2010 Fiscal Year Final Research Report
A new speech coding system based on ML-BEATS
| Project/Area Number |
20760229
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Communication/Network engineering
|
|---|
| Research Institution | The University of Tokushima |
|---|
Principal Investigator |
SUZUKI Motoyuki The University of Tokushima, 大学院・ソシオテクノサイエンス研究部, 准教授 (30282015)
|
|---|
| Research Collaborator |
ADACHI Masashi 徳島大学, 大学院・先端技術科学教育部システム創成工学専攻, 博士前期課程学生
|
|---|
| Project Period (FY) |
2008 – 2010
|
| Keywords | ML-BEATS法 / 音声符号化 / セグメント量子化 / HMM |
|---|
| Research Abstract |
In this research, a new speech coding system based on ML-BEATS has been developed. ML-BEATS method can split an input sequence into many sub-sequences and cluster these. This method has been applied to segment quantization. Experimental results showed the proposed method decreased bit rate, but increased spectral distortion compared with the G.729.
|
