VDK 5.3.0 - 4th August 2023

Summary

We are pleased to introduce VDK 5.3.0 which is a feature release.

When using the Voice Synthesis technology, you can now retrieve more information about what is being generated, such as the text markers which will allow you, for example, to display spoken words as they are read.

When several SDKs are available for the same technology, you can now switch from one another to get the appropriate one for your target language.

C++ Samples are now less confusing to use with the Conan package manager.

Other fixes are described below, under relevant components.

Versions

Studio

Component	Version
Studio	5.3.1 UPDATED
Dictionary manager	2.0.0
Free-speech	5.0.0
Grammar editor	5.0.1UPDATED
Phonetic editor	5.0.0
Project editor	2.1.4 UPDATED
Simple assistant maker	5.0.2 UPDATED
TTS manager	5.0.1
Voice biometrics	4.3.2 UPDATED
Nlu Editor	2.0.2 UPDATED
Vsdk-daemon	1.11.0 UPDATED

Samples

Component	Version
Chained grammars	5.0.1 UPDATED
Dynamic grammar	5.0.0 UPDATED
Simple application	5.0.0 UPDATED
Tts	5.0.1 UPDATED
Voice biometrics	5.1.0 UPDATED
Voice Commands Language Understanding	2.0.0 UPDATED

VSDK (C++)

Component	Version
Vsdk	8.1.0 UPDATED
Vsdk-vasr	5.0.0 UPDATED
Vsdk-csdk	4.0.3 UPDATED
Vsdk-tnl	3.0.1 UPDATED
Vsdk-baratinoo	4.0.0 UPDATED
Vsdk-vtapi	3.0.0 UPDATED
Vsdk-idvoice	4.1.0 UPDATED
Vsdk-tssv	3.1.0 UPDATED
Vsdk-vnlu	2.0.0 UPDATED
Vsdk-audio-portaudio	3.3.0 UPDATED

VSDK (JAVA)

Component	Version
Vsdk	5.1.0 UPDATED
Vsdk-csdk	5.1.0 UPDATED
Vsdk-tnl	2.5.0
Vsdk-baratinoo	4.0.0 UPDATED
Vsdk-vtapi	3.1.0 UPDATED
Vsdk-idvoice	2.4.0
Vsdk-tssv	2.5.0
Vsdk-vasr	2.1.3 UPDATED
Vsdk-vnlu	1.0.1

Details

Studio

Removed:

Embedded tutorials in favor of the online documentation

Grammar Editor

Bug-fixes:

Model names containing . were wrongly handled and could lead to weird name modification after compilation

Project Editor

Bug-fixes:

Vsdk-vtapi’s configuration generation only applied for the last selected voice
Exporting several NLU models would result in only exporting the last one

Simple Assistant Maker

Bug-fixes:

The script directory was still placed according to the old directory hierarchy and could not be used anymore

Voice Biometrics

Bug-fixes:

Enrolling a new user while recording could cause conflicts

Nlu Editor

Bug-fixes:

The placeholder could sometimes show even when some intents were already created

Vsdk Daemon

Bug-fixes:

The Freespeech could sometimes use the wrong engine upon switching languages

VSDK (C++)

Vsdk

Improvements:

Offer a way to retrieve more information about what is being played for the voice synthesis technology
Voice syntheses are now implemented in a more consistent way with the others technologies, by subscription
Voice syntheses can now be used in an asynchronous way

Vsdk-vasr

Improvements:

Now supports the gapless recognition between two different recognition result to be able to speak directly after saying for example a wake up word

Vsdk-audio-portaudio

Improvements:

Now contains a new PaMicrophoner producer and PaPlayer consumer that supports the play/pause features

Vsdk-csdk, Vsdk-vtapi, Vsdk-baratinoo

Improvements:

Implement the new voice synthesis features for more audio played information

VSDK (Java)

Vsdk

Improvements:

Offer a way to retrieve more information about what is being played for the voice synthesis technology

Vsdk-csdk, Vsdk-vtapi, Vsdk-baratinoo

Improvements:

Implement the new voice synthesis features for more audio played information