Speech To Text Conversion

Abstract

ABSTRACT:

Though Arabic language is a widely spoken language, research done in the area of Arabic Speech Recognition is limited when compared to other similar languages. This paper concerns with convert Arabic spoken word into text using Mel-frequency Cepstrum Coefficient (MFCC) and Vector Quantization (VQ). This has been realized by first recording teachers’s voices for each word in a noisy environment. Secondly these words have been used to extract their features using the Mel Frequency Cepstral Coefficients (MFCC) technique which are taken as input data to the Vector Quantization to construct codeword for each word . Finally ,in the conversion stage each codeword was indexed with the corresponding text. The system targeting deaf students to help them solve some of the problems which face them in the university environment. The system Word Error Rate was 20%.