Interface: AdaptiveVADConfig

Defined in: packages/agentos/src/core/audio/AdaptiveVAD.ts:49

Configuration options for the AdaptiveVAD.

Properties

energySmoothingFrames?

optional energySmoothingFrames: number

Defined in: packages/agentos/src/core/audio/AdaptiveVAD.ts:74

Number of past frames to consider for smoothing energy calculations (if smoothing is applied).

Default

5

---

maxSilenceDurationMsInSpeech?

optional maxSilenceDurationMsInSpeech: number

Defined in: packages/agentos/src/core/audio/AdaptiveVAD.ts:61

Maximum duration of silence in milliseconds within a speech segment before it's considered ended. e.g., a pause between words.

Default

500

---

minSpeechDurationMs?

optional minSpeechDurationMs: number

Defined in: packages/agentos/src/core/audio/AdaptiveVAD.ts:55

Minimum duration in milliseconds that a sound segment must have to be considered speech. Helps filter out very short, non-speech noises.

Default

150

---

thresholdRatio?

optional thresholdRatio: number

Defined in: packages/agentos/src/core/audio/AdaptiveVAD.ts:81

Ratio of speech_threshold / silence_threshold. Helps in creating a hysteresis effect. speech_threshold = silence_threshold * thresholdRatio

Default

1.5

---

vadSensitivityFactor?

optional vadSensitivityFactor: number

Defined in: packages/agentos/src/core/audio/AdaptiveVAD.ts:69

Sensitivity adjustment factor, further fine-tunes thresholds from EnvironmentalCalibrator. Values > 1.0 make VAD less sensitive (require louder input for speech). Values < 1.0 make VAD more sensitive. This is applied ON TOP of the sensitivity factor in EnvironmentalCalibrator.

Default

1.0