hp.telephony.media
Interface SpeechDetectorConstants

All Known Subinterfaces:

ASRRecorderEvent, FaxEvent, Recorder, RecorderEvent, SMSRecorderEvent, SpeechDetector

All Known Implementing Classes:

AsyncMediaGroup

public interface SpeechDetectorConstants

Defines the Symbols used by SpeechDetector and RecorderEvent.

Since:

OCMP 2.2 with ASR/TTS extensions

See Also:

ResourceSpec.basicSpeechDetector

Field Summary

static Symbol	p_BargeIn Controls whether the caller can start speaking before prompts have ended.
static Symbol	p_FinalTimeout Controls the length of a period of silence after callers have spoken to conclude they finished.
static Symbol	p_InitialTimeout Controls how long the recognizer should wait after the end of the prompt for the caller to speak before sending a Recorder event (ev_Record, q_NoSpeechTimeout, null, e_OK).
static Symbol	p_InPromptSensitivityPercent Controls how loudly the caller must speack to interrupt prompts (barge-in) in order for speech to be detected.
static Symbol	p_Sensitivity Sensitivity of the speech detector when looking for speech.
static Symbol	p_Type Speech detector type.
static Symbol	q_EndOfSpeechDetected SpeechDetector stopped because end of speech has been detected.
static Symbol	q_NoSpeechTimeout SpeechDetector stopped because no speech has been detected before the initial timout popped.
static Symbol	q_SpeechDetected SpeechDetector detected speech.
static Symbol	rtca_PromptDone Advise the Speech Detector that the prompt is finished.
static Symbol	rtcc_EndOfSpeechDetected Trigger when end of speech has been detected (ie. final timeout popped).
static Symbol	rtcc_NoSpeechTimeout Trigger when no speech has been detected before the initial timout popped.
static Symbol	rtcc_SpeechDetected Trigger when speech has been detected.
static Symbol	v_HP Value for p_Type.
static Symbol	v_Nvad Value for p_Type.
static Symbol	v_SwiEp Value for p_Type.
static Symbol	v_Telisma Value for p_Type.

Field Detail

v_Nvad

public static final Symbol v_Nvad

Value for p_Type. Nuance algorithm based speech detector.

See Also:

p_Type

v_SwiEp

public static final Symbol v_SwiEp

Value for p_Type. Speech Works International algorithm based speech detector.

See Also:

p_Type

v_Telisma

public static final Symbol v_Telisma

Value for p_Type. Telisma DSR algorithm based speech detector.

See Also:

p_Type

v_HP

public static final Symbol v_HP

Value for p_Type. HP algorithm based speech detector.

See Also:

p_Type

p_Type

public static final Symbol p_Type

Speech detector type.

Value is v_Telisma, v_SwiEp, v_Nvad or v_HP.

See Also:

v_Telisma, v_SwiEp, v_Nvad, v_HP

rtca_PromptDone

public static final Symbol rtca_PromptDone

Advise the Speech Detector that the prompt is finished. A signal is more likely to occur.

Default : managed internally.

p_InitialTimeout

public static final Symbol p_InitialTimeout

Controls how long the recognizer should wait after the end of the prompt for the caller to speak before sending a Recorder event (ev_Record, q_NoSpeechTimeout, null, e_OK).

This parameter only takes effect after the speech detector has been informed that the prompt is done (managed internally).

Value is an Integer.
Default: 7000 ms
Range: 0 to 19999 (ms)

See Also:

q_NoSpeechTimeout, rtcc_NoSpeechTimeout

p_FinalTimeout

public static final Symbol p_FinalTimeout

Controls the length of a period of silence after callers have spoken to conclude they finished.

Value is an Integer.
Default: 4000 ms
Range: 0 to 9999 (ms)

p_BargeIn

public static final Symbol p_BargeIn

Controls whether the caller can start speaking before prompts have ended.

Value is a Boolean.
Default: Boolean.TRUE

p_Sensitivity

public static final Symbol p_Sensitivity

Sensitivity of the speech detector when looking for speech.

Higher values (more sensitive) tend to detect people who speak quietly, but risk false detection of background noise as speech. Lower values (less sensitive) tend to avoid accidently confusing background noise as speech, but may require some callers to speak more loudly.

Value is a Float.
Default: 0.5
Range: 0 to 1

p_InPromptSensitivityPercent

public static final Symbol p_InPromptSensitivityPercent

Controls how loudly the caller must speack to interrupt prompts (barge-in) in order for speech to be detected.

Whenever the caller is speaking, the recognizer and speech detector are continuously adapting their acoustic states to the current speech volume (as well as telephone line conditions). For most applications, the caller must speak slightly more loudly when the prompt is playing than when there is no prompt in order for speech to be detected. After the prompt is done, the sensitivity parameter is used to determine the sensitivity. While the prompt is playing, the InPromptSensitivityPercent parameter is multiplied by the sensitivity parameter to determine the sensitivity. Barge-in must be enabled for this setting to have any effect.

Value is an Integer.
Default: 40
Range: 0 to 100 (%)

q_NoSpeechTimeout

public static final Symbol q_NoSpeechTimeout

SpeechDetector stopped because no speech has been detected before the initial timout popped.

See Also:

p_InitialTimeout

q_SpeechDetected

public static final Symbol q_SpeechDetected

SpeechDetector detected speech.

q_EndOfSpeechDetected

public static final Symbol q_EndOfSpeechDetected

SpeechDetector stopped because end of speech has been detected.

See Also:

p_FinalTimeout

rtcc_NoSpeechTimeout

public static final Symbol rtcc_NoSpeechTimeout

Trigger when no speech has been detected before the initial timout popped.

See Also:

p_InitialTimeout

rtcc_SpeechDetected

public static final Symbol rtcc_SpeechDetected

Trigger when speech has been detected.

rtcc_EndOfSpeechDetected

public static final Symbol rtcc_EndOfSpeechDetected

Trigger when end of speech has been detected (ie. final timeout popped).

See Also:

p_FinalTimeout