|
ReaderDemo Application
The ReaderDemo application is a simple speech synthesizer that demonstrates the features of JSAPI. The application presents five modes of operation:
- Read (no JSML): Reads out some plain text with no formatting.
- Read (JSML): Reads out text formatted with JSML. In this case, the text is read in a more natural way.
- Features Test (JSML): You will get a feel of most of the features of JSML in this mode including volume, speech rate, pitch, etc.
- Test (no JSML): In this mode, you can enter some plain text and see how the engine generates speech. Enter additional spaces, new lines, periods, commas, and check the differences they make in speech generation.
- Test (JSML): In this mode, you can enter JSML formatted text and test how the output varies (click here for more information on the JSML syntax).
The application demonstrates, among other things, the use of speech events and event listeners.
To Run the demo applications use java speechdemo.TellTime and java speechdemo.ReaderDemo.
You can also run the demos using FreeTTS. For this, you need to put the speech.properties file (loactedin the FreeTTS package) in either the lib folder of your 'java.home' (the location of your JRE) or your 'user.home' (usually "c:\documents and settings\username" on Windows 2000 and XP machines). To run the above demos in FreeTTS, use the TellTimeFTTS and ReaderDemoFTTS Java classes. Unfortunately, FreeTTS does not render JSMLit reads all the JSML as plain text.
Speech Engines
Speech Technology is implemented as software, hardware, or a combination of both. When implemented using software, speech applications never directly interact with the audio components of a computer. Instead, they are abstracted from the underlying hardware through the speech engine. This ensures that speech applications are not tied down to specific hardware implementations and are portable across all platforms.
The speech engine is the heart of any speech-based application. It takes speech input or speech output and converts the speech to a standard format so that speech applications can process them and produce desired results. Speech synthesizers, speech recognizers, speaker verification systems, speaker identification systems, etc. are examples of speech engines.
Each of the systems below is a specialized speech engine that does some predefined processing with the speech input or output:
- Speech Synthesis: This engine handles conversion of textual input to synthetic spoken output. This is often referred to as "text-to-speech" (TTS) conversion.
- Speech Recognition: This engine performs the conversion of spoken input to digital output, such as text. Note that speech recognition does not mean understanding speech.
- Speaker Identification: This engine identifies a person by the sound of their voice. Identified is performed by comparing the person's voice against existing voices in a voice database.
- Speaker Verification: This engine authenticates a person based on their voice.
The basic functionality of speech engines is abstracted in the classes and interfaces of the javax.speech package. However, specific speech engine functionalities can be added using additional packages. The Java speech API provides built-in support for the synthesizer and recognizer speech engines. The additional classes which support these functionalities are defined in the javax.speech.synthesis and javax.speech.recognition packages.
Speech Synthesis
A speech synthesizer converts text to speech. The synthesizer supports two different formats of feeding text and getting speech output. The first format is plain text (which is self-explanatory) and the second format uses the Java Speech Markup Language (JSML).
- The Plain Text Format: When input is fed to the synthesizer as plain text, the speech rendered is machine-like, without any emotion attached to it. Moreover, plain text synthesis does not distinguish and recognize known notations like date and time. For instance, 12:15PM is read as "twelve colon fifteen PM" which is obviously not how a user would expect to hear the text. To improve the readability of the synthesizer, the speech API supports annotations, which are embedded in the speech text. These annotations are supported through JSML.
- The JSMLFormat: JSML is an XML-based markup language that allows you to embed structural and presentational information within the text. Synthesizers interpret these annotations and convert the text to speech appropriately. For instance, the annotation <SAYAS class=time> 12:15PM </SAYAS> in JSML will make sure the synthesizer interprets the time as twelve fifteen PM.
New on the Java Boutique:
New Review:
Time Management Made Easy with the Quartz Enterprise Job Scheduler
Why not just use the Java timer API? This open source scheduling
API boasts simplicity, ease-of-integration, a well-rounded feature
set, and it's free!
New Applet:
Reverse Complement
Reverse Complement is a simple applet that converts DNA or RNA
sequences into three useful formats.
Elsewhere on internet.com:
WebDeveloper Java
Lots of Java information on webdeveloper.com
WDVL Java
Thorough Java resource at the Web Developer's Virtual Library.
ScriptSearch Java
Hundreds of free Java code files to download.
jGuru: Your View of the Java Universe
Customizable portal with online training, FAQs, regular news updates, and tutorials.
|