Difference between revisions of "Voice Activity Detection (VAD)"

Latest revision as of 14:47, 25 October 2023

Voice Activity Detection (VAD) or Sound Detection is a feature available for the IP Station range. Voice Activity Detection samples the peak amplitude detected at the microphone every 30 ms and converts it to dB. If the sampled amplitude is continously above a trigger amplitude for a set duration, then a DAK key is triggered.

Sound Detection (VAD)

The VAD feature is available only when the station operates in ICX-AlphaCom mode.

Station Web interface

The Sound Detection configuration is done from the IP-station web interface at:

Advanced ICX-Alphacom --> VAD for INCA Stations
Advanced ICX-Alphacom --> Sound Detection for Turbine Stations

The following settings are possible to configure:

Sound Detection configuration page in IP Stations

Sound Detection status: Enable or Disable the feature
- In Turbine v. 4.7 the options are expanded:
  - Disable
  - Enable (Disable When Audio Out) - The Sound Detection is enabled when the station is in idle, and disabled when the station is in conversation, listening to Simplex Conference or Audio Program, or any other audio mechanism. This is to prevent the station from trigger by its own loudspeaker audio
  - Enable (Always) - The Sound Detection is always active, also when in conversation, listening to Simplex Conference or Audio Program, or any other audio mechanism. Can be used on devices without speaker (e.g. TKIE kit acting as mobile radio interface)
Minimum amplitude (dBA): Choose the trigger amplitude
Minimum duration of audio (ms): Choose for how long the sampled amplitude must be above the "Trigger ampltiude" before triggering the DAK key
DAK key to activate: Choose which DAK key to trigger. The station will simulate that this DAK key is pressed.
Minimum time before reactivation (ms): Choose how long to wait before enabling VAD after triggering the DAK key
Report DAK key off after (ms): This parameter sets the duration of the DAK key press.

AlphaPro Configuration

When the Sound Detection feature in the station is triggered, it will simulate that a DAK key (Call Button) is pressed. By default this is DAK 5. In the DAK configuration in AlphaPro enter the number to call on the corresponding DAK Key entry.

DAK configuration in AlphaPro - Here the Sound Detection will call number 234

Modifying parameters by Data Command

The parameters used by the Sound Detection feature can be modified by an ACDP data command. This command can also be triggered from the Event Handler.

$DIP L301 "vad 75 350 5 2000 200 1"

This command will change the parameters on station 301 to the values shown in the image above. The last parameter ("1") enables the feature. Value "0" would disable the feature.

Notes on operation

The sound pressure SPL is dropping by 6 dB (four times) for every doubling of the distance from the microphone.
The audio detection is before the noise reduction circuit, so in case Active Noise Cancellation is activated, noise will always cause the detector to trigger.
The VAD is working only when the station is in idle. It will not work if the Station is for example in conversation or in conference.
- Turbine devices with firmware 4.7 or later can also trigger during conversation or while in conference.
The VAD feature is only available when the station is used in ICX-AlphaCom mode. The feature is not available in IC-Edge or SIP mode.

Software requirement

INCA Station software 01.09.3.0 or later.
Turbine: available in all versions

@@ Line 1: / Line 1: @@
-'''Voice Activity Detection''' (VAD) is a feature available for the IP Station range. Voice Activity Detection samples the peak amplitude detected at the microphone every 30 ms and converts it to dB. If the sampled amplitude is continously above a trigger amplitude for a set duration, then a [[DAK|DAK key]] is triggered.
+{{AI}}'''Voice Activity Detection''' (VAD) or '''Sound Detection''' is a feature available for the IP Station range. Voice Activity Detection samples the peak amplitude detected at the microphone every 30 ms and converts it to dB. If the sampled amplitude is continously above a trigger amplitude for a set duration, then a [[DAK|DAK key]] is triggered.
-In the [[:Category:Turbine|Turbine range of stations]], the feature is called '''Sound Detection'''.
+<table style="float:right; max-width:80%; border-style: double; border-color: #c7c7c7;">
+<tr style="rowspan:2;">
+<td style="display: block; vertical-align: bottom;">[[Image:VAD principle.png|200px]]</td>
+</tr>
+<tr>
+<td style='background-color:#efefef;'>Sound Detection (VAD)</td>
+</tr>
+</table>
-== User interface ==
+{{note|The VAD feature is available only when the station operates in '''ICX-AlphaCom mode'''. }}
-The Voice Activity Detection configuration is done from the IP-station web interface at:
-* '''Advanced Alphacom''' --> '''VAD''' for [[:Category:INCA Stations|Inca]] stations
+== Station Web interface ==
-* '''Advanced Alphacom''' --> '''Sound Detection''' for [[:Category:Turbine|Turbine]] stations
+The Sound Detection configuration is done from the IP-station web interface at:
+* '''Advanced ICX-Alphacom''' --> '''VAD''' for [[:Category:Stations#INCA_stations|INCA Stations]]
+* '''Advanced ICX-Alphacom''' --> '''Sound Detection''' for [[:Category:Stations#Turbine_stations|Turbine Stations]]
 The following settings are possible to configure:
-[[File:VAD Inca.png|left|thumb|500px|Voice Activity Detection configuration page in Inca stations]]
-<br style="clear:both;" />
-[[File:VAD Turbine.png|left|thumb|500px|Sound Detection configuration page in Turbine stations]]
+<table style="max-width:80%; border-style: double; border-color: #c7c7c7;">
-<br style="clear:both;" />
+<tr style="rowspan:2;">
+<td style="display: block; vertical-align: bottom;">[[Image:VAD2.png|700px]]</td>
+</tr>
+<tr>
+<td style='background-color:#efefef;'>Sound Detection configuration page in IP Stations</td>
+</tr>
+</table>
-* '''Sound Detection status:''' Choose whether to enable to disable VAD
+* '''Sound Detection status:''' Enable or Disable the feature
+** In Turbine v. 4.7 the options are expanded:
+*** Disable
+*** Enable (Disable When Audio Out) - The Sound Detection is enabled when the station is in idle, and disabled when the station is in conversation, listening to Simplex Conference or Audio Program, or any other audio mechanism. This is to prevent the station from trigger by its own loudspeaker audio
+*** Enable (Always) - The Sound Detection is always active, also when in conversation, listening to Simplex Conference or Audio Program, or any other audio mechanism. Can be used on devices without speaker (e.g. [[TKIE-2|TKIE kit]] acting as mobile radio interface)
 * '''Minimum amplitude (dBA):''' Choose the trigger amplitude
 * '''Minimum duration of audio (ms):''' Choose for how long the sampled amplitude must be above the "Trigger ampltiude" before triggering the DAK key
-* '''DAK key to activate:''' Choose which DAK key to trigger
+* '''DAK key to activate:''' Choose which DAK key to trigger. The station will simulate that this DAK key is pressed.
 * '''Minimum time before reactivation (ms):''' Choose how long to wait before enabling VAD after triggering the DAK key
-* '''Report DAK key off after (ms):''' With this parameter it is possible to delay sending DAK key off for some time
+* '''Report DAK key off after (ms):''' This parameter sets the duration of the DAK key press.
+<br>
+== AlphaPro Configuration ==
+When the Sound Detection feature in the station is triggered, it will simulate that a DAK key (Call Button) is pressed. By default this is DAK 5. In the [[Users_%26_Stations_(AlphaPro)#DAK|DAK configuration in AlphaPro]] enter the number to call on the corresponding DAK Key entry.
+<table style="max-width:80%; border-style: double; border-color: #c7c7c7;">
+<tr style="rowspan:2;">
+<td style="display: block; vertical-align: bottom;">[[Image:VAD DAK5.png|1000px]]</td>
+</tr>
+<tr>
+<td style='background-color:#efefef;'>DAK configuration in AlphaPro - Here the Sound Detection will call number 234</td>
+</tr>
+</table>
+<br>
+== Modifying parameters by Data Command ==
+The parameters used by the Sound Detection feature can be modified by an [[ACDP|ACDP data command]]. This command can also be triggered from the [[Event Handler]].
+{{code|$DIP L301 "vad 75 350 5 2000 200 1"}}
+This command will change the parameters on station 301 to the values shown in the image above. The last parameter ("1") enables the feature. Value "0" would disable the feature.
+<br>
 == Notes on operation ==
 *The sound pressure SPL is dropping by 6 dB (four times) for every doubling of the distance from the microphone.
-*The audio detection is before the noise reduction circuit, so in case [[Active Noise Cancelling|Noise Reduction]] is activated, noise will always cause the detector to trigger.
+*The audio detection is ''before'' the noise reduction circuit, so in case [[Active Noise Cancellation (ANC)|Active Noise Cancellation]] is activated, noise will always cause the detector to trigger.
-*The VAD is working only when the station is in idle. It will not work if the Station is for example in conversation or in conference. (Station software 2.2.3.1)
+*The VAD is working only when the station is in idle. It will not work if the Station is for example in conversation or in conference.
-* The VAD feature is only available when the station is used in AlphaCom mode. The feature is not available in Pulse or SIP mode.
+** Turbine devices with firmware 4.7 or later can also trigger during conversation or while in conference.
+* The VAD feature is only available when the station is used in ICX-AlphaCom mode. The feature is not available in IC-Edge or SIP mode.
 == Software requirement ==
@@ Line 32: / Line 76: @@
 == Related articles ==
-* [[Scream Alarm using VAD]]
+* [[Silent Alarm]]
+* [[Scream Alarm]]
-[[Category:Applications]]
+[[Category:INCA Station Configuration Guide]]
-[[Category: IP Stations]]
+[[Category:Turbine Configuration]]
-[[Category: AlphaCom features]]
+[[Category: ICX-AlphaCom features]]
+[[Category:AlphaCom features]]
+[[Category: Audio Settings and Audio Features]]