we're having a pretty huge problem here:
when for instance the built in Windows7 speech recognizer is running along with Cloudgarden code, both recognize commands at the same time. When I say something e.g. "What time is it?" from your test program, Cloudgarden will hear "What time is it" and Windows SR will hear e.g. "What can I say" at the same time and both commands will be triggered from the same speech.
This problem remains the same, no matter if we use the default recognizer or select a specific one and also occurs under Vista and even XP when using other SAPI5 recognition programs. With other recognition programs we don't have this: every program receives only the commands it registered and there are no double recognitions.
From how it behaves, it seems that Cloudgarden is launching an inprocess recognizer instead of a shared recognizer as Win7 and our other recognition programs are doing. Is that so and if yes, is there any way to influence that?
If we can't get that to work properly, we can't use Cloudgarden as we need several recognizers running side by side with some global commands in each...(and I'd really hate to write my own bridge between Java, SAPI5 and Dragon)
Thanks in advance for your help.