Picovoice Platform
Microcontroller Quick Start

Platforms

• Arm Cortex-M4
• Arm Cortex-M7

Requirement

• C99-compatible compiler

Picovoice Account & AccessKey

Signup or Login to Picovoice Console to get your AccessKey. Make sure to keep your AccessKey secret.

Quick Start

Setup

1. Clone the repository:

git clone --recurse-submodules https://github.com/Picovoice/picovoice.git

Usage

1. Include the public header files .
2. Link the project to an appropriate library file .
3. Construct the Picovoice object:

#define MEMORY_BUFFER_SIZE ${MEMORY_BUFFER_SIZE}

static const char* ACCESS_KEY = "${ACCESS_KEY}";

static uint8_t memory_buffer[MEMORY_BUFFER_SIZE] __attribute__((aligned(16)));

static const uint8_t *keyword_array = ${KEYWORD_ARRAY};
const float porcupine_sensitivity = 0.5f;

static void wake_word_callback(void) {
    // logic to execute upon detection of wake word
}

static const uint8_t *context_array = ${CONTEXT_ARRAY};
const float rhino_sensitivity = 0.75f;
const float rhino_endpoint_duration_sec = 1.0f;
const bool rhino_require_endpoint = true;

static void inference_callback(pv_inference_t *inference) {
    // `inference` exposes three immutable properties:
    // (1) `IsUnderstood`
    // (2) `Intent`
    // (3) `Slots`
    // ..
    pv_inference_delete(inference);
}

pv_picovoice_t *picovoice = NULL;

const pv_status_t status = pv_picovoice_init(
        ACCESS_KEY,
        MEMORY_BUFFER_SIZE,
        memory_buffer,
        sizeof(keyword_array),
        keyword_array,
        porcupine_sensitivity,
        wake_word_callback,
        sizeof(context_array),
        context_array,
        rhino_sensitivity,
        rhino_endpoint_duration_sec,
        rhino_require_endpoint,
        inference_callback,
        &picovoice);

if (status != PV_STATUS_SUCCESS) {
    // error handling logic
}

Pass in frames of audio to the pv_picovoice_process function:

extern const int16_t *get_next_audio_frame(void);

while (true) {
    const int16_t *pcm = get_next_audio_frame();
    const pv_status_t status = pv_picovoice_process(picovoice, pcm);
    if (status != PV_STATUS_SUCCESS) {
        // error handling logic
    }
}

Release resources explicitly when done with Picovoice:

pv_picovoice_delete(picovoice);

Create Custom Keyword and Context Models

1. Obtain the UUID of the chipset.
2. Go to Picovoice Console to create models for Porcupine wake word engine and Rhino Speech-to-Intent engine.
3. Select Arm Cortex-M as the platform when training the model.
4. Select appropriate board type.
5. Train your models.
6. Download your custom voice model(s).
7. Decompress the zip file. The model file is either .ppn for Porcupine wake word or .rhn for Rhino Speech-to-Intent. Both zip archives also contain a .h header file containing the C array version of the binary model.
8. Copy the contents of the arrays inside the .h header files and update the keyword_array and context_array values.

Non-English Languages

Use the corresponding library file (.a) to process non-English wake words and contexts. The library files for all supported languages are available on the Picovoice GitHub repository .

Demo

For the Picovoice mcu SDK, we offer demo projects for several evaluation boards to demonstrate how to use the Picovoice Platform on microcontrollers. The full list of supported boards is available on the Picovoice GitHub repository .

Setup

Clone the repository:

git clone --recurse-submodules https://github.com/Picovoice/picovoice.git

Usage

• Arduino Portenta H7 (Arm Cortex-M7)
• Arduino Nano 33 BLE Sense (Arm Cortex-M4)
• IMXRT1050-EVKB (Arm Cortex-M7)
• STM32F407G-DISC1 (Arm Cortex-M4)
• STM32F411E-DISCO (Arm Cortex-M4)
• STM32F769I-DISCO (Arm Cortex-M7)

Resources