'fltp': The output is torch.float32 type. Otherwise, the sample is converted and the output dtype is changed

If None, the output chunk has dtype corresponding to To list decoder options for a decoder, you can use To list the available decoders, you can use ffmpeg -decoders command.ĭecoder_option ( dict or None, optional) – When provided, use the specified decoder instead of the default one. If omitted, default_audio_stream is used. Stream_index ( int or None, optional) – The source audio stream index. When the number of chunks buffered exceeds this number, old frames are If the source stream is exhausted before enough frames are buffered, add_basic_audio_stream ( frames_per_chunk : int, buffer_chunk_size : int = 3, stream_index : Optional = None, decoder : Optional = None, decoder_option : Optional ] = None, format : Optional = 'fltp', sample_rate : Optional = None ) ¶įrames_per_chunk ( int) – Number of frames returned as one chunk. Timestamp ( float) – Target time in second. Seek the stream to the given timestamp Parameters OutputStream seek ( timestamp : float ) ¶ Get the metadata of output stream Parameters SourceStream get_out_stream_info ( i : int ) → torchaudio.io._stream_reader.StreamReaderOutputStream ¶ Get the metadata of source stream Parameters Optional get_src_stream_info ( i : int ) → torchaudio.io._stream_reader.StreamReaderSourceStream ¶ Number of output streams configured by client code. Number of streams found in the provided media source. Such as microphone, camera and screen, or a virtual device. This allows to load media stream from hardware devices,

This is useful when the input data do not contain a header. This argument serves two different usecases.

Override the input format, or specify the source sound device.ĭefault: None (no override nor device input). Please refer to the following for the expected signature and behavior If file-like object, it must support read method with the signatureĪdditionally, if the file-like object has seek method, it uses The supported value depends on the FFmpeg found This includes a file path, URL, device identifier orįilter expression. If string-type, it must be a resource indicator that FFmpeg can StreamReader ( src : str, format : Optional = None, option : Optional ] = None, buffer_size : int = 4096 ) ¶įetch and decode audio/video streams chunk by chunk.įor the detailed usage of this class, please refer to the tutorial.

Ffmpeg python documentation