Release notes 3.1.0

Release date: 6th August 2022

Summary

This page highlights all the changes, new features and bugs addressed within the LumenVox Containers version 3.1.0 release. This change affects both Speech and Voice biometric products however voice biometric changes are not for general release.

Highlights

 

  1. Update of the ASR engine providing improved accuracy and performance

  2. Included are the following ASR languages: Dutch; English: All dialects including US, UK, Indian, Australian, and NZ; French: All dialects including European and Canadian; German; Italian; Portuguese: All dialects including European and Brazilian; Russian; Spanish: All dialects including Castilian (European), Mexican, and Latin American; Swiss German.

  3. Optimized audio streaming methods added

  4. Improvements to container multi-tenancy

  5. An admin portal is now available for cluster administrators to manage deployments, configurations and run diagnostic tests

  6. Various bugs identified in version 3.0.0 are resolved

  7. Integration of new DNN engine for voice biometrics and vb containers moved into speech stack

  8. Refactoring Biometric engine configuration

  9. Database configuration changes & updates to config handler

  10. Binary storage changes to support data archiving

  11. License service updates 

  12.  Improvements to session manager and cleanup of session data

     

Whatโ€™s new on LumenVox Cloud 3.1.0  

New Features

  • Optimized audio streaming methods added

  • New administration web portal is now available for cluster administrators to manage deployments, configurations and run diagnostics

Updates

  • Improvements have been made to the following:

    • ASR engine performance and accuracy

    • Multi-tenancy functionality

    • The new DNN engine for voice biometrics and vb containers have been moved into the speech stack

    • Biometric engine configurations have been refactored

    • Database configuration has been changed & updates made to the configuration handler

    • Binary storage changes made to support data archiving

    • License service updates 

    •  Improvements made to session manager and cleanup of session data. This should assist with the cleanup of redis data

Bug fixes

Various including issues with the following:

  • UTF-8 encoding issues resolved

  • SessionManager crash with RedisClient         

  • Error response messages were standardized     

  • stream_start_time was not configurable

Installation notes

The following helm chart can be used 

Helm Chart

Note that cache must be deleted

Upgrade procedures

Upgrade or migration from versions 1.x ,version 2.x  are not supported. Version 3.1.0 must be a fresh install. Version 3.0.0 can be upgraded to version 3.1.0 by updating the protobuf files, all existing grammar cache files must be deleted prior to upgrade.

NOTES

Certain API names have also changed:

  • AudioSteamCreate is now called AudioCreate

  • AudioStream is new and used to stream audio into system

  • AudioStreamPush is now called AudioPush

  • AudioStreamPull is now called AudioPull

  • VADMessage.audio_stream_offset is now called VADMessage.audio_offset

ASR output 

  • Note that the ASR output has changed from uppercase to lowercase.

Updated API guide

LumenVox Speech API

 

Further information can be obtained here:

LumenVox Containers API documentation



Was this article helpful?
Copyright (C) 2001-2024, Ai Software, LLC d/b/a LumenVox