Tajuk projek ini adalah Pengekod Suara Bit Rendah Menggunakan Pemproses Isyarat Digit TMS320C6416 (Low Bit Rate Speech Coding Using TMS320C6416 DSP Processor). Skop projek ini boleh di bahagikan kepada dua bahagian. Bahagian pertama melibatkan pengenalan kepada pemproses TMS320C6416. Bahagian ke dua projek ini melibatkan pensampelan isyarat suara pada frekuensi pensampelan yang berlainan dan mengenalpasti perbezaan di antara kualiti suara yang dihasilkan selepas pensampelan. Pada mulanya, Matlab dan Simulink digunakan untuk proses pensampelan dan mengenalpasti kesan frekuensi pensampelan terhadap kualiti isyarat yang dihasilkan. Kemudian, proses pensampelan diteruskan dengan menggunakan pemproses TMS320C6416. Tiga frekuensi pensampel di pilih iaitu 8000 Hz, 4000 Hz dan 2000 Hz. Bahagian keputusan di bahagikan kepada dua bahagian iaitu sebelum masa nyata dan selepas masa nyata (menggunakan TMS320C6416). Perbandingan di buat di antara kualiti isyarat suara yang dihasilkan pada frekuensi pensampelan yang berlainan. Dua kaedah telah digunakan untuk mengukur kualiti isyarat yang dihasilkan. Kaedah pertama ialah dengan meminta 15 orang pelajar memberikan skor mereka terhadap kualiti isyarat yang dihasilkan. Julat skor yang dipilih ialah di antara 1(sangat tidak memuaskan) hingga 5(sangat bagus). Kaedah ke dua ialah dengan menggunakan osiloskop untuk memaparkan bentuk isyarat yang dihasilkan dan membandingkannya dengan isyarat asal. Keputusan memunjukkan bahawa kualiti isyarat berkurangan dari 8000 Hz ke 2000 Hz.
_________________________________________________________________________________________
The title of the project is Low Bit Rate Speech Coding Using TMS320C6416 DSP Processor. The scope of this project is divided into two main parts. Part one involves the study of the TMS320C6416 DSP processor. My task was to understand the architecture of this board and complete the tutorials in Code Composer Studio (CCS). The second part is concerned with the sampling of speech signal (analog signal) at different sampling frequencies and to study its effects on the quality of the reconstructed speech signal. Initially MATLAB and SIMULINK were used to sample the speech file and to study the effect of variation in sampling frequency on the quality of the speech signal and its waveform. Later, the sampling process is implemented in real time using the TMS320C6416 DSP Processor. Three sampling frequencies were chosen which are 8000 Hz, 4000 Hz and 2000 Hz. The results were divided into two sections; before real-time implementation and after real-time implementation. The comparison of the quality of the sampled audio signal was carried out for the three sampling frequencies as mentioned earlier. Two methods were used to measure the quality of the reconstructed audio signal. First, fifteen students were chosen to rate their score for the quality of the reconstructed signal. The score range was from 1(bad) to 5(excellent). Secondly, scope was used to display the waveform of the original and reconstructed signal. The results showed that the quality of the sound degrades from 8000 Hz to 2000 Hz.
Low bit rate speech coding using tms320c6416_Mohamad Habib Mahamad Haniffah_E3_2005_NI