Microsoft Bing Speech Recognition (BingSR) Plugin 1.6.0 to the UniMRCP Server (UMS) has been released.
The plugin is based on the following components:
- UniMRCP Server 1.5.0
- Microsoft Speech WebSocket Protocol
- Libevent 2.1.8
- Rapidjson 1.1.0
The plugin supports the following endpoints:
- Bing Speech
- Speech Service
The binaries are currently available for the following Linux distributions:
- Red Hat / CentOS 7
- Ubuntu 16.04 LTS
This release adds support for new regional Speech Service endpoints by retaining the support for the non-regional Bing Speech endpoint. The release also fixes a few minor issues and allows for new alternate ways of specifying recognition language and referencing speech context.
Everyone is encouraged to upgrade and start gradually migrating to the new Speech Service endpoints. While nothing has to be changed in the configuration or subscription data, check out the new data format described in Section 5 of the corresponding Installation Guide.
The detailed list of changes introduced in this release follows.
New Features
- Added support for new regional Speech Service endpoints. See Section 5 in the Installation Guide.
- Implemented an alternate way of referencing built-in grammars via 'special' ruleref in SRGS XML grammars. See Section 5.1 in the Usage Guide.
- Added support for 'xml:lang' attribute optionally specified in SRGS XML grammars. This method has the highest precedence in selection of recognition language. The language can be otherwise specified either by the global configuration parameter or the header field 'Speech-Language'. See Section 4.2 in the Usage Guide.
- Added a new attribute 'language' to individual speech contexts, which allows to define speech contexts for various languages regardless of the global configuration parameter.
Fixed Problems
- Speech input is considered complete only when an interim result uniquely matches one of the phrases in the used speech context.
- All the internal common class names are declared under a unique namespace not to cause a conflict with other speech recognition plugins being loaded in to the same instance of UniMRCP server.
- In case of DTMF input, allow both 'maxlength' and 'term char' be specified at the same time. Before, 'malxength' was not observed, if 'term char' was specified.
Configuration Parameters
- Added a new attribute 'language' to individual speech contexts.
Miscellaneous
- Fixed the digit format output in a log statement 'Detected Start of Event'.
- Updated the Installation Guides to reflect the changes introduced in this release.
- Updated the Usage Guide to reflect the changes introduced in this release.
Visit the BingSR plugin page for more information.
http://www.unimrcp.org/bingsr
Thank you for using UniMRCP.
--
Arsen Chaloyan
Author of UniMRCP
http://www.unimrcp.org