Arabic Speech Rhythm Corpus: Read and Spontaneous Speaking Styles

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 5337–5342 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 5337 Arabic Speech Rhythm Corpus: Read and Spontaneous Speaking Styles Omnia Ibrahim 1&2,Homa Asadi 1, Eman Kassem 2, Volker Dellwo 1 1 Institute of Computational Linguistics, University of Zurich, 2 Phonetics and linguistics department, Alexandria University 1 Andreasstrasse 15, 8050 Zürich, Switzerland, 2 Qism Bab Sharqi, Alexandria, Egypt ibrahim@ifi.uzh.ch, eman.qasem@alexu.edu.eg, homa.asadi@cl.uzh.ch, volker.dellwo@uzh.ch Abstract Databases for studying speech rhythm and tempo exist for numerous languages. The present corpus was built to allow comparisons between Arabic speech rhythm and other languages. 10 Egyptian speakers (gender-balanced) produced speech in two different speaking styles (read and spontaneous). The design of the reading task replicates the methodology used in the creation of BonnTempo corpus (BTC). During the spontaneous task, speakers talked freely for more than one minute about their daily life and/or their studies, then they described the directions to come to the university from a famous near location using a map as a visual stimulus. For corpus annotation, the database has been manually and automatically time-labeled, which makes it feasible to perform a quantitative analysis of the rhythm of Arabic in both Modern Standard Arabic (MSA) and Egyptian dialect variety. The database serves as a phonetic resource, which allows researchers to examine various aspects of Arabic supra-segmental features and it can be used for forensic phonetic research, for comparison of different speakers, analyzing variability in different speaking styles, and automatic speech and speaker recognition. Keywords: Speech corpus, Arabic rhythm, Egyptian dialect, stress-timed language 1 Introduction The successful collection of data is a key stage to obtain reliable and valid results in phonetic research. In this paper, we report on work-in-progress about the construction of a speech corpus for Arabic in Egyptian dialect. The primary intention behind designing such a corpus is to provide a homogeneous database of Arabic speech recordings, which investigates broadly/narrowly acoustic parameters of Arabic speech rhythm for forensic voice comparison (FVC) research and casework application. In a typical FVC casework, two samples of voices, a known and an unknown (disputed) sample, are compared to estimate the probability that the same speaker has produced the speech samples (same-speaker hypothesis) versus the probability that the speech samples have come from two different speakers (different-speaker hypothesis) (Rose, 2002). To objectively estimate this probability, having access to speech corpus containing samples from Arabic native speakers that contributes to the knowledge about between-speaker variability and within-speaker variability (Morrison et al., 2012; Kinoshita, 2001). The motivation for exploring speaker-specific acoustic properties of speech of Arab speakers stems precisely from the lack of population statistics for the Arabic language in FVC. As yet, to the best of our knowledge, there is no forensically relevant Arabic speech corpus available being capable of utilizing in FVC research. The current corpus fitted for FVC research aims to fill this gap and it can be considered as the first database of its kind in Arabic …

Content truncated to 500 words