toolbox/src/exampleCode/speech/inst README
Both the execution end development environments for speech recognition
are included on this (v5.1) developers' CD. The capabilities of
speech recognition are discrete-utterance, speaker-independent, small
vocabulary. Online help and man pages are included for speech.
A reboot after installing speech is required before starting speech.
(Technically, all that is needed is an X-server process restart.
However there is no standard UI to do this (/etc/killall Xsgi
assumes xdm is running to restart the server). A reboot is simpler
and safer.)
Speech's API spec may be found in /usr/share/data/speech/misc
and is a showcase document call recog.api. A simple example may be
found there as well, (colors.c) - install speech_dev.sw.misc for these.
Current inst image explanation:
speechManager.sw.base
This includes the client GUI and the server recognition algorithms.
All the stuff needed for speech and the server to execute
correctly (DSO's, executables, and data files).
speechManager.sw.applications
This image places a file called <application> in
/usr/share/data/speech/bindings from which speech can read
predefined associations between words and keystrokes (on a
per-app basis). Look for the menu selection "Vocabulary"
off speech's main menu. Current examples are MediaMail
(and Zmail), Jot (and Zip), and global (4Dwm). This also
includes some trained templates for those vocbaularies.
speechManager.books.SpeechMgr_UG
Online help and user guide.
speechManager.man.relnotes
Man pages and release notes for speech.
speech_dev.sw.dev
This is speech's development environment consisting of headers
and libraries. It currently does not include the vocabulary
development system (VDS) from Scott Instruments which is
required for commercial applications.
speech_dev.sw.misc
This includes an API spec (recog.api - a showcase document) and
some examples in /usr/share/data/speech/misc.
Documentation
Data
OR, ...