Microsoft Azure Speech Recognition (SR) Plugin 1.17.0 to the UniMRCP Server (UMS) has been released.
The plugin is based on the following components:
- UniMRCP Server 1.7.0
- Microsoft Speech WebSocket Protocol
- Libevent 2.1.9
- Rapidjson 1.1.0
The plugin supports the following endpoints:
- Speech Service (regional)
- Custom Speech Service (regional)
- Speech Service (containerized)
- Custom Speech Service (containerized)
The binaries are currently available for the following Linux distributions:
- Red Hat / CentOS 7 (unimrcp-azure-sr-1.7.9-1.el7.x86_64.rpm)
- Red Hat / CentOS 8 (unimrcp-azure-sr-1.7.9-1.el7.x86_64.rpm)
- Ubuntu 16.04 LTS (unimrcp-azure-sr_1.7.9-xenial_amd64.deb)
- Ubuntu 18.04 LTS (unimrcp-azure-sr_1.7.9-bionic_amd64.deb)
This release provides a few fixes and enhancements to the existing functionality.
The detailed list of changes introduced in this release follows.
- Reworked processing of STOP requests to reliably release all the associated resources and respond to the client regardless of the WS connection and/or request state.
- Reconnect, when the WS connection is unexpectedly closed by the service after RECOGNIZE is received but before streaming is started.
- If an active recognition request is stopped, trailing audio data could remain in the buffer, if the resampler is used, and be processed with the next recognition request placed in the scope of the same MRCP session.
- The statement "Failed to Connect to Service" is now logged only when a connection to the service endpoint cannot be established, but not when connection establishment is cancelled due to processing of a STOP request.
Visit the Azure SR plugin page for more information.
Thank you for using UniMRCP.
Author of UniMRCP