Release notes 3.2.0

Release date: 21st July 2022

Summary

This page highlights all the changes, new features and bugs addressed within the LumenVox Containers version 3.2.0 release. This change affects both Speech and Voice biometric products for general release.

Highlights

1. MRCP connection interface into speech products

2. New DNN engine implemented for Active voice biometric use cases

3. Active voice biometrics integrated into the new speech ASR DNN engine for text validation

4. Licensing services updated to ensure accurate license counting

5. Bug fixes identified in version 3.1.0 are addressed

6. Web portal enhancements

What’s new on LumenVox Cloud 3.2.0  

New Features

• MRCP connection interface into speech products

• Integrated new DNN active voice biometrics engine

• Integrated new ASR into active voice biometrics for phase validation

• New Admin Portal added for cluster administrators to manage deployments

• New Deployment Portal added for deployment (tenant) managers to manage settings, etc.

Updates

• Improvements have been made to the following:

  • Improved intra-service reconnect strategy

  • Added Prometheus and Health probe endpoints and handlers to ASR

  • Implemented new confidence scoring algorithm that results in more predictable and expected ASR confidence score results

Bug fixes

Various including issues with the following:

• Unsuccessful ASR grammar loads did not trigger final response correctly

• Fixed issue related to ASR grammar handling of certain UTF-8 encoded characters

• Fix minor issues related to running diagnostics running in admin portal

• Fixed “Edit Deployment” handling and validation in admin portal

• Fixed issue where SSML requests in TTS were not resolving referenced audio URI

Installation notes

The following helm chart can be used

Helm Chart

Note that for MRCP there is no helm chart but a docker compose file. MRCP will run on its own Docker virtual machine which will integrate into the Kubernetes cluster. 

Upgrade procedures

Upgrade or migration from versions 1.x ,version 2.x are not supported. Version 3.2.0 must be a fresh install. Version 3.1.0 can be upgraded to version 3.2.0 by updating the protobuf files, all existing grammar cache files must be deleted prior to upgrade.

Updated API guide

Speech: LumenVox Speech API

Voice biometrics: LumenVox API Documentation

Further information can be obtained here:

LumenVox Containers API documentation

NOTES: Voice biometric API changes

The following changes have been made to the voice biometric APIs

• SampleTooLong has been added as a possible output in the statusHint table when utterances are longer than the configured maxUtteranceDuration

• audioFormat in the response has been renamed to “SampleEncoding”

• SpeechScore and SpeechQualityScore are currently not displayed in the result

• the enrollment and verification repoinse contains 2 additional fields, AsrTranscript and AsrConfidence e.g. 

• Init enrollment and init verification are now requiring a new parameter called audioFormat

• Audio files used for voice biometrics should now be headerless