torchwrench.extras.audio module¶

torchwrench.extras.audio.AudioMetaData¶

alias of _AudioMetaData

class torchwrench.extras.audio.AudioMetaDataDict[source]¶

Bases: TypedDict

bits_per_sample : int¶

encoding : str¶

num_channels : int¶

num_frames : int¶

sample_rate : int¶

class torchwrench.extras.audio.CodecConfig(*args, **kwargs)[source]¶

Bases: Placeholder

torchwrench.extras.audio.audio_metadata_to_dict(meta: _AudioMetaData) → AudioMetaDataDict[source]¶

torchwrench.extras.audio.dump_audio(src: Tensor, uri: BinaryIO | str | Path | PathLike | None, sample_rate: int, channels_first: bool = True, format: str | None = None, encoding: str | None = None, bits_per_sample: int | None = None, buffer_size: int = 4096, backend: str | None = None, compression: CodecConfig | float | int | None = None, *, overwrite: bool = True, make_parents: bool = True) → bytes[source]¶

Dump tensors to audio waveform file. Requires torchaudio package installed.

torchwrench.extras.audio.dump_with_torchaudio(src: Tensor, uri: BinaryIO | str | Path | PathLike | None, sample_rate: int, channels_first: bool = True, format: str | None = None, encoding: str | None = None, bits_per_sample: int | None = None, buffer_size: int = 4096, backend: str | None = None, compression: CodecConfig | float | int | None = None, *, overwrite: bool = True, make_parents: bool = True) → bytes[source]¶

Dump tensors to audio waveform file. Requires torchaudio package installed.

torchwrench.extras.audio.info_audio(uri: BinaryIO | str | PathLike | Path, **kwargs) → _AudioMetaData[source]¶

torchwrench.extras.audio.load_audio(uri: BinaryIO | str | PathLike | Path, frame_offset: int = 0, num_frames: int = -1, normalize: bool = True, channels_first: bool = True, format: str | None = None, buffer_size: int = 4096, backend: str | None = None) → tuple[Tensor, int][source]¶

torchwrench.extras.audio.load_with_torchaudio(uri: BinaryIO | str | PathLike | Path, frame_offset: int = 0, num_frames: int = -1, normalize: bool = True, channels_first: bool = True, format: str | None = None, buffer_size: int = 4096, backend: str | None = None) → tuple[Tensor, int][source]¶

torchwrench.extras.audio.save_audio(src: Tensor, uri: BinaryIO | str | Path | PathLike | None, sample_rate: int, channels_first: bool = True, format: str | None = None, encoding: str | None = None, bits_per_sample: int | None = None, buffer_size: int = 4096, backend: str | None = None, compression: CodecConfig | float | int | None = None, *, overwrite: bool = True, make_parents: bool = True) → bytes[source]¶

Dump tensors to audio waveform file. Requires torchaudio package installed.