Estimates word-level timestamps by mapping words to acoustic energy regions (pauses/silences) in the synthesized waveform, falling back to character-proportional spacing within each phrase.
Signature:
export declare function computeWordTimestamps(text: string, samples: Float32Array, sampleRateHz: number, startOffsetSec?: number): WordTimestamp[];
Parameters
| Parameter | Type | Description |
|---|---|---|
| text | string | The input text corresponding to the synthesized audio samples. |
| samples | Float32Array | Mono PCM audio samples in the range [-1.0, 1.0]. |
| sampleRateHz | number | Sample rate of the audio samples in Hz. |
| startOffsetSec | number | Optional time offset in seconds to add to all computed timestamps (defaults to 0). |
Returns: