IBM WebSphere Voice Toolkit
V6.1
Readme
August 2008
(C) Copyright IBM Corporation 2001, 2008. All Rights Reserved.
Note to U.S. Government Users Restricted Rights - Use,
duplication or disclosure restricted by GSA ADP Schedule
Contract with IBM Corp.
Note: Before using this information and the product it supports, read the
general information under Notices
in this document.
Contents
This document contains the following topics for the IBM(R) WebSphere(R) Voice Toolkit:
Introduction
This readme provides information on known problems and software limitations for the WebSphere Voice Toolkit.
For installation information, see the IBM WebSphere Voice Toolkit V6.1 Installation Readme, located in <Eclipse_install_dir>\eclipse\plugins\com.ibm.voicetools.doc_6.1.0
Getting Started
To launch the WebSphere Voice Toolkit, start the Eclipse workbench into which the toolkit has been installed.
For information on using the toolkit, refer to the online Help.
From the Help menu, click Help Contents.
When the Help window appears, select Developing Voice applications from the list.
For additional information about the samples included with the toolkit, see
the sample readme files that are located in each of the sample directories
in <Eclipse_install_dir>\eclipse\plugins\com.ibm.voicetools.samples_6.1.0
- Developing Voice applications
In addition to the toolkit help topics, this package contains related
publications in PDF format. These documents are available from the Help > Help Contents pane.
To view or print PDF files, use Adobe Reader, Version 4.0.5 or later. For
optimum viewing, use Adobe Reader, Version 7.0, available at http://www.adobe.com.
What's new
The following features and functions are new or updated in this release:
- Eclipse and Web Tools Platform Pre-requisite
- This version of the WebSphere Voice Toolkit can be used with any installation of Eclipse 3.2 and the Web Tools Platform 1.5.2, including IBM Rational SDP 7.0.
- Transcription Tool
- Listen to audio segments (local files or URIs) and provide transcriptions for them to be used for recognition accuracy analysis. A transcription file can be exported from and used by either the Voice Trace Analyzer or the Batch Grammar Test Tool.
- Grammar Weighting
- Ability to add weights to the rules of a grammar file to improve the recognition accuracy of a grammar.
- Lexicon Migration
- The latest version of the Latin American Spanish (es-MX) acoustic model has a different ASR phonology from the previous versions included with WebSphere® Voice Server V.5.1.X. If you previously created lexicon files that you want to use with the latest version of WebSphere Voice Server, then the toolkit provides a way for you to migrate these lexicon files using the Lexicon Migration wizard. The converted files contain the new phonemes in the phoneme tags.
- Batch Grammar Test Tool
- Allows you to automate the testing of a large number of grammars and audio files (local files or URIs), analyze the recognition accuracy of the test, and compare the results with previous tests.
- Local Voice Server Environment Removed
- The WebSphere Voice Toolkit no longer installs speech technology, an MRCP server, or a VoiceXML browser in the development environment. All speech-related functions, such as grammar testing and pronunciation generation, will now require an MRCP connection to a WebSphere Voice Server installation. The features to start and stop the Voice Server ("Start Voice Server" and "Stop Voice Server" menu items) are removed from the Run menu.
- Lexicon Editor (updated)
- Support the .lexml and .pls file extensions and provide more useful information to the Outline view.
- Voice Trace Analyzer (updated)
- Displays information and statistics on the grammar recognition events that have been logged on a WebSphere Voice Server system, with additional analysis of recognition accuracy for tuning of the grammars and server settings.
Allows you to export the list of audio files to be transcribed, perform that transcription offline, and import the transcriptions into this tool for further statistical analysis of the accuracy. Gathers the Synthesis events from the collected logs and displays them in a table in a new Synthesis tab, as well as interleaving them in with the Recognition events in the Call Flow tab.
Known problems and software limitations
General
- If after installation of the toolkit some functions do not appear or work as expected, use Window > Reset Perspective to reset the Voice Perspective to its default configuration.
- If resetting the perspective does not fix the problem, please close your Eclipse workbench and restart it using the
–clean
command line parameter. You can add this parameter by viewing the properties of the startup file and just adding –clean
to the executable (example: eclipse.exe –clean
). This increases the startup time slightly and needs to be done only once. After you have restarted, you can remove the –clean
parameter so that it is not executed again.
Note:
If you use this parameter more than once, the only effect is a longer start-up time for your Eclipse workbench.
- If your program stops responding for an extended period of time, close the Eclipse workbench and use the Windows Task manager to close any javaw.exe processes. If this fails to resolve the problem, restart your computer.
- In some configurations, one or more javaw.exe processes might continue to run after you exit the toolkit, particularly after an abnormal termination of the program.
Workaround:
Use the Windows Task Manager to locate and end these processes.
- The projects and files that you create using the
WebSphere Voice Toolkit are saved in the workspace used by the Eclipse workbench. If you uninstall the toolkit, the workspace remains on
your system so that the files are still available.
- Your microphone can pick up output from your computer speakers and
interpret it as words in a grammar. To avoid this, turn down the volume
on your speakers or use a headset.
- If you deploy your grammars and lexicon files to a web server for testing, you might find that updates to those files are not being loaded by the grammar testing tools. This occurs because the WebSphere Voice Server honors the caching policies of the web server to avoid having to retrieve the same document repeatedly. To circumvent this, you should set the caching policies for the web server such that the files are not cached, or assign an immediate expiration in the caching properties of your application documents when deploying them to your web server.
- If you create an audio file from a text file or input text field that contains a large amount of text, you might find the length incorrectly shows 0.0 seconds. The audio file is created correctly, however the initial display of the length data is incorrect. To display the correct length, press Play or close and re-open the audio file in the editor.
- If you experience JavaScript related errors, it might be due to the incorrect version of the Rhino JavaScript Interpreter. The WebSphere Voice Toolkit has been optimized to work with Rhino JavaScript Interpreter v1.5 release 3. It is recommended that this version, Rhino JavaScript Interpreter v1.5 release 3, be downloaded from http://www.mozilla.org/rhino/download.html using the download link rhino15R3.zip.
Important: Do not install any other version than v1.5 release 3, as this may cause error conditions with this product.
To install the Rhino JavaScript Interpreter download the rhino15R3.zip, extract the file js.jar, and copy to the
following directory:
<Eclipse_install_dir>\eclipse\plugins\com.ibm.voicetools.es_6.1.0
- Aside from your system TEMP directory, the toolkit creates log files in this directory, which can be removed after uninstalling the WebSphere Voice Toolkit:
- <Eclipse_install_dir>\eclipse\logs
When you change the Workbench or Workbench > Appearance
preferences, the Restore Defaults button might not restore
the original defaults. Close and
restart the Eclipse workbench for changes to take effect.
If you receive an "Internal Error - Widget Disposed" error
message, close and restart the Eclipse workbench to recover from the
error.
Use only alphanumeric characters and spaces for
directory, folder, and file names. Names should not include
non-alphanumeric characters. For example: ~!@#$%^&*)
To import or use files from a network resource, map a drive letter to that network
resource and import from the network drive.
The "Validate" context menu item in the Navigator view invokes any registered validators
on the selected resource(s). If no validators exist for the chosen resource(s), no validation
will occur, and no message will be shown. To see which validators are registered with the
Eclipse workbench, click Window > Preferences > Validation.
Language-specific considerations
- WebSphere Voice Toolkit 6.1 does not support Lexicons for AP Languages. This
is because a required feature, part-of-speech, is not supported by the
WebSphere Voice Server V5.x and higher.
- If you choose to install non-IBM Text-to-Speech (TTS) languages for
use on your voice application server, such as WebSphere Voice Server,
you should consult the vendor's documentation for details on the vendor's
implementation of VoiceXML 2.1 and Speech Synthesis Markup Language
(SSML). In addition, if you use another vendor's languages, you should
also use that vendor's tools for generating pronunciations. If you use
IBM tools to produce SSML for TTS output, the resulting code might not
run as expected in a deployment environment for that language.
- Do not create a project name using DBCS characters. If you do, you
receive validation errors and XML files do not validate.
- When using the Unknown Pronunciation view, Cantonese words are flagged
as "unknown." For correct Cantonese functionality of the tools, your
Windows regional setting should be "Chinese-Taiwan", rather than "Chinese-HongKong."
Editors
- The Content Assist pop-up window in the CCXML and LXML editors does not list all valid content when typing
a less than symbol (<) (as set by user preferences). It only proposes the appropriate closing tag.
Workaround:
To view the full list of valid tags, type a backspace. Selecting an item from the list also inserts the removed < symbol. Alternatively, press Ctrl + Space to open the Content Assist.
- In an SRGS-XML editor, the Outline view displays the graphical zoomed-out image, instead of displaying the XML source outline.
Workaround:
Select the Graphics page and then select the Source (SRGXML) page. This will resynch the Outline view with the Editor.
- The Content Assist in the editor does not display initially.
Workaround:
In all the Multi-page editors, the Content Assist Processor is not associated with the Editor until after the Design View is used. Open the Design/Graphics Page and then go back to the Source Page. If the editor actually starts in the Design/Graphics page you will
have to do this twice before content assist works.
- In the Lexicon editor, the Outline view does not always refresh as the file is being updated.
Workaround:
To refresh the Outline view, close the lexicon file being updated and then reopen the file. This allows the Lexicon editor to fully refresh the Outline View. Another option is to change the <lexicon
tag to <lexic and then change it back to <lexicon. This forces an update of the Outline view.
- When you select text in large VoiceXML files, it might take a few seconds for the highlighting to appear.
- If you have a file open in an editor, and you Undo all the
changes you have made to the file since your last Save,
the file is still considered to have changes pending.
- To pass validation, a VoiceXML document must have at least
one child of the VXML element.
- You might experience a problem with long lines of text
(such as lines that are thousands of characters in length)
in any of the Voice Toolkit editors. These conditions can cause the Eclipse workbench to hang.
If this occurs, shut down your system and restart it.
Open the file with the Default Text Editor (File > Open With or Open With from the pop-up menu)
to split the long lines into smaller ones, then save it
and re-open the file in the appropriate Voice Toolkit editor.
- In the SRGS grammar editor, if a rule is defined
but not used, no warning is generated.
- If you see the syntax error, "An error that cannot be
determined exists in the SISR of this grammar," with no
line information, it might be due to using invalid SISR syntax
in SRGS-XML or ABNF grammars. This error can occur in either the
grammar or VoiceXML editor if you declare an SISR object, and then
reference an undeclared child object of that object.
- After closing an editor, there might be a
delay before you can delete the file.
- In working with the SRGS-XML grammar editor, two messages for the
same error appear in the Tasks view. To turn off the XML parsing in the SRGS-XML
editor, follow these steps:
- In the Navigator view, right click on the project, and on the pop-up
menu, click Properties.
- On the Properties dialog, select Validation. Also select the
Override validation preferences check box, and clear the SRGS-XML
Form Validator check box.
- Click Apply, and then click OK.
Note:
This turns off the SRGS-XML editor validation, but the compiler's
XML validation still occurs.
Multi-page SRGS-XML Grammar Editor
Grammar conversions
Communication Flow Builder
Reusable Dialog Component (RDC) release notes:
- If you receive a message stating that Windows cannot open the JSP file, click Window > Preferences > Workbench > Capabilities and clear the Web Developer (typical) check box. Next, generate the JSP file again.
- If you use the group.rdg file, do not change the strategy and submit default values.
- By default, the order of the RDCs in a group is determined by the ID of the RDC objects. To order objects differently, you must drag in a Start object and then use Connect objects to achieve the desired order.
- In a Group Communication Flow Model, you can drag in only RDC files, not VoiceXML palette objects.
- When you launch the Tomcat server from the Workbench, an HTTP Status 500 error is displayed in the Web browser.
Workaround:
- Install J2SDK1.4.2, which is specified in the RUNNING.txt file in the Tomcat installation directory.
- Copy the tools.jar file from <J2SDK_install_dir>/lib to the <jakarta_tomcat_installation_dir>\common\lib.
- Restart the Tomcat server in the Workbench.
Grammar Test Tool
WebSphere Voice Toolkit documentation
- If you have Adobe Acrobat(R) 7.0 installed on your system
and you open a PDF file in the toolkit's Help window, the Help
window might stop responding when you close the PDF file or
change to another topic. If this occurs, use the Task Manager
to end the Help application, and then reopen the Help window.
Notices
The following paragraph does not apply to the
United Kingdom or any country where such
provisions are inconsistent with local law:
INTERNATIONAL BUSINESS MACHINES CORPORATION
PROVIDES THIS PUBLICATION "AS IS" WITHOUT WARRANTY
OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING,
BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR
A PARTICULAR PURPOSE. Some states do not allow
disclaimer of express or implied warranties in
certain transactions, therefore, this statement
may not apply to you.
This information could include technical
inaccuracies or typographical errors. Changes are
periodically made to the information herein; these
changes will be incorporated in new editions of
the publication. IBM may make improvements and\or
changes in the product(s) and\or the program(s)
described in this publication at any time without
notice.
Any references in this information to non-IBM Web
sites are provided for convenience only and do not
in any manner serve as an endorsement of those Web
sites. The materials at those Web sites are not
part of the materials for this IBM product and
use of those Web sites is at your own risk.
Information concerning non-IBM products was
obtained from the suppliers of those products,
their published announcements or other publicly
available sources. IBM has not tested those
products and cannot confirm the accuracy of
performance, compatibility or any other claims
related to non-IBM products. Questions on the
capabilities of non-IBM products should be
addressed to the suppliers of those products.
Trademarks
The following terms are trademarks or registered
trademarks of the International Business Machines
Corporation in the United States, other countries,
or both:
IBM
Rational
WebSphere
Intel and Pentium are trademarks or registered
trademarks of Intel Corporation in the United
States, other countries, or both.
Microsoft and Windows are trademarks of Microsoft
Corporation in the United States, other countries,
or both.
Java and all Java-based trademarks and logos are
trademarks or registered trademarks of Sun
Microsystems, Inc. in the United States, other
countries, or both.
Other company, product, and service names may be
trademarks or service marks of others.