Document License

 This work is licensed under the Creative Commons Attribution-Share Alike 3.0
 United States License. To view a copy of this license, visit 
 http://creativecommons.org/licenses/by-sa/3.0/us/
 or send a letter to Creative Commons, 171 Second Street, Suite 300,
 San Francisco, California, 94105, USA.

 
Contributors to this document

 Copyright (C) 2010 Texas Instruments Incorporated - http://www.ti.com/


------------------
This directory contains data associated with the TIesr Speech Recognizer and
with test programs released as part of the open source. This directory also acts
as a repository for files used by or generated by the TIesr test programs. 

There are three sub-directories as described below.

The OffDT_GenDictPhbVR_LE_MQ directory contains the American English word pronunciation data, a general American dictionary, and phonetic acoustic model data. This data is used by the TIesrFlex API when creating a grammar network and acoustic model data set for a particular user-defined grammar. See the detailed ReadMe documentation in the directory for further information specific to this data, and the TIesr Documentation directory for further information regarding TIesr data.

The AudioDir directory is used by the testtiesrfa program. The testtiesrfa program utilizes the TIesrFA API to record a duration of audio data and write it to a file in the AudioDir directory. The data is in raw 16-bit single-channel format, normally sampled at 8kHz sampling rate. A user of TIesr should execute the testtiesrfa program to record audio data, and examine the data to ensure the proper operation of the TIesrFA API and the quality of the recorded audio signal. It is very important to obtain a good quality audio signal in order for TIesr to operate successfully. There should not be noticeable distortion or clipping in the audio signal.

The GramDir is used by the testtiesrflex program.  The testtiesrflex program utilizes the TIesrFlex API to generate a set of binary grammar network and model files based on an input user-supplied grammar. The program outputs the binary data to the GramDir directory for subsequent use by the TIesrSI API when performing speech recognition based on the input grammar. For further information on creating an input grammar, consult the TIesr Documentation directory.

Various files are also located in this directory. The TestTIesrDTWords.txt file contains words, one per line, that the testtiesrdt program uses. The program outputs the phonetic  pronunciation of each word as determined by the TIesrDT API, which is based on spelling alone. It is likely that there will be pronunciation errors, since determining pronunciation from word spelling alone is not reliable. Consult documentation for the TIesrDT API for further information.

Several output files generated while running test programs may be output in this directory. The TIesr Speech Recognizer keeps track of the state of the recognizer and environment in the "JAC" file. The testtiesrsi program writes the JAC file to this directory for use in testing. Consult documentation for the TIesrSI API for further information.

