`scipy.io.wavfile:_read_data_chunk`

source: /scipy/io/wavfile.py :447

Signature

   def     _read_data_chunk (    fid    ,    format_tag    ,    channels    ,    bit_depth    ,    is_big_endian    ,    is_rf64    ,    block_align    ,    mmap    =  False   ,    rf64_chunk_size    =  None     )

Notes

Assumes file pointer is immediately after the 'data' id

It's possible to not use all available bits in a container, or to store samples in a container bigger than necessary, so bytes_per_sample uses the actual reported container size (nBlockAlign / nChannels). Real-world examples:

Adobe Audition's "24-bit packed int (type 1, 20-bit)"

nChannels = 2, nBlockAlign = 6, wBitsPerSample = 20

http://www-mmsp.ece.mcgill.ca/Documents/AudioFormats/WAVE/Samples/AFsp/M1F1-int12-AFsp.wav is:

nChannels = 2, nBlockAlign = 4, wBitsPerSample = 12

http://www-mmsp.ece.mcgill.ca/Documents/AudioFormats/WAVE/Docs/multichaudP.pdf gives an example of:

nChannels = 2, nBlockAlign = 8, wBitsPerSample = 20

Aliases

scipy.io.wavfile._read_data_chunk