ریتم گفتار

This study investigates rhythmic patterns in Persian-English bilingual speech, focusing on duration-based measures in a sample of late bilingual adult males. Using various rhythmic measures, including consonantal and vocalic duration, we explored cross-linguistic differences, individual consistency, and speaker identification potential. The results revealed significant differences between Persian (L1) and English (L2), particularly in vocalic measures, whereas consonantal measures exhibited greater consistency. Cross-linguistic correlations were stronger for consonantal measures than for vocalic measures, suggesting higher individual consistency in consonant timing. Speaker identification, conducted through linear discriminant analysis, achieved the highest accuracy with consonantal measures, with stronger performance in L1 than in L2. These findings indicate that while bilingual speakers adjust their rhythmic patterns to suit L2 demands, they retain individual characteristics, especially in consonant timing. This research has implications for understanding bilingual speech production and enhancing speaker recognition technology.

Scientific study of speech rhythm for more than eight decades proves that it is not only an important part of music and poetry, but has become a challenging subject in all branches of linguistics. Moreover, acoustic study of speech rhythm has provided practical results in various topics, including speaker identification. On the other hand, some studies indicate that speech rhythm varies greatly between languages, dialects and speakers. They concluded that the between-speaker variations in speech rhythm depends on the linguistic discriminations in the sentences. In this study, in addition to examining Qashqai speech rhythm, the effect of each sentence on the rhythmic features of this linguistic dialect was investigated. To this end, the acoustic durational rhythmic correlations of Qashqai in a reading text were examined by calculating different measures of phonetic intervals including segmental intervals, vowel and consonant intervals, vocalic and intervocalic intervals, voiced and unvoiced intervals, syllable intervals, and syllable peaks intervals. Furthermore, in order to study the effect of each sentence on Qashqai rhythmic features, between-sentence speech variability was considered. The results showed that Qashqai rhythm was placed in the middle the rhythm continuum of languages close to the syllable-time languages; however, in comparison to Persian language it was closer to the stress-time languages. The results of durational rhythmic measures at different phonetic intervals demonstrated that the Syllabic normalize Pairwise Variability Index (nPVI-Syl), the standard deviation of Variation Coefficient in consonant intervals (VarcoC) and the Syllable rate (rateSyl) were respectively the robust measures in identifying between-sentence variabilities in Qashqai speech rhythm.

Rhythmic characteristics of speech based on consonantal and vocalic intervals as well as syllabic intervals vary between speakers of the same language. Nonetheless, the rhythmicity of a speech signal is not solely dependent on the durational variability of phonetic intervals but it is also associated with the variability of the intensity patterns as well. Acoustic parameter of intensity is largely determined by the articulatory behaviors of the speech organs such as lip movement or mouth aperture. Therefore, it is plausible that speaker idiosyncrasy in movement of speech articulators and anatomical differences in individual’s vocal tracts may influence the energy distribution across a speech signal which subsequently leads to the variability in the values of the intensity measures. Using experimental phonetics tools and from an explicitly speaker-specific perspective, the present research attempts to explore potential speaker-specific acoustic parameters of speech rhythm which are extracted from the intensity contours across Persian speakers. This research aims to discover whether intensity-based measures of speech rhythm are able to discriminate between speakers in Persian. Two types of acoustic rhythmic measures based on the mean syllable intensity (stdevM, varcoM, rPVIm, nPVIm) and peak syllable intensity (stdevP, varcoP, rPVIp, nPVIp))  were selected for this study. Speech data from 12 Persian male speakers were recorded non-contemporaneously in laboratory environment on two different occasions separated by one to two weeks. Speech tokens were acoustically measured with PRAAT version 5.2.34 and statistical analyses were carried out with SPSS version 21 and R version 3.3.3. Results of the study indicated that speech rhythm measures based on intensity fluctuations play an important role in between-speaker rhythmic variability. In addition, discriminatory power of intensity-based measures is not affected by the language-dependent characteristics of Persian. The results also showed that the peak syllable intensity measures carry more speaker-specific information compared to the mean syllable intensity measures

Using the auditory-acoustic approach, the present study examines the possibility of using durational acoustic parameters of speech rhythm for detecting the Persian speakers’ use of Azeri as a form of voice disguise. To do so, continuous speech of 5 speakers of Standard Persian and 5 speakers of Azeri, Tabrizi variety, were chosen for the acoustic and statistical analysis after completing the validity procedures. All Persian speakers were monolingual and all Azeri speakers were bilingual speakers of Azeri and Persian, who spoke Azeri as their mother tongue and Persian as their second language. Each Persian speaker was asked to narrate a lifetime experience once in Persian (Persian- Persian data), and once as an imitation of Azeri (Persian- Azeri data). Azeri speakers were also asked to narrate a lifetime experience once in Azeri (Azeri- Azeri data) and once in Persian (Azeri-Persian Data). Persian-Azeri data is the type that is considered as the disguised data in this survey. The recorded data were then annotated in five tiers: segment, CV-segment, CV-segment interval, CV-interval and syllable. In order to control the effect of any unwanted variable, one minute (±5 seconds SD) of each sound file was extracted for further acoustic and statistical analyses. A Praat script, DurationAnalyzer, was used to automatically calculate the acoustic correlates of durational parameters of speech rhythm. These parameters are: %V (the proportion which speech is vocalic), ΔC (ln) (standard deviation of the natural-log normalized duration of consonantal intervals), ΔV (ln) (standard deviation of the natural-log normalized duration of vocalic intervals), nPVI- V (rate-normalized averaged durational differences between consecutive vocalic intervals) and syllable rate.  Results revealed there was a significant difference between the proposed types of data and that%V and syllable rate best discriminated between them; however, none of the above-mentioned parameters were significantly different between Persian-Azeri and Azeri-Persian data.

The durational variability of phonetic intervals is considered as one of the properties of speech rhythm. These intervals include segmental, vowel, consonantal, vocalic, intervocalic, voiced, unvoiced, syllable, and syllable peak intervals. The durational variability measure for some of these intervals, such as vowel, consonantal, vocalic, intervocalic intervals, determines the classification of languages based on their rhythm. Besides, in some cases, the speaker identification is only possible through the person's voice. The segmental and suprasegmental properties of a language can be used to identify the speaker. In this study, the acoustic correlations of Persian speech rhythm in a reading text are calculated by various durational measures. Also, the between-speaker rhythmic variability is considered to find the best rhythmic measures for Persian speaker identification. The results confirm that Persian is near to the syllable-based languages. Moreover, the results from segmental and supra-segmental consideration demonstrate significant between-speaker variability in Persian. Among phonetic intervals, nPVI-VC and V% (percentage of vocalic intervals) best discriminate between-speaker variability in Persian.