audio_afe_vcmd

# Audio AFE Voice Command Example

## Introduction

This example extends [audio_afe](../audio_afe) with MultiNet voice command detection. After wake word detection, the device listens for voice commands within a timeout window and logs recognized commands.

Flow: `audio_recorder` (AFE + wakeup + vcmd) → `audio_feeder` loopback, with wakeup prompt playback on `WAKEUP_START`.

## Hardware Requirements

Same as [audio_afe](../audio_afe). Tested on boards with microphone array and speaker, such as **ESP32-S3-KORVO2-V3** or **ESP32-S31-KORVO1**.

## Feature Description

Compared to `audio_afe`, this example additionally enables:

- `AV_PROCESSOR_AFE_FLAG_VCMD_DETECT_ENABLE` in AFE config
- WakeNet model: `Hi,乐鑫` (`CONFIG_SR_WN_WN9_HILEXIN`)
- MultiNet model: `mn7_cn` (`CONFIG_SR_MN_CN_MULTINET7_QUANT`)

### Usage Flow

1. Say the wake word (e.g. "Hi,乐鑫") → prompt tone plays
2. Within the vcmd timeout window (default 5 s), speak a MultiNet command
3. Recognized commands are logged as `VCMD_DETECTED`
4. Processed audio continues to loop back through the speaker

## Usage Instructions

1. Run `. ./prebuild.sh` and select the development board
2. Select a board, e.g. `idf.py gen-bmgr-config -b esp32_s31_korvo1`
3. Build and flash: `idf.py flash monitor -p PORT`
4. SPIFFS prompt files are shared from [audio_afe/spiffs](../audio_afe/spiffs)

## Configuration Parameters

- Sample rate: 16000 Hz
- Bit depth: 16 bit
- Channels: 1 (recorder output)
- Frame duration: 20 ms
- Vcmd timeout: 5000 ms (from `DEFAULT_AV_PROCESSOR_AFE_CONFIG()`)
- MultiNet language: `cn`
- Prompt file: `file://spiffs/I_comeon_wakeup.wav`