Codebook excited linear predictive coding
In standard Codebook Exited Linear Coding (CELP) the encoding method uses random values for excitation vectors. The excitation is modelled by the codebook. Each time a segment of speech is encoded and an excitation vector is matched to minimise the error between original and encoded signal to maint...
| Main Authors: | , |
|---|---|
| Format: | Book Chapter |
| Language: | English |
| Published: |
IIUM Press
2011
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/29350/ http://irep.iium.edu.my/29350/ http://irep.iium.edu.my/29350/1/Codebook_Excited_Linear.pdf |
| Summary: | In standard Codebook Exited Linear Coding (CELP) the encoding method uses random values for excitation vectors. The excitation is modelled by the codebook. Each time a segment of speech is encoded and an excitation vector is matched to minimise the error between original and
encoded signal to maintain quality of speech. This chapter narrates the CELP encoding algorithm. The conceptual block diagram of twoO time varying filters and a Gaussian codebook
shown in figures, fc encoder and decoder respectively. This simplified block diagram mainly consists ofthree blocks, Short-Term Prediction, Long-Term Prediction and a random code-book. The parameters of these predictors SIP, LIP and the codebook are optimised and estimated in many ways. If these estimated parameters are accurate, the synthesized speech \vilI sound the same as original speech. Because ofthe limitations ofthe coding build blocks the estimated filter parameters case as the estimation errors in a result speech quality suffers degradations. The standard CELP algorithm explained gradually as follows and its block diagram shown in figure |
|---|