Specifying raw type to build_array()

[SamuelLarkin]

Hi,
I'm attempting to translate

sox -r {SAMPLE_RATE} -b {BITS} -e {ENCODING} -c {NUM_CHANNEL} -t raw  {raw_audio_file.name} {wav_audio_file.name}

to a call to build_array() but I can't seem to be able to specify the equivalent of -t raw.

Code

tfm2wav = sox.Transformer()
tfm2wav.set_input_format(
    file_type="raw",
    rate=SAMPLE_RATE,
    bits=BITS,
    channels=NUM_CHANNEL,
    encoding=ENCODING,
    )
tfm2wav.set_output_format(
      file_type="wav",
      rate=SAMPLE_RATE,
      bits=BITS,
      channels=NUM_CHANNEL,
      encoding=ENCODING,
      )

audio = tfm2wav.build_array(
    input_filepath=raw_audio_file.name,
    )
tfm2wav.build_file(
     input_array=audio,
     output_filepath=wav_audio_file.name,
     sample_rate_in=SAMPLE_RATE,
     )

Error Message

{
  "errorMessage": "list index out of range",
  "errorType": "IndexError",
  "requestId": "d89c028b-e894-4596-89fd-0a9f19017103",
  "stackTrace": [
    "  File \"/var/task/lambda_function.py\", line 103, in handler\n    raise e\n",
    "  File \"/var/task/lambda_function.py\", line 91, in handler\n    denoise(noisy_audio_file.name, denoised_audio_file.name)\n",
    "  File \"/var/task/lambda_function.py\", line 157, in denoise\n    audio = tfm2wav.build_array(\n",
    "  File \"/opt/python/sox/transform.py\", line 793, in build_array\n    encoding_out = [\n"
  ]
}

[ERROR]	2022-11-01T14:34:13.338Z	d89c028b-e894-4596-89fd-0a9f19017103	list index out of range

[ERROR] IndexError: list index out of range
Traceback (most recent call last):
  File "/var/task/lambda_function.py", line 103, in handler
    raise e
  File "/var/task/lambda_function.py", line 91, in handler
    denoise(noisy_audio_file.name, denoised_audio_file.name)
  File "/var/task/lambda_function.py", line 157, in denoise
    audio = tfm2wav.build_array(
  File "/opt/python/sox/transform.py", line 793, in build_array
    encoding_out = [END RequestId: d89c028b-e894-4596-89fd-0a9f19017103

[zoldaten]

just remove '[0]' in /opt/python/sox/transform.py", line 793

[jsalsman]

Can confirm in v1.4.1:

Python 3.10.11 (main, Apr  4 2023, 22:10:32) [GCC 12.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import sox
>>> tfm=sox.Transformer()
>>> tfm.set_input_format(file_type='raw', rate=16000, bits=16, 
...              channels=1, encoding='signed-integer')
>>> tfm.silence(min_silence_duration=0.25,
...     buffer_around_silence=True)
<sox.transform.Transformer object at 0x7fdb8e043460>
>>> pcm = tfm.build_array('static/audio-22212508.raw')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/runner/webrec/.pythonlibs/lib/python3.10/site-packages/sox/transform.py", line 793, in build_array
    encoding_out = [
IndexError: list index out of range
>>> 

just remove '[0]' in /opt/python/sox/transform.py", line 793

If you mean at https://github.com/rabitt/pysox/blob/master/sox/transform.py#L825 that doesn't work, sorry:

>>> pcm = tfm.build_array('static/audio-22212508.raw')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/runner/webrec/.pythonlibs/lib/python3.10/site-packages/sox/transform.py", line 829, in build_array
    raise ValueError("invalid n_bits {}".format(n_bits))
ValueError: invalid n_bits 0
>>>