{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

Ch5-Speech&amp;Audio_Coding_Examples

# Ch5-Speech&amp;Audio_Coding_Examples - Speech Audio...

This preview shows pages 1–9. Sign up to view the full content.

Click to edit Master subtitle style  6/3/09  Veton K ë puska Speech & Audio Processing Speech & Audio Coding Examples

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
6/3/09  Veton K ë puska 22 A Simple Speech Coder u LPC Based Analysis Structure Pre- emphasis Windowing Analysis Auto- Correlation Levinson- Durbin Linear Prediction Analysis Audio Input Analysis Filter Residual Filter  Coeffs Residual Filter  Coeffs Quantization
6/3/09  Veton K ë puska 33 Windowing Analysis Stage N – Length of the  Analysis Window 10-30 msec

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
6/3/09  Veton K ë puska 44 Some Analysis Windows
6/3/09  Veton K ë puska 55 MATLAB Useful Functions u wintool n Use “doc wintool” for more information u window n Use “>doc window” for the list of supported windows u Define your own window if needed e.g: n Sine window and Vorbis window

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
6/3/09  Veton K ë puska 66 LPC Analysis Stage u LPC Method Described in: n Ch5-Analysis_&_Synthesis_of_Pole- Zero_Speech_Models.ppt u Summary: n Perform Autocorrelation n Solve system of equations with Durbin- Levinson Method u MATLAB help
6/3/09  Veton K ë puska 77 Example of MATLAB Code function myLPCCodec(wavfile, N) % % wavfile - input MS wav file % N - LPC Filter Order % ŝ [n] [ ] [ ] [ ] = + - = p k k n ge k n s n s 1 ˆ ˆ α Impulse Train Noise Generator Voiced Unvoiced Vocal Tract Model Gain

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
6/3/09  Veton K ë puska 88 Analysis of Quantization Errors u Use MATLAB functions to research the effects of  quantization errors introduced by precision of the arithmetic  operations and representation of the filter and error signal: n Double (float64) representation (software emulation) n Float (float32) representation (software emulation) n Int (int32) representation (hardware emulation) n Short (int16) representation (hardware emulation).
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}