Examples

Several new examples have been added to the examples.synthesis and examples.recognition packages to demonstrate the IO features of the com.cloudgarden.audio package. Also, new example packages, examples.applet and examples.remote, have been added. Please consult these examples for usage of the classes. Also, and since it was found that the javax.sound.sampled package can sometimes hang a system (mainly when trying to open a TargetDataLine and SourceDataLine at the same time), it might save time to use the code from some of the examples as a starting point for new code.

Running the Examples

The examples can be run:

• using the batch files supplied with the implementation, of the form TestXXX.bat - eg. TestHelloWorld.bat, or
• from a console window by typing (assuming Java 2 is being used) "java -cp .;cgjsapi.jar examples.synthesis.HelloWorld" etc. from the same directory as the "examples" folder is in.
• by running the TestExamples.bat script - this starts up a Java app which lets you run 12 of the test demos with output going to a scrolling Java window - useful for those poor soles working on Windows 98 or Me boxes who want to see more than 50 lines of text.
• from the Java installer supplied with the implementation - see the "Using the Java Installer" part of the Installation section. After it extracts the jar file it shows twelve buttons which run twelve examples and show the output inside a scrolling Java frame.
• The Applet examples can be run in a browser, provided the JSAPI files are installed correctly (see the applets page).

Description of the Examples

Some of the examples are described here - all of the examples are described briefly in the javadoc.

• examples.userinterface
• examples.userinterface.VoicePad - A sample Swing-based application which uses the Mouth component and a JEditorPane to display a text document and highlight words currently being spoken. When the "Speak" button is pressed it will start speaking either the text currently selected, or will start speaking from the current cursor position if no text is selected. Also allows synthesized speech to be saved to a MPEG file.
• examples.userinterface.TestInterfaces - Displays a GUI with buttons which pop-up the Recognition and Synthesis engines' native user interfaces.
• examples.applet
• BrowserApplet - open the browser.html file in IE or Netscape for a simple voice-enabled browsing experience.
• SpeechApplet - a simple TTS applet.
• examples.audio
• Contains examples demonstrating the com.cloudgarden.audio package.
• ConversionTest - demonstrates the AudioFormatConverter's capabilities, converting from 8 to 16kHz with and without linear interpolation.
• FileToFileToLine - Demonstrates copying an mp3 file to a gsm file, then playing the gsm file to an AudioLineSink.
• FileToFiles - Demonstrates reading from one file and writing to two different ones with different formats. Uses the AudioMediaURLSource, AudioSplitter, AudioFormatConverter classes.
• WAVtoMPEG - Demonstrates reading from a wave file and writing to an mp3 file.
• examples.recognition
• examples.recognition.DictationTest - Tests recognition of phrases from the full vocabulary - after it has recognized three phrases it will stop.
• examples.recognition.WhatTimeIsIt - Tests recognition of 'what time is it' and 'what date is it' and responds with a spoken  answer
• examples.recognition.LoadJSGFFromURL - Tests recognition of phrases loaded from the JSGF file ../examples/grammars/helloWorld.gram - example phrases are 'pause and finish and stop', 'I'd like five bananas','I'd like twenty five bananas and twelve oranges', 'hello my name is <your name here>', 'open the bottle please' - after it has recognized three phrases it will stop.
• examples.recognition.DictationFromFile - Audio data read in from WAV file and sent to a Recognizer
• examples.recognition.DictationFromMPEGFile - Audio data read in from MP3 file and sent to a Recognizer
• examples.recognition.ParserTest - Demonstrates use of the RuleGrammar.parse method
• examples.recognition.GramCommitTest - Tests activating rules and grammars in response to spoken commands. One grammar contains the words "alpha","bravo","charlie","delta", but only one rule is active at a time and the other grammar contains "one","two","three","four" - again, only one rule active at a time. Also, only one of the grammars is active at a time. Switch between grammars with the command "switch" and activate the next rule in each grammar with the command "next".
• examples.remote
• FileToLineRemote - Demonstrates network transmission of audio data from an AudioFileSource to a AudioLineSink in compressed (GSM) and uncompressed (RAW) formats, from server to client and from client to server - for ease of demonstration incorporates both server and client since both run on the localhost.
• contains several examples demonstrating transfer of audio speech data between a server machine running a Recognizer/Synthesizer and a client machine without speech engines. Start up the Server on the server machine and start the Client on the client machine with a command-line argument of the server machine's network name (or no argument if Client and Server are on the same machine).
• RemoteAudioServer and RemoteAudioClient - doesn't use speech engines - just transfers audio data from a client machine to a server machine where it is played to a SourceDataLine.
• RemoteDictationServer and RemoteDictationClient- sends audio data from the client to the server machine running a Recognizer with a DictationGrammar.
• RemoteSynthesisServer and RemoteSynthesisClient - sends synthesized speech to a client machine.
• WhatTimeIsItServer and WhatTimeIsItClient audio data "what time is it" sent from client to server and synthesized speech sent from server to client.
• examples.rtp
• CaptureAndPlay - This class opens two RTP steams - one receives audio data from the url "rtp://<localhost>:12346/audio" and plays it to the local output device, and the other sends audio data (captured from the local audio capture device) to the url "rtp://<localhost>:12344/audio". This class should generally be started before running any of the other classes in this package, since the other classes either send data to the url "rtp://<localhost>:12346/audio" or receive data from the url "rtp://<localhost>:12346/audio", or do both.
• DictationFromRTP - allocates a recognizer, listening on an RTP port, and sends the contents of a WAV file to the RTP port the recognizer is listening on.
• FileToRTP - Plays file to RTP stream (doesn't use JSAPI - it's just an RTP demo) - to hear output stream, start up CaptureAndPlay class, then start this class
• RTPCommandVoiceServer - Creates a Recognizer which listens to an incoming RTP stream and replies with synthesized speech on a separate RTP stream - to send the RTP stream to the recognizer and hear it's response, start up the CaptureAndPlay class, then this class, then say "what date/time is it", or "goodbye computer".
• RTPDictationVoiceServer - Creates a Recognizer which listens to an incoming RTP stream and replies with synthesized speech on a separate RTP stream - to send the RTP stream to the recognizer and hear it's response, start up the CaptureAndPlay class, then speak into the microphone - the recognizer should respond with what it thinks you said.
• SpeakToRTP - Uses a synthesizer to send audio data to an rtp url. To hear the output, run the CaptureAndPlay class then run this example.
• examples.synthesis
• examples.synthesis.HelloWorld - Tests basic speech synthesis with a 'Hello World', and demonstrates use of the SpeechEngineChooser and Mouth classes.
• examples.synthesis.DuplexTest - Tests whether a duplex sound card is being used. Keep saying one of the six commands "Nice day", "Hello", "How are you" ,"What time is it","Would you like a cup of tea" or "Goodbye" while the computer is replying - the computer should hear what you said while it was still talking and reply when it has finished it's current reply.
• examples.synthesis.SynthesizerTest - Tests more complex speech synthesis
• examples.synthesis.SpeakToFile - Tests saving of synthesized speech to two WAV files in the extracted directory
• examples.synthesis.SpeakToMPEGFile - Tests saving of synthesized speech to a MP3 file in the extracted directory
• examples.synthesis.ReadAliceToWAVEFile - Tests saving of synthesized "Alice in Wonderland, Chapter 1" to a WAVE file in the extracted directory - for SAPI4 engines you will need to allocate about 128MB of memory for the JVM with the "-Xmx128m" java command-line option.
• examples.synthesis.ReadAliceToMPEGFile - as ReadAliceToWAVEFile but to a (about 3MB) MPEG file.
• examples.synthesis.SpeakToLine - Uses the com.cloudgarden.audio classes to send audio to an AudioLineSink.
• examples.synthesis.SpeakToMultiLine - Uses the com.cloudgarden.audio classes to demonstrate splitting of synthesizer output to three separate AudioLineSink s.
• examples.ListEngineTest
• Lists all the available Synthesizers (for all languages) and their Voices, says "hello" with each of the Voices, then lists all the Recognizers and SpeakerProfiles.
• examples.vocab.VocabTest
• Demonstrates changing the pronunciation of "time" to "aw r" (hour) for both speech synthesis and recognition.
• examples.vocab.Pronunciations
• Demonstrates the addition of a word (with pronunciation) to the VocabManager, and by using JSML tags in the speech string, and oddities of the SAPI4 engines.