Recognizer RECOGNITION-START-TIMERS

RECOGNITION-START-TIMERS is the MRCPv1 method for starting the recognizer’s no-input timers, sent from the client once it knows a kill-on-barge-in prompt has finished playing. It is useful when the recognizer and synthesizer are not in the same session.

While a kill-on-barge-in prompt plays, the client typically wants a RECOGNIZE request active so it can detect and act on barge-in, but it does not want the recognizer to start the no-input timers until the prompt has finished. The Recognizer-Start-Timers header field on the RECOGNIZE request controls this: when the timers are deferred, the recognizer holds them until the client sends a RECOGNITION-START-TIMERS request.

This is an MRCPv1-only method. It should not be confused with the MRCPv2 START-INPUT-TIMERS method, which performs the same function under the newer protocol.

MRCPv1 example

Once the prompt has finished, the client sends RECOGNITION-START-TIMERS to start the timers:

Related articles

• Recognizer Methods
• Recognizer RECOGNIZE
• Recognizer START-INPUT-TIMERS (MRCPv2 equivalent)
• Recognizer SET-PARAMS