Dissertation > Excellent graduate degree dissertation topics show

DSP Realization of Research on Speech Enhancement Algorithm in High Noise Environment

Author: WangZuoLin
Tutor: TianXueLong
School: Chongqing University
Course: Biomedical Engineering
Keywords: Speech enhancement Adaptive filtering Variable step size LMS Tongue-like curve De-correlation
CLC: TN912.35
Type: Master's thesis
Year: 2013
Downloads: 95
Quote: 0
Read: Download Dissertation

Abstract


Noise seriously affects the quality of speech signal under complex environment,leading us cannot convey the semantic correctly, so speech enhancement processingbecomes very necessary, and has already become a hot research topic on speech signalprocessing at home and abroad. At present, there are lots of algorithms on speechenhancement technology, and the noise cancellation technology based on adaptivefiltering technique has become a widely used speech enhancement method for its strongself-learning characteristic and excellent tracking performance. However, the traditionaladaptive filtering technology exist problems as follows: poor adaptability, slowconvergence speed when the input signals were heavily correlated, which limit itsdevelopment on a certain degree. In order to solve this problem, an improved speechenhancement algorithm was proposed after analysis the principle and performance ofthe standard least mean square (LMS) algorithm, it uses adaptive filter theory andcomprehensives the advantages of variable step size LMS algorithm and de-correlationprinciple. The new algorithm uses tongue-like curve to update the step size of adaptivefiltering, at the same time it also introduces the de-correlation principle for speechde-noising. Moreover, compared the algorithm proposed in this paper with otheralgorithms on traceability, computational complexity and speech enhancement effect,which proved that the algorithm proposed in this paper can improve the adaptability andconvergence rate and reduces the steady-state error of traditional algorithm.Considering the hardware implementation, this article also takes the DSP chip withthe high speed frame processing ability to be the control core of the system, meanwhile,uses its multi-channel buffered serial port (McBSP) to connect with the extended audiointerface chip TLV320AIC23, which can realize seamless connection. So thecomplexity of the system design is simplified, also it makes the system with high speedacquisition、de-noising processing and output.At the same time, de-noising effect was test with different types of speech signalsbased on the DSP system. The final experimental results show that, the signal-to-noise(SNR) rate is improved significantly with the new algorithm. At the meantime, thelistening test also indicates that the voice has become clearer and intelligibility has been improved greatly after the system process, that is to say the purpose to separate theuseful speech is realized.

Related Dissertations

  1. The Speech Enhancement System Based on Binary Mask and Perceptual Wavelet Packet Transform,TN912.35
  2. Design of Degital Effect Processor by DSP Platform,TN912.2
  3. Research on Speech Enhancement Method Based on Spherical Microphone Array Beamforming,TN912.35
  4. Research on Vermicularity of Ductile Iron Based on Machine Learning,TG143.5
  5. Single channel speech enhancement algorithm,TN912.35
  6. Multi- directional tube array speech enhancement system development,TN912.35
  7. Noisy environments Microphone Array Speech Enhancement Technology,TN912.35
  8. Complex environment speech enhancement method,TN912.35
  9. Research on airborne noise reduction technology in digital audio,TN912.2
  10. Research on High-Precision GPS Data Processing and Integer Ambiguity Search Algorithm,P228.4
  11. Analysis and Calibration of Three-axis Fluxgate Magnetometers Error,TH762.3
  12. Study on Control Strategy and Error Compensation Method for Atomic Force Microscopy Positioning System,TH742
  13. Research of Active Noise Cancellation Earphone Algorithms,TN911.7
  14. The Rearch on the Technology of MELPe Speech Coder,TN912.3
  15. Research and Implementation on Single-Channel Speech Enhancement Algorithm in Digital Hearing Aid Chips,TN912.35
  16. Research of Adjustment Theory Based on Extended Correlation Least Mean Square Algorithm,TN713
  17. Comparative Analysis of Ambiguity Resolving Method Based on LAMBDA Method,P228.4
  18. Research on Integrated Navigation System of High Altitude / Long Endurance Uavs Base on the Astronomy Sensor,V249.328
  19. Research on Subspace-Based Speech Enhancement,TN912.35
  20. Research on Speech Enhancement Method at Low Signal-to-Noise Ratio,TN912.35
  21. Based on DSP acoustic echo cancellation system,TN911.7

CLC: > Industrial Technology > Radio electronics, telecommunications technology > Communicate > Electro-acoustic technology and speech signal processing > Speech Signal Processing > Speech Enhancement
© 2012 www.DissertationTopic.Net  Mobile