in domestic cats is my way of combining two of my favorite interests: cats and phonetics! In 2010 I began to record and analyse different cat vocalisations to learn more about how domestic cats communicate vocally with humans as well as with other cats. Based on previous research (mainly Moelk, 1944 and McKinley, 1982), I am currently trying to update and further develop the typology of cat vocalisations. In the project Melody in Human–cat commmunication, I study how the prosody in both human speech and cat vocalisations influence our communication. Specifically, I am interested in the following question:
• How do cats perceive human speech?
• How do humans perceive cat vocalisations? Do we use the same cues as when we perceive human speech? I have studied how human listeners perceive intonation and emotion in cat vocalisations.
• What vocalisation types can cats produce in different situations, what characterizes the different types, and how do they vary?
•Estonian Swedish is still spoken by a handful of speakers in Estonia and Sweden. Although nearly extinct, there are still no recordings of Estonian Swedish that have been adapted for phonetic studies. I would like to help document the last native speakers of Estonian Swedish and analyse this unique variety of Swedish phonetically.
•Vowel articulation in Swedish dialects (see Articulography).
•The accent distribution of compound words in Scanian (skånska) dialects. Being a native Scanian speaker, I have noticed that my accent distribution in compound words is changing towards a slightly more standard pattern (which is always accent II in compound words). I would like to investigate this distribution further, and find out why accent I compounds are becoming increasingly rare in the Scanian dialects.
•The Swedish allophone Viby-i and how it is produced in different dialects. I am currently doing a pilot study of Viby-i in the Humanities Lab with the articulograph AG 500.
• Using electromagnetic articulography (EMA), we are studying speech movements of Swedish vowels using articulography in the research project Exotic Vowels in Swedish: an articulographic study of palatal vowels [VOKART].
• In the project FDA Analysis of Speech Movements (part of the the Linnaeus centre for Cognition, Communication and Learning (CCL)), we investigated lip movement variability as a function of age.
• Speaker age is a paralinguistic and speaker-specific quality which is always present in speech. I am interested in human perception, acoustic analysis, machine recognition and simulation (using speech synthesis) of speaker age, Speaker age is also the theme of my doctoral thesis. Some sound examples of my experiments with synthesis of speaker age can be found here (coming soon).
• More knowledge about dialectal variation can be used to improve speech technology applications, such as speech recognition and speech synthesis. In the research projects SweDia 2000 and SIMULEKT, I have been given excellent opportunities to pursue my interest in Swedish dialects and dialectal prosody, especially intonation in regional varieties of Swedish.
• I would like to contribute to the research in the area of human-machine communication, (1) by improving automatic recognition of speaker-specific qualities in speech, such as age, dialect (or accent), emotion and physiological state, and (2) by making synthetic speech sound more natural.
• Other paralinguistic and nonlinguistic features - I would like to use similar methods that I have used in my experiments with speaker age in studies of other paralinguistic and speaker-specific properties in speech – such as emotions, attitudes and health state– in order to further increase the naturalness of synthetic speech.