VDK 4.2.0 - 9th Dec 2022

Summary

We are pleased to introduce VDK 4.2.0 which is a feature and bug-fix release.

In this release we have optimized the integration of voice recognition models.

In addition to the version unification of the previous release, we have locked widgets and binaries updates for a fixed version of the VDK Studio. It means that updates for those packages will become available only once you have updated your VDK Studio to that particular matching version.

Versions

Studio

Component	Version
Studio	4.1.23 UPDATED
Dictionary manager	1.0.5 UPDATED
Free-speech	4.1.4 UPDATED
Grammar editor	4.2.12 UPDATED
Phonetic editor	4.1.3 UPDATED
Project editor	1.17.5 UPDATED
Simple assistant maker	4.0.4 UPDATED
TTS manager	4.2.2 UPDATED
Voice biometrics	4.1.2 UPDATED

VSDK (C++)

Component	Version
Vsdk	7.1.4 UPDATED
Vsdk-vasr	3.1.1 UPDATED
Vsdk-csdk	3.1.6 UPDATED
Vsdk-tnl	2.0.7 UPDATED
Vsdk-baratinoo	3.0.6 UPDATED
Vsdk-vtapi	2.1.5 UPDATED
Vsdk-idvoice	3.1.4 UPDATED
Vsdk-tssv	2.0.7 UPDATED

VSDK (JAVA)

Component	Version
Vsdk	4.6.1UPDATED
Vsdk-csdk	4.5.1
Vsdk-tnl	2.5.0
Vsdk-baratinoo	3.3.0
Vsdk-vtapi	2.3.2
Vsdk-idvoice	2.4.0
Vsdk-tssv	2.5.0
Vsdk-vasr	2.0.2 UPDATED

Details

Studio

Bug-fixes:

Searching in the resources view with no results do not uncheck filters anymore

Dictionary manager

Bug-fixes:

File paths now use native directory separator

Free-speech

Bug-fixes:

Some translations were missing

Grammar editor

Improvements:

Quick testing now takes into account the model’s advanced settings if any

Bug-fixes:

File paths now use native directory separator

Phonetic editor

Bug-fixes:

Use a cancel button instead of abort when closing the widget
File paths now use native directory separator

Project editor

Bug-fixes:

Deleting a voice biometrics model could in some cases result in a crash of the Studio
File paths now use native directory separator
Exporting ASR technologies now only process checked ones
Some ASR recognizers had missing languages when exported

TTS manager

Improvements:

Can now export generated audio as raw file or wave file

Voice biometrics

Improvements:

Can now import multiple audio files
The training’s progress bar now displays unit of recordings
The user’s name now gets cleared when added

Bug-fixes:

File paths now use native directory separator

VSDK (C++)

Vsdk

Bug-fixes:

Some minor compilation warnings

Improvements:

Internal improvements

Vsdk-vasr

Improvements:

Performances are greatly increased on all platforms

VSDK (Android)

Vsdk

Improvements:

Add ASR result json parser