Improving Speech Signal Quality in Meeting Room using FIR Filter

  • Aekkarat Suksukont นักศึกษา หลักสูตรวิศวกรรมศาสตรมหาบัณฑิต ภาควิชาวิศวกรรมอิเล็กทรอนิกส์และโทรคมนาคม คณะวิศวกรรมศาสตร์ มหาวิทยาลัยเทคโนโลยีราชมงคลธัญบุรี
  • Jakkree Srinonchat ผู้ช่วยศาสตราจารย์ ภาควิชาวิศวกรรมอิเล็กทรอนิกส์และโทรคมนาคม คณะวิศวกรรมศาสตร์ มหาวิทยาลัยเทคโนโลยี ราชมงคลธัญบุรี
Keywords: Speech processing, Fast fourier transform, Finite impulse response filter

Abstract

This research proposes an improving the speech signal quality in the meeting room using the Fast Fourier Transform (FFT) with Finite Impulse Response (FIR) filter. In the experiment, speech signal is recorded from 54 locations in the same meeting room, which these speech is generated from the same source. The band pass filter is then designed by using FIR filter for filtering the distort frequencies and compare to the original speech in each position. The experiment results show that the FFT technique, which is used to analysis of the speech frequencies, provides more accuracy when uses the frame-to-frame comparison. Moreover, it shown that the lower 500 Hz and higher 2.5 KHz are the error bandwidth for the low and high frequency respectively. Which designed by using FIR filter, is then used to comparison and adjustment these speech signal with the original speech signal. It modifies these speech frequency to similar to original speech approximately 70.77%.

 

References

1. Woudenberg E, Frank K, Soong BH. Juan. A block least squares approach to acoustic echo cancellation. International Conference on Acoustics Speech and Signal Processing; 1999 Mar 15-19; Phoenix, USA; 2002.

2. Ogawa A, Aikawa N, Sato M. A design method of low delay FIR band pass filters. International Symposium on Circuits and Systems (ISCAS2000); 2000 May 28-31; Geneva, Switzerland; 2000.

3. Kodrasi I, Doclo S. Joint derverberation and noise reduction based on acoustic multichannel equalization. 14th International Workshop on Acoustic Signal Enhancement (IWAENC 2014); 2016 Apr.

4. Pandey V. Equiripple band pass FIR filter design for speech signals order optimization for frequency range of 300 Hz to 4000 Hz. International Conference on Computing Communication and Automation (ICCCA 2015); 2015 May 15-16; Noida, India; 2015.

5. Suksukont A, Srinonchat J. Frequency Analysis of Speech and Resonance Speech using FFT. Proceedings of the 8th Conference of Electrical Engineering Network of Rajamangala University of Technology (EENET 2016); 2016 May 25-27; Patong, Phuket, Thailand; 2016.

6. Vairetti G, Holdt S, Jensen, Sena ED, Moonen M, et al. Multi-channel identification of room acoustic system with adaptive filters based on orthonormal basic functions. International Conference on Acoustics Speech and Signal Processing (ICASSP 2016); 2016 Mar 20-25; Shanghai, China; 2016.

7. Paleologu C, Benesty J, Ciochin S. A variable step-size affine projection algorithm designed for acoustic echo cancellation. Transaction on Audio Speech and Language Processing. 2008; 16-18.

8. Kruthiventi S, Srinivas S, Prahallad K. An FIR implementation of zero frequency filtering of speech signals. Transactions on Audio Speech and Language Processing. 2012; 20-29.

9. Suksukont A, Srinonchat J. Improving the quality of the speech signal using a FIR band pass filter with Fast Fourier Transform. 5th International Electrical Engineering Congress (IEECON 2017); 2017 March 8-10; Chonburi, Thailand; 2017.
Published
2018-12-20
Section
บทความวิจัย