Javad Sheykhzadegan
Abstract
In this paper, the efficiency of Persian speech phonemes from the point of view of efficiency in speaker recognition has been studied, and then with due attention to efficiencies, the ...
Read More
In this paper, the efficiency of Persian speech phonemes from the point of view of efficiency in speaker recognition has been studied, and then with due attention to efficiencies, the ranking of phonemes has been done. For estimating the efficiencies of phonemes, we have introduced one criterion that has been defined in the form of phonemes “Inter speaker distance” to “Intra speaker distance” ratio, referred to as “Speaker Affectability Ratio: SAR”. The necessary experiments and computations have been done for all Persian speech phonemes (with the exception of /À/) using the Persian speech database “Farsdat” and then on the basis of the results of these experiments and computations, the ranking of single phonemes and phoneme groups has been done. The results have shown that in the phoneme groups ranking, vowels and semi-vowels are first, nasals, fricatives and liquids are second and, obstructions and plosives are third from the point of view of efficiency in speaker recognition. Likewise, the ranking of single phonemes shows that the phoneme /∂/ is first and the phoneme /t/ is last from the point of view of efficiency in speaker recognition. The results of this research in line with research on other spoken languages such as English, Germanic and Dutch show high agreement for phoneme groups ranking but noticeable differences in details of rankings are also noted.