Note: Your custom schedule will not be saved unless you create a new account or login to an existing account.
| Paper: | SS-1.1 |
| Session: | Multilinguality in Speech Processing |
| Time: | Tuesday, May 18, 13:00 - 13:17 |
| Presentation: |
Special Session Lecture |
| Topic: |
Special Sessions: Multilinguality in Speech Processing |
| Title: |
MULTILINGUAL ACOUSTIC MODELS FOR SPEECH RECOGNITION AND SYNTHESIS |
| Authors: |
Siegfried Kunzmann; IBM Pervasive Computing | | |
| | Volker Fischer; IBM Pervasive Computing | | |
| | Jorge Gonzalez; IBM Pervasive Computing | | |
| | Ossama Emam; IBM Pervasive Computing | | |
| | Carsten Günther; IBM Pervasive Computing | | |
| | Eric Janke; IBM Pervasive Computing | | |
| Abstract: |
In this paper we review the design of a common phone alphabet forup to fifteen languages and describe its application in twoimportant components of a seamless multilingual conversationalsystem, namely speech recognition and synthesis. We report onexperiments that demonstrate the advantages of multilingualacoustic models both for the recognition of foreign names andnon-native speech, and describe the usefulness of a commonphone alphabet for the construction of unit selection based mono-and bilingual speech synthesis |
| |
| Back | |