INNER CODE UNIT · Python

split_silence

voicepaw/so-vits-svc-fork · src/so_vits_svc_fork/inference/core.py:56

def split_silence(
    audio: ndarray[Any, dtype[float32]],
    top_db: int = 40,
    ref: float | Callable[[ndarray[Any, dtype[float32]]], float] = 1,
    frame_length: int = 2048,
    hop_length: int = 512,
    aggregate: Callable[[ndarray[Any, dtype[float32]]], float] = np.mean,
    max_chunk_length: int = 0,
) -> Iterable[Chunk]:
    non_silence_indices = librosa.effects.split(
        audio,
        top_db=top_db,
        ref=ref,
        frame_length=frame_length,
        hop_length=hop_length,
        aggregate=aggregate,
    )
    last_end = 0

View source record →