Maestro voice detection sample using VIT

Description

Records data from microphone (alias dmic0 in devicetree) and detects voice commands from selected language model. Detected commands are printed via UART.

Language model may be changed via Kconfig using CONFIG_MAESTRO_EXAMPLE_VIT_LANGUAGE selection. For other configuration options, see example’s Kconfig and prj.conf.

This project requires an NXP board supported by the VIT library.

The example has to be modified if a new board needs to be added. Please create an issue in that case.

User Input/Output

• Input:

  None.

• Output:

  UART Output:

  • List of voice commands the model can detect (printed immediately after start)

  • <Specific voice command> if voice command was detected

  • Demo result: FAIL otherwise

Dependencies

• VIT library: https://www.nxp.com/design/design-center/software/embedded-software/voice-intelligent-technology-wake-word-and-voice-command-engines:VOICE-INTELLIGENT-TECHNOLOGY

Supported platforms

Currently tested for:

• RD_RW612_BGA.