Unable to Initialize Audio Capture Device in Whisper Library**

Issue Summary: Unable to Initialize Audio Capture Device in Whisper Library


  • Attempting to use the Whisper library for transcription with the provided stream.py script.
  • Successfully loaded the model (tiny.en) without errors.
  • Specified the correct --device_id for the C-Media CM108 Audio Controller.

Error Details:

  • Encounter the following error during execution:
    ERROR: Failed to initialized SDL: dsp: No such audio device
    RuntimeError: Failed to initialize audio capture device.

Steps Taken:

  1. Verified the correct --device_id using --list_audio_devices.
  2. Installed SDL library with sudo apt-get install libsdl1.2-dev.
  3. Checked and confirmed the C-Media CM108 Audio Controller as the default audio device with aplay -l.
  4. Tried running the script without specifying --device_id to use the default audio device.

Additional Observations:

  • SDL configurations were explored, but the issue persists.
  • Model loading and other aspects of the script seem to work correctly.


  • Raspberry Pi (Assuming based on username).
  • Whisper library version and other relevant information not provided.

Request for Assistance:

  • Seeking guidance from the community to resolve the SDL initialization issue and successfully capture audio with Whisper.
  • Open to suggestions, troubleshooting steps, or insights from users experienced with Whisper on similar environments.