OpenAI SS Plugin 1.3.0 Released

[Arsen Chaloyan]

OpenAI Speech Synthesis (SS) Plugin 1.3.0 to the UniMRCP Server (UMS) has been released.

The plugin is based on the following components:

• OpenAI Text-to-Speech REST API
• Libevent 2.1.9
• Rapidjson 1.1.0

The plugin is available for the following UMS versions:

• 1.7.0
• 1.8.0

The binaries are currently available for the following Linux distributions:

• Red Hat 8
  • unimrcp-openai-ss-1.7.3-1.el8.x86_64.rpm
  • unimrcp-openai-ss-1.8.3-1.el8.x86_64.rpm
• Red Hat 9
  • unimrcp-openai-ss-1.7.3-1.el9.x86_64.rpm
  • unimrcp-openai-ss-1.8.3-1.el9.x86_64.rpm
• Ubuntu 22.04 LTS
  • unimrcp-openai-ss_1.7.3-jammy_amd64.deb
  • unimrcp-openai-ss_1.8.3-jammy_amd64.deb
• Ubuntu 24.04 LTS
  • unimrcp-openai-ss_1.7.3-noble_amd64.deb
  • unimrcp-openai-ss_1.8.3-noble_amd64.deb

This release adds support for flatting supplied SSML content into plain text segments.

The detailed list of changes introduced in this release follows.

New Features

• Since the OpenAI TTS API does not currently support SSML, allow to flatten the received SSML content and pass plain text segments to the service endpoint for speech synthesis.

Fixed Problems

• None.

Configuration Parameters

• Added a new attribute 'flatten-ssml' to the element 'synth-settings'. Defaults to 'true'.

Miscellaneous

• Updated the Usage Guide to reflect the changes introduced in this release.

Visit the OpenAI SS plugin page for more information.

• https://unispeech.io/openaiss
  
• https://www.unimrcp.org/openaiss
  

Thank you for using UniMRCP.

--

Arsen Chaloyan
Author of UniMRCP
http://www.unimrcp.org