Nalazite se na CroRIS probnoj okolini. Ovdje evidentirani podaci neće biti pohranjeni u Informacijskom sustavu znanosti RH. Ako je ovo greška, CroRIS produkcijskoj okolini moguće je pristupi putem poveznice www.croris.hr
izvor podataka: crosbi !

Naive listener’s perception: who is better in recognition? (CROSBI ID 679021)

Prilog sa skupa u zborniku | sažetak izlaganja sa skupa | međunarodna recenzija

Biočina, Zdravka Naive listener’s perception: who is better in recognition? // Proceedings of 28th Conference of the International Association for Forensic Phonetics and Acoustics / Önder Gürpinar, Burcu (ur.). Istanbul, 2019. str. 24-25

Podaci o odgovornosti

Biočina, Zdravka

engleski

Naive listener’s perception: who is better in recognition?

Local speeches of the island of Brač (Croatia) can be divided into four language groups: a) central Čakavian settlements, b) west Čakavian settlements, c) east Čakavian settlements with Štokavian influence and d) east Štokavian settlement Sumartin (Sujoldžić et al., 1988 ; Šimunović, 2006) (Figure 1). The aim of this paper was to examine the ability of young naive listeners to distinguish speeches of the island of Brač in relation to this diversity. Therefore, the first goal was to gather a speech corpus of a greater number of native speakers. The spontaneous spoken utterance of 150 speakers was recorded with a high-quality recorder and microphone. Fifteen speakers from each of the ten selected settlements (Bol, Gornji Humac, Milna, Nerežišća, Postira, Pražnica, Pučišća, Selca, Sumartin, and Supetar) were recorded and divided into groups according to age and gender. These settlements were chosen to represent all larger Čakavian, Cakavian and Štokavian-Čakavian speeches of the island of Brač. The criteria for selecting the speakers were the following: they had to be originally from the settlement, they had to live there for the last ten years, and both of their parents had to be from the settlement. Based on five minutes of spontaneous speech of each speaker, 23–second samples were created and played randomly for recognition. Shorter samples were chosen to reflect the forensic scenario. All toponyms and other indicators that could reveal the origin of the speaker were cut out from the samples. Apart from speakers from Brač, two speakers from Split (Čakavian – Štokavian speech) were also included in the recognition, to examine if the listeners distinguish the speech of the city nearby from the speeches of Brač. The listeners (N=81) were native speakers from Brač, who were born and live there as well. They were students from two high schools on the island of Brač that are located in Supetar and Bol. The recognition test had in total 304 questions. The listeners needed to decide if the speaker is from Brač or Split (1st question). If they chose Brač, they had to circle the exact settlement (2nd question). Results show that the listeners were more successful in detecting whether speech is from Brač or Split than placing the speaker in the exact place of origin. They have detected for almost all speeches with more than 80% of accuracy that they are from Brač, which shows that they differentiate between speeches from Brač and that from Split. The only exception is the speech of Sumartin, which was recognized to be from Brač in only 60.04% of cases. Brač was detected best in the speech from Pučišća which is also the speech with the most accurate detections (56.33%) in this experiment. That's because it has the most prominent markers in the vowel systems of speeches of Brač - diphthongs [ie] and [uo]. The second-best recognized speech is the speech of Pražnica (53.01%) that is accurately detected due to its specific tonal intonation acoustically described in Biočina, Bašić, and Varošanec-Škarić (2018). The results of multiple regression analysis have shown that students from the high school in Bol, senior high-school students, students from the vocational secondary school and listeners with both parents from Brač were more successful in recognizing the speaker’s origin. Better performance of these groups of students can be explained by their greater geographical mobility and numerous lin-guistic contacts, which enhance recognition based on existing studies (Nolan, 2012 ; Jacewicz, 2016). Regression analysis showed that only gender was not a significant predictor, i.e. female and male students were equal in the detection. In the recognition, they relied mostly on the segmental level and the average basic tone of the voice, so they recognized better speakers with phonetic markers in their speech and higher values of F0. Considering the growing importance of naive listeners in forensics and LADO (Köster, Schiller, and Künzel, 1995 ; Köster and Schiller, 1997 ; Cambier-Langeveld, 2010 ; Foulkes and Wilson, 2011 ; Nolan, 2012), this research can also contribute to a better understanding of their abilities.

sociophonetics ; forensic phonetics ; speech recognition ; naive listeners ; speeches of the island of Brač

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

Podaci o prilogu

24-25.

2019.

objavljeno

Podaci o matičnoj publikaciji

Podaci o skupu

28th Annual Conference of the International Association for Forensic Phonetics and Acoustics (IAFPA 2019)

predavanje

14.07.2019-17.07.2019

Istanbul, Turska

Povezanost rada

Filologija

Poveznice