Towards an Automatic Speech-to-Text Transcription System: Amazigh Language (original) (raw)
Related papers
Comparative Analysis of Arabic Vowels using Formants and an Automatic Speech Recognition System
Arabic, the world’s second most spoken language in terms of number of speakers, has not received much attention from the traditional speech processing research community. This study is specifically concerned with the analysis of vowels in modern standard Arabic dialect. The first and second formant values in these vowels are investigated and the differences and similarities between the vowels explored using consonant-vowels-consonant (CVC) utterances. For this purpose, a Hidden Markov Model (HMM) based recognizer is built toclassify the vowels and the performance of the recognizer analyzed to help understand the similarities and dissimilarities between the phonetic features of vowels. The vowels are also analyzed in both time and frequency domains, and the consistent findings of the analysis are expected to enable future Arabic speech processing tasks such as vowel and speech recognition and classification.
Bulletin of Electrical Engineering and Informatics, 2023
Vowels are the primary units of a sound system of a language. The classification of these vowels is therefore very important for the recognition and synthesis of speech. In this paper, we propose a normalized energy-based approach in formants and pitch to characterize Arabic vowels (short vowels: / a /, / i /, / u /; long vowels: / a: /, / i: /, / u: /). The classification was performed using a developed algorithm on records extracted from an Arabic corpus after the extraction of the pitch and the first three formants and the computation of the normalized energy in these bands. The results showed that the algorithm distinguishes Arabic vowels by analyzing the normalized energy in the nucleus of F1, F2, and F3 formants and pitch F0 with a rate of 88.7% for long vowels and a rate of 90% for short vowels.