pyflink.multimodal.expression.AudioExpressionAccessor.detect_speech

AudioExpressionAccessor.detect_speech(*, aggressiveness: int = 0, frame_ms: int = 30, min_speech_ms: int = 0, merge_gap_ms: int = 0, max_segments: int = 1024, concurrency: Optional[int] = None) → pyflink.table.expression.Expression

Detect speech activity ranges in waveform values.

Equivalent to audio_detect_speech(). See that function for full parameter details.

Parameters

• aggressiveness – WebRTC VAD aggressiveness, from 0 to 3.

• frame_ms – WebRTC frame size in milliseconds.

• min_speech_ms – Drop detected speech ranges shorter than this value.

• merge_gap_ms – Merge ranges separated by this gap or less.

• max_segments – Maximum number of ranges to return.

• concurrency – Optional execution concurrency for this operation. None uses the framework default.

Returns

A DataFrame expression that produces ARRAY<ROW<start_ms, end_ms, duration_ms>> speech ranges.