Streaming speech recognition (Android)

NOTE

The Android version of the AidVoice SDK is currently in internal testing. Presently, only the Demo APP is available for download to experience the features.

Introduction

This case study demonstrates how to use the AidVoice SDK to implement real-time streaming recognition of audio information transmitted via a microphone on the Android system.

• Device: Rhino Pi-X1
• Microphone: Jabra Speak 410 Conference Speaker
• System: Android 13
• Model: SenseVoiceSmall

Supported Platforms

Platform	Running Mode
Rhino Pi-X1	Android 13

Prerequisites

1. Rhino Pi-X1 Hardware
2. Android 13 System
3. USB Microphone

Deployment Steps

Step 1: Microphone Recording Verification

Check if the USB microphone is recognized.

adb shell
# When the USB microphone is NOT plugged in:
lsusb # Output will be similar to:

# kalama:/ $ lsusb
# Bus 003 Device 001: ID 1d6b:0002
# Bus 001 Device 001: ID 1d6b:0002
# Bus 004 Device 001: ID 1d6b:0003
# Bus 002 Device 001: ID 1d6b:0003

# When the USB microphone IS plugged in:
lsusb # Output will include an additional device:

# aidlux@kalama:~$ lsusb
# kalama:/ $ lsusb
# Bus 003 Device 001: ID 1d6b:0002
# Bus 001 Device 001: ID 1d6b:0002
# Bus 003 Device 002: ID 0b0e:0412
# Bus 004 Device 001: ID 1d6b:0003
# Bus 002 Device 001: ID 1d6b:0003

Access the Android interface, open the Recorder APP, and start recording. You should see audio waveform activity as shown below:

Step 2: Demo APP Download

Download the AidVoice Android Demo APP

Step 3: Install Demo APP

Connect the PC and the board via a wired connection and install via ADB:

adb install aidvoice-release.apk

Step 4: Launch Demo APP Experience

It is recommended to use QtScrcpy screen mirroring software to control the Android interface.

Open the APP named ASR Real-time Transcription. You will see the following screen:

NOTE

Speech recognition performance is related to device sound pickup, noise reduction, and model capabilities. This Demo specifically showcases the model performance for real-time streaming recognition.