=?utf-8?Q?At_the_start_of_recognition_I_us?=

Hi all,



I am developing an application that uses “system.speech” in “windows
vista”. 
The problem I face is sometimes at the start of recognition I used to get 
the following

“e b 2 b e & s l t z n w i r @ w n @ t 3 ) 7 , - n i % x t r u 6 4 % s l
t” 
junk and by restarting the speech engine works fine  and most of the times 
the things are recognized correctly. I could not find out what is the real 
cause of the problem.



Thanks in advance