add arg for sampling rate

[karasjoh000]

Some input devices give errors (e.g. scarlett ) if frequency is not set to device frequency.

python -m sounddevice
#    0 HDA Intel PCH: ALC1220 Analog (hw:0,0), ALSA (2 in, 8 out)
#    1 HDA Intel PCH: ALC1220 Digital (hw:0,1), ALSA (0 in, 2 out)
#    2 HDA Intel PCH: ALC1220 Alt Analog (hw:0,2), ALSA (2 in, 0 out)
#    3 HDA NVidia: HDMI 0 (hw:1,3), ALSA (0 in, 8 out)
#    4 HDA NVidia: PHL 272P7VU (hw:1,7), ALSA (0 in, 2 out)
#    5 HDA NVidia: HDMI 2 (hw:1,8), ALSA (0 in, 8 out)
#    6 HDA NVidia: HDMI 3 (hw:1,9), ALSA (0 in, 8 out)
#    7 HDA NVidia: HDMI 4 (hw:1,10), ALSA (0 in, 8 out)
#    8 HDA NVidia: HDMI 5 (hw:1,11), ALSA (0 in, 8 out)
#    9 HDA NVidia: HDMI 6 (hw:1,12), ALSA (0 in, 8 out)
#   10 Scarlett 2i2 USB: Audio (hw:2,0), ALSA (2 in, 0 out)   <-- attempting to use this one
#   11 sysdefault, ALSA (128 in, 128 out)
#   12 front, ALSA (0 in, 8 out)
#   13 surround21, ALSA (0 in, 128 out)
#   14 surround40, ALSA (0 in, 8 out)
#   15 surround41, ALSA (0 in, 128 out)
#   16 surround50, ALSA (0 in, 128 out)
#   17 surround51, ALSA (0 in, 8 out)
#   18 surround71, ALSA (0 in, 8 out)
#   19 iec958, ALSA (0 in, 2 out)
#   20 spdif, ALSA (0 in, 2 out)
# * 21 default, ALSA (128 in, 128 out)
#   22 dmix, ALSA (0 in, 2 out)

whisper-ctranslate2 --live_transcribe True --live_input_device="10" --device=cuda --model medium
# Expression 'paInvalidSampleRate' failed in 'src/hostapi/alsa/pa_linux_alsa.c', line: 2048
# Expression 'PaAlsaStreamComponent_InitialConfigure( &self->capture, inParams, self->primeBuffers, hwParamsCapture, &realSr )' failed in 'src/hostapi/alsa/pa_linux_alsa.c', line: 2718
# Expression 'PaAlsaStream_Configure( stream, inputParameters, outputParameters, sampleRate, framesPerBuffer, &inputLatency, &outputLatency, &hostBufferSizeMode )' failed in 'src/hostapi/alsa/pa_linux_alsa.c', line: 2842
# Consider specifying the language using `--language`. It improves significantly prediction in live transcription.
# �[32mLive stream device: �[37mScarlett 2i2 USB: Audio (hw:2,0)�[0m
# �[32mListening.. �[37m(Ctrl+C to Quit)�[0m
# 
# �[93mQuitting..�[0m
# Traceback (most recent call last):
#   File "/lsiopy/bin/whisper-ctranslate2", line 8, in <module>
#     sys.exit(main())
#              ^^^^^^
#   File "/lsiopy/lib/python3.12/site-packages/src/whisper_ctranslate2/whisper_ctranslate2.py", line 206, in main
#     ).inference()
#       ^^^^^^^^^^^
#   File "/lsiopy/lib/python3.12/site-packages/src/whisper_ctranslate2/live.py", line 170, in inference
#     self.listen()
#   File "/lsiopy/lib/python3.12/site-packages/src/whisper_ctranslate2/live.py", line 158, in listen
#     with sd.InputStream(
#          ^^^^^^^^^^^^^^^
#   File "/lsiopy/lib/python3.12/site-packages/sounddevice.py", line 1440, in __init__
#     _StreamBase.__init__(self, kind='input', wrap_callback='array',
#   File "/lsiopy/lib/python3.12/site-packages/sounddevice.py", line 909, in __init__
#     _check(_lib.Pa_OpenStream(self._ptr, iparameters, oparameters,
#   File "/lsiopy/lib/python3.12/site-packages/sounddevice.py", line 2796, in _check
#     raise PortAudioError(errormsg, err)
# sounddevice.PortAudioError: Error opening InputStream: Invalid sample rate [PaErrorCode -9997]

# to get the sample rate: 
arecord -D "hw:2,0" -f cd 
#Recording WAVE 'stdin' : Signed 16 bit Little Endian, Rate 44100 Hz, Stereo
#arecord: set_params:1352: Sample format non available
#Available formats:
#- S32_LE

# with this PR
python -m src.whisper_ctranslate2.whisper_ctranslate2 --live_transcribe True --live_input_device=10 --live_input_device_sample_rate=44100 --device=cuda --model medium
# Consider specifying the language using `--language`. It improves significantly prediction in live transcription.
# Live stream device: Scarlett 2i2 USB: Audio (hw:2,0)
# Listening.. (Ctrl+C to Quit)
# ....
#  Shhh, shhh, shhh, shhh, shhh.
# ...............
#  This is a test. This is a test.
# 
#

[jordimas]

Looks good.
If you can please do:

"make dev"

and commit the changes. This will fix the formatting. Thanks

[karasjoh000]

@jordimas amended the changes from make dev

[jordimas]

Thanks, some of the tests are failing

----------------------------------------------------------------------
Traceback (most recent call last):
File "/home/runner/work/whisper-ctranslate2/whisper-ctranslate2/tests/testlive.py", line 8, in test_constructor
live = Live(
TypeError: Live.init() missing 1 required positional argument: 'input_device_sample_rate'

They probably just need to be update it

Thanks

[jordimas]

This has been merged 4d8a052

Thanks!