(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Design and implementation of suboptimal and optimal linear-phase digital fir filters for audio signal processing

Design and implementation of suboptimal and optimal linear-phase digital fir filters for audio signal processing / Ko Wei Jie
Secara umumnya, istilah audio dirujuk sebagai bunyi, terutamanya apabila bunyi itu direkod, dihasil semula, dihantar atau disimpan. Isyarat audio boleh diwakili dalam digital atau analog. Fokus dalam pemprosesan isyarat audio kebanyakannya ialah analisis matematik yang sebahagian daripada isyarat yang boleh didengar. Dalam projek ini, isyarat audio berada dalam julat frekuensi di antara 300 - 4000 Hz yang merupakan isyarat suara manusia. Salah satu masalah dalam aplikasi isyarat audio ialah penambahan isyarat bunyi bising yang mengganggu isyarat asal dan menyebabkan kualiti audio menurun. Sebagai penyelesaian, penapis audio seperti penapis analog dan penapis digital telah dibina dalam pemprosesan isyarat audio. Dalam projek ini, penapis FIR digital digunakan kerana ia mempunyai fasa linear, kestabilan yang baik, ketepatan dan fleksibiliti yang tinggi. Penapis FIR digital fasa-linear suboptimum dan optimum akan direka dengan menggunakan kaedah yang paling popular iaitu kaedah tetingkap (kaedah Hanning) dan algoritma Parks-McClellan dalam perisian dengan menggunakan Matlab dan dalam perkakasan pada TMS320VC5510 DSK. Fungsi reka bentuk penapis juga disahkan dengan pendengaran, dan diluluskan dengan merujuk kepada nilai SNR dan spektrum kuasa isyarat audio dalam simulasi. Analisis perbandingan kecekapan penapis untuk penapis FIR suboptimum dan optimum telah dibuat. Melalui analisis ini, tetingkap Hanning penapis FIR didapati mempunyai prestasi yang sama baik dengan Parks-McClellan penapis FIR walaupun urutan penapisnya adalah 2.67 kali lebih tinggi. _______________________________________________________________________________________________________ Generally, the term audio refers to sound, particularly when such sound is recorded, reproduced, transmitted or stored. The audio signal can be represented in digital or analog. The focus in audio signal processing is topically a mathematical analysis of which part of the signal is audible. In this project, the audio signal is in the frequency ranges of between 300 – 4000 Hz which is the human voice. One of the problems in audio signal applications is that the addition of noise signal that interfere with the original signal, result in poor quality of audio. In order to solve it, audio filters such as analog filter and digital filter have been developed in audio signal processing. In this project, digital FIR filter is employed as it has linear phase, good stability, high accuracy and flexibility. Suboptimal and optimal linear-phase digital FIR bandpass filters will be designed using the most popular window (Hanning) method and Parks-McClellan algorithm respectively in software using Matlab and hardware on TMS320VC5510 DSK. The functionality of the filter designs is also verified audibly, and are justified with the SNR values and power spectra of the audio signals in simulation. A comparative analysis of the filter competency of the suboptimal and optimal FIR filters is made. Through the analysis, the window Hanning FIR bandpass filter is found to have equally good performance with the PM FIR bandpass filter, despite the filter order is 2.67 times higher than the latter filter design.
Contributor(s):
Ko Wei Jie - Author
Primary Item Type:
Final Year Project
Identifiers:
Accession Number : 875005805
Language:
English
Subject Keywords:
audio; sound; frequency
First presented to the public:
6/1/2015
Original Publication Date:
3/7/2019
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 129
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2019-03-07 12:35:28.629
Submitter:
Mohd Jasnizam Mohd Salleh

All Versions

Thumbnail Name Version Created Date
Design and implementation of suboptimal and optimal linear-phase digital fir filters for audio signal processing1 2019-03-07 12:35:28.629