********** DISREGARD… I FIGURED IT OUT!!! ****************
(In case anyone was following this)- I got the application I described below to work. As it turns out I was on the right track originally, using only one trigger word to start, then building the hierarchy of command menus in the application code. The original code was good all along, only it was not responding because I did not account for the half second or so it was taking for the application to play the response chirp when I commanded the sub-menus. The chirp was therefore still playing when the application started listening for the next command, and that was causing it not to recognize it. I placed a one-second delay just before the T2SI command, and bingo- it recognized the sub-command on the first try… whew… I thought I would never get past that.
While I am waiting for my EasyVR, I am continuing work on my SmartVR Project. I have a nice trigger and command menu routine working very well, but it is just a simple trigger word and a long monolithic response menu, which is programmed to trigger an MP3 player for a given response. It is all going well as a single list of commands, but I am having trouble figuring out how to start splitting this into sub-menus. I basically want the tree as such:
main trigger–>command 1
-----------------command 2
-----------------command 3
-----------------command 4 (or trigger?)------------->Sub Command 1
-----------------------------------------------------------Sub Command 2, etc.
-----------------command 5 (or trigger?)------------->Sub Command 1
---------------------------------------------------------->Sub Command 2 etc.
I am trying to emulate a popular interactive toy. The trigger word is the robot’s name, then there are a half dozen commands you could give it. Two of those commands should lead to their own separate list of commands, say “game mode” then the list of games, and “command mode” and then a list of actions.
I am working with three acoustic models which I split into the main menu, the game mode menu, and the command mode menu. I used the acoustic model combiner to combine all of these in Quick T2SI. This is where I am lost. When we want to get into sub-menus and such, do we create separate trigger words for each? If so, how do we use the _T2SI((long) &nn_combotest, (long)&gsearch_trig_test1, T2SI_DEFAULT_KNOB, T2SI_DEFAULT_TIMEOUT, T2SI_DEFAULT_TRAILING, &results) dialogue to search for more than one trigger? or to search for a trigger for a second list while at the same time listening for commands from the current list?
Is there only one “official” trigger phrase allowed in a given application, with the other sub-menus being created in the source code? I really wish there were an example out there. I found one example of a routine that used the acoustic model combiner (for the T2SIMath example) but it does not really deal with multiple levels of menus, so I could not find an answer there.
I hope someone out there is familiar enough with SmartVR that they can help, or at least point me to a tutorial or example or something. Any help is much appreciated.
Rob