Unchanged high-quality Built-in pocket-sized speech synthesis engine  microAITalk®3

micr2

Unchanged high-quality Built-in pocket-sized speech synthesis engine

microAITalk® is the minimized version of the human-like, natural, high-quality speech synthesis AITalk® built to integrate into devices. This previous version has been updated with additional functions as microAITalk®3.
Provision type Equipment Integration Main usage Built-in Equipment/ Navigation and PND/ Games/ Smartphone Application
Provision type
Equipment Integration
Main usage
Built-in Equipment/ Navigation and PND/ Games/ Smartphone Application

Introduction of Functions

機能1
Document free text-to-speech

High quality speech synthesis based on human voice

機能2
Voice effect tag function

Specifying various types of parameter values and adding bookmarks is now possible within Japanese texts.

機能3
Bookmark function

You can place a bookmark in any place in the document. You can find information about the bookmark in the callback function.

機能4
Ruby function

You may now set the way of reading using the ruby function when same kanji characters have a different way of reading.

機能5
Loading and unloading word dictionary

You can load the word dictionary by setting the file path. Also, you can unload, then load a new word dictionary.

機能6
Low-volume and low-resource

Implementation of low-volume and low-resources, making it possible to integrate with various devices

機能7
Input Support

Input supports both a mixture of kanji and kana Japanese text and middle notation (AI special standard, JEITA TT-6004 standard)

機能8
User dictionary function

The user dictionary function registers and saves names of people and places that are read in special ways.

機能9
Porting

Porting available for various OS

microAITalk® Application Examples

Configuration

Supported Platforms

OS CPU
Android 4 or later ARM, Intel
iOS8 or later Apple A5 or later

Please inquire us for more details on the Windows Embedded CE 6.0 ver. and Linux ver.

Basic specification table

Analysis in Japanese available

Normal Version Compact Version
Memory (RAM) Over 40MB (In the case of 1 speaker)
Approx. an additional 5MB per addition of 1 speaker
More than 30 – 35 MB (In the case of 1 speaker)
Approx. an additional 1.5MB per addition of 1 speaker
Data Size(ROM) Japanese Dictionary:11.0MB
Program: 2.0MB
Voice sound dictionary: 40MB
Approx. an additional 40MB per addition of 1 speaker
Japanese Dictionary: 2.8 MB
Program: 2.0MB
Voice sound dictionary: 10MB
Approx. an additional 10MB per addition of 1 speaker
Output Data Format 16KHz 16bit Linear PCM  Monaural

Please inquire us for more details on the Windows Embedded CE 6.0 ver. and Linux ver.

Analysis in Japanese not available

Normal Version Compact Version
Memory (RAM) Over 26MB (In the case of 1 speaker)
Approx. an additional 5MB per addition of 1 speaker
More than 24MB (In the case of 1 speaker)
Approx. an additional 1.5MB per addition of 1 speaker
Data Size (ROM) Program: 2.0MB
Program: 2.0MB
Voice sound dictionary: 40MB
Approx. an additional 40MB per addition of 1 speaker
Program: 2.0MB
Voice sound dictionary: 10MB
Approx. an additional 10MB per addition of 1 speaker
Output Data Format 16KHz 16bit Linear PCM  Monaural

*1 Japanese analytical engine uses theBerkeley Software Distribution
A licensed software (Chasen) is used based on License (BSD License)
* Attachment tool is offered only to devices compatible with Windows.

Product Introduction


Inquiry Form

 

 

TelephoneInquiry