Silence Segmenter

speechline.segmenters.silence_segmenter.SilenceSegmenter (Segmenter)

Source code in speechline/segmenters/silence_segmenter.py

class SilenceSegmenter(Segmenter):
    def chunk_offsets(
        self,
        offsets: List[Dict[str, Union[str, float]]],
        silence_duration: float,
        **kwargs,
    ) -> List[List[Dict[str, Union[str, float]]]]:
        """
        Chunk transcript offsets based on in-between silence duration.

        ### Example
        ```pycon title="example_silence_segmenter.py"
        >>> from speechline.segmenters import SilenceSegmenter
        >>> segmenter = SilenceSegmenter()
        >>> offsets = [
        ...     {"start_time": 0.0, "end_time": 0.4},
        ...     {"start_time": 0.4, "end_time": 0.5},
        ...     {"start_time": 0.6, "end_time": 0.8},
        ...     {"start_time": 1.1, "end_time": 1.3},
        ...     {"start_time": 1.5, "end_time": 1.7},
        ...     {"start_time": 1.7, "end_time": 2.0},
        ... ]
        >>> segmenter.chunk_offsets(offsets, silence_duration=0.2)
        [
            [
                {"start_time": 0.0, "end_time": 0.4},
                {"start_time": 0.4, "end_time": 0.5},
                {"start_time": 0.6, "end_time": 0.8},
            ],
            [
                {"start_time": 1.1, "end_time": 1.3}
            ],
            [
                {"start_time": 1.5, "end_time": 1.7},
                {"start_time": 1.7, "end_time": 2.0}
            ],
        ]
        >>> segmenter.chunk_offsets(offsets, silence_duration=0.1)
        [
            [
                {"start_time": 0.0, "end_time": 0.4},
                {"start_time": 0.4, "end_time": 0.5}
            ],
            [
                {"start_time": 0.6, "end_time": 0.8}
            ],
            [
                {"start_time": 1.1, "end_time": 1.3}
            ],
            [
                {"start_time": 1.5, "end_time": 1.7},
                {"start_time": 1.7, "end_time": 2.0}
            ],
        ]
        ```

        Args:
            offsets (List[Dict[str, Union[str, float]]]):
                Offsets to chunk.
            silence_duration (float):
                Minimum in-between silence duration (in seconds) to consider as gaps.

        Returns:
            List[List[Dict[str, Union[str, float]]]]:
                List of chunked/segmented offsets.
        """
        # calculate gaps in between offsets
        gaps = [
            round(next["start_time"] - curr["end_time"], 3)
            for curr, next in zip(offsets, offsets[1:])
        ]
        # generate segment slices (start and end indices) based on silence
        slices = (
            [0]
            + [idx + 1 for idx, gap in enumerate(gaps) if gap >= silence_duration]
            + [len(offsets)]
        )
        # group consecutive offsets (segments) based on slices
        segments = [offsets[i:j] for i, j in zip(slices, slices[1:])]
        return segments

chunk_offsets(self, offsets, silence_duration, **kwargs)

Chunk transcript offsets based on in-between silence duration.

Example

example_silence_segmenter.py

>>> from speechline.segmenters import SilenceSegmenter
>>> segmenter = SilenceSegmenter()
>>> offsets = [
...     {"start_time": 0.0, "end_time": 0.4},
...     {"start_time": 0.4, "end_time": 0.5},
...     {"start_time": 0.6, "end_time": 0.8},
...     {"start_time": 1.1, "end_time": 1.3},
...     {"start_time": 1.5, "end_time": 1.7},
...     {"start_time": 1.7, "end_time": 2.0},
... ]
>>> segmenter.chunk_offsets(offsets, silence_duration=0.2)
[
    [
        {"start_time": 0.0, "end_time": 0.4},
        {"start_time": 0.4, "end_time": 0.5},
        {"start_time": 0.6, "end_time": 0.8},
    ],
    [
        {"start_time": 1.1, "end_time": 1.3}
    ],
    [
        {"start_time": 1.5, "end_time": 1.7},
        {"start_time": 1.7, "end_time": 2.0}
    ],
]
>>> segmenter.chunk_offsets(offsets, silence_duration=0.1)
[
    [
        {"start_time": 0.0, "end_time": 0.4},
        {"start_time": 0.4, "end_time": 0.5}
    ],
    [
        {"start_time": 0.6, "end_time": 0.8}
    ],
    [
        {"start_time": 1.1, "end_time": 1.3}
    ],
    [
        {"start_time": 1.5, "end_time": 1.7},
        {"start_time": 1.7, "end_time": 2.0}
    ],
]

Parameters:

Name	Type	Description	Default
offsets	List[Dict[str, Union[str, float]]]	Offsets to chunk.	required
silence_duration	float	Minimum in-between silence duration (in seconds) to consider as gaps.	required

Returns:

Type	Description
List[List[Dict[str, Union[str, float]]]]	List of chunked/segmented offsets.

Source code in speechline/segmenters/silence_segmenter.py

def chunk_offsets(
    self,
    offsets: List[Dict[str, Union[str, float]]],
    silence_duration: float,
    **kwargs,
) -> List[List[Dict[str, Union[str, float]]]]:
    """
    Chunk transcript offsets based on in-between silence duration.

    ### Example
    ```pycon title="example_silence_segmenter.py"
    >>> from speechline.segmenters import SilenceSegmenter
    >>> segmenter = SilenceSegmenter()
    >>> offsets = [
    ...     {"start_time": 0.0, "end_time": 0.4},
    ...     {"start_time": 0.4, "end_time": 0.5},
    ...     {"start_time": 0.6, "end_time": 0.8},
    ...     {"start_time": 1.1, "end_time": 1.3},
    ...     {"start_time": 1.5, "end_time": 1.7},
    ...     {"start_time": 1.7, "end_time": 2.0},
    ... ]
    >>> segmenter.chunk_offsets(offsets, silence_duration=0.2)
    [
        [
            {"start_time": 0.0, "end_time": 0.4},
            {"start_time": 0.4, "end_time": 0.5},
            {"start_time": 0.6, "end_time": 0.8},
        ],
        [
            {"start_time": 1.1, "end_time": 1.3}
        ],
        [
            {"start_time": 1.5, "end_time": 1.7},
            {"start_time": 1.7, "end_time": 2.0}
        ],
    ]
    >>> segmenter.chunk_offsets(offsets, silence_duration=0.1)
    [
        [
            {"start_time": 0.0, "end_time": 0.4},
            {"start_time": 0.4, "end_time": 0.5}
        ],
        [
            {"start_time": 0.6, "end_time": 0.8}
        ],
        [
            {"start_time": 1.1, "end_time": 1.3}
        ],
        [
            {"start_time": 1.5, "end_time": 1.7},
            {"start_time": 1.7, "end_time": 2.0}
        ],
    ]
    ```

    Args:
        offsets (List[Dict[str, Union[str, float]]]):
            Offsets to chunk.
        silence_duration (float):
            Minimum in-between silence duration (in seconds) to consider as gaps.

    Returns:
        List[List[Dict[str, Union[str, float]]]]:
            List of chunked/segmented offsets.
    """
    # calculate gaps in between offsets
    gaps = [
        round(next["start_time"] - curr["end_time"], 3)
        for curr, next in zip(offsets, offsets[1:])
    ]
    # generate segment slices (start and end indices) based on silence
    slices = (
        [0]
        + [idx + 1 for idx, gap in enumerate(gaps) if gap >= silence_duration]
        + [len(offsets)]
    )
    # group consecutive offsets (segments) based on slices
    segments = [offsets[i:j] for i, j in zip(slices, slices[1:])]
    return segments