IEEE Global Telecommunications Conference, 1989, and Exhibition. 'Communications Technology for the 1990s and Beyond
DOI: 10.1109/glocom.1989.64153
|View full text |Cite
|
Sign up to set email alerts
|

Backward adaptation for low delay vector excitation coding of speech at 16 kbit/s

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
10
0

Publication Types

Select...
4
2
1

Relationship

2
5

Authors

Journals

citations
Cited by 24 publications
(10 citation statements)
references
References 7 publications
0
10
0
Order By: Relevance
“…For codebook based systems like LD-CELP and LD-VXC, the WMSE is computed for each entry in the excitation codebook and M is the vector dimension. Codebook design algorithms are discussed in [7,83. For tree coding, all possible path map sequences to depth M are considered for computing the reconstructed speech y ( n ) and the path giving the minimum distortion is chosen.…”
Section: Excitation Sequence Optimization the Weighted Mean-square Errormentioning
confidence: 99%
See 3 more Smart Citations
“…For codebook based systems like LD-CELP and LD-VXC, the WMSE is computed for each entry in the excitation codebook and M is the vector dimension. Codebook design algorithms are discussed in [7,83. For tree coding, all possible path map sequences to depth M are considered for computing the reconstructed speech y ( n ) and the path giving the minimum distortion is chosen.…”
Section: Excitation Sequence Optimization the Weighted Mean-square Errormentioning
confidence: 99%
“…The update of the coefficients k r ) may be done using the LMS algorithm with leakage factors for improving noisy channel performance. However, a significant reduction in the computational complexity may be obtained by using the so-called sign algorithm [8] which avoids the division operations required by the standard LMS algorithm: kp+')-, -(l-p)kr)+ hsign(ej(n))sign(rj(n-l)) (7) where l<$.IM. A good choice for the constants is…”
Section: Fig 2 Lattice Short-term Predictormentioning
confidence: 99%
See 2 more Smart Citations
“…The recursive adaptive algorithms used to update the predictor coefficients are similar to those used in the CCITT 32 kb/s algorithm [5,71. This solution leads to relatively low computational complexity; however, it has the disadvantage of a rough spectral modeling of the speech signal.…”
Section: Introductionmentioning
confidence: 99%