5.1 Voice Activity Detector Operating Modes

The voice activity detector (VAD) is a component in the CSP software that examines a caller's incoming signal and determines if the signal contains significant energy and is likely to be speech rather than a click, for example.

When the voice activity detector (VAD) detects audio energy, you can specify that it act on this energy in one or more of the following ways:

• send a VAD event to the host application when speech is detected
• stop play when speech is detected (barge-in) or allow play to continue
• record/stream data to the host application only after energy is detected (voice-activated record/stream) or constantly record/stream

Table5-1 summarizes the types of operation, the modes and the settings required to enable them.

Table 5-1. VAD and Barge-In Operating Modes 

Operation	Modes	Required Setting
send VAD event to host when speech is detected	on	DM_VADEVTS event bit mask set in dx_setevtmsk( )
	off (default)	DM_VADEVTS bit mask not set
stop play when speech is detected	on	DXCH_BARGEIN = 1 in ec_setparm( ) On SpringWare boards, to use barge-in, you must also set ECCH_VADINITIATED to 1.
	off (default)	DXCH_BARGEIN = 0
record/stream data	constant	ECCH_VADINITIATED = 0 in ec_setparm( ) Use ec_reciottdata( ) or ec_stream( ) to record echo-cancelled data
	voice activated (default)	ECCH_VADINITIATED = 1 in ec_setparm( ) Use ec_reciottdata( ) or ec_stream( ) to record echo-cancelled data

The following topics provide more information on the VAD operating modes:

• Sending a VAD Event to the Host Application
• Stopping Play When Speech is Detected (Barge-In)
• Voice-Activated or Constant Recording
• Sample VAD Scenarios

Click here to contact Dialogic Customer Engineering