Changes in the channel’s characteristics increase the number of errors during automatic speech recognition. A new method, which uses the wavelet packet transform, is described in this text. This transform is used to approximate the critical bands of human auditive system. RASTA-PLP coefficients are produced from the energy of the approximated critical bands. Test results, using those coefficients, show an improvement of the robustness to channel’s variation when compared to coefficients produced from a Fourier transform filters bank.
| Date | 8 Jan 2010 |
|---|
| Original language | French |
|---|
| Awarding Institution | - École de technologie supérieure
|
|---|
| Supervisor | Gheorghe Marcel Gabrea (Supervisor) |
|---|
Rondeau Beauchamp, M.-A. (Author),
Gabrea (Supervisor),
8 Jan 2010Student thesis: Master's thesis › Master in Engineering: Electrical Engineering