|
|
|
|
|
The following table describes the parameters that can be used to control the behavior of the recognition functions.
Parameter |
Description |
IRP_RECOG_TYPE |
Indicates which type of recognition to use. Speech recognition type: IRD_WHOLE_WORD. |
IRP_RECOG_GRAMMAR |
Depends on IRP_RECOG_TYPE, valid values are found with the grammar header files associated with the recognition type. |
IRP_RECOG_GRAMNAME |
Specifies the recognition grammar name for WholeWord speech recognition. |
IRP_RECOG_PRETIME |
Specifies a timeout to wait for caller input, measured in milliseconds. |
IRP_RECOG_INTERTIME |
Specifies a timeout to wait for caller input between digits, measured in milliseconds [applies only to touch tone (DTMF)]. |
IRP_ECHOCAN_TYPE � IRD_SP_ECHO |
Do not change the IRP_ECHOCAN_TYPE parameter. Changing it renders the echo canceler inoperable. |
All recognition parameters should be set before calling any recognition functions and retain those settings until the functions complete.
Using the WholeWord recognizer during prompting without echo cancellation results in premature responses from the recognizer with nonsense values.
Header files for WholeWord speech recognition
WholeWord recognition grammar numbers are defined in header files. Each WholeWord recognition language has a specific header file named according to the following template:
/att/asr/grammar_hs/{XX}.gram.h
where {XX} is the country specific designation. For US English, {XX} is US. Applications using US English WholeWord speech recognition would include the grammar header file into their programs as follows:
#include "/att/asr/grammar_hs/US.gram.h"
