Skip to content

Flutter audio recording plugin designed for real-time speech audio processing

License

Notifications You must be signed in to change notification settings

Picovoice/flutter-voice-processor

Repository files navigation

Flutter Voice Processor

GitHub release GitHub

Pub Version

Made in Vancouver, Canada by Picovoice

Twitter URL

YouTube Channel Views

The Flutter Voice Processor is an asynchronous audio capture library designed for real-time audio processing on mobile devices. Given some specifications, the library delivers frames of raw audio data (16-bit, mono PCM) to the user via listeners.

Table of Contents

Requirements

Compatibility

  • Flutter 1.20.0+
  • Android 5.0+ (API 21+)
  • iOS 13.0+

Installation

Flutter Voice Processor is available via pub.dev. To import it into your Flutter project, add the following line to your pubspec.yaml:

dependencies:
  flutter_voice_processor: ^<version>

Permissions

To enable recording with the hardware's microphone, you must first ensure that you have enabled the proper permission on both iOS and Android.

On iOS, open the Info.plist file and add the following line:

<key>NSMicrophoneUsageDescription</key>
<string>[Permission explanation]</string>

On Android, open the AndroidManifest.xml and add the following line:

<uses-permission android:name="android.permission.RECORD_AUDIO" />

See our example app for how to properly request this permission from your users.

Usage

Access the singleton instance of VoiceProcessor:

import 'package:flutter_voice_processor/flutter_voice_processor.dart';

VoiceProcessor? _voiceProcessor = VoiceProcessor.instance;

Add listeners for audio frames and errors:

VoiceProcessorFrameListener frameListener = (List<int> frame) {
    // use audio
}

VoiceProcessorErrorListener errorListener = (VoiceProcessorException error) {
    // handle error
}

_voiceProcessor?.addFrameListener(frameListener);
_voiceProcessor?.addErrorListener(errorListener);

Ask for audio record permission and start recording with the desired frame length and audio sample rate:

final int frameLength = 512;
final int sampleRate = 16000;
if (await _voiceProcessor?.hasRecordAudioPermission() ?? false) {
    try {
        await _voiceProcessor?.start(frameLength, sampleRate);
    } on PlatformException catch (ex) {
        // handle start error
    }
} else {
    // user did not grant permission
}

Stop audio capture:

try {
    await _voiceProcessor?.stop();
} on PlatformException catch (ex) {
    // handle stop error
}

Once audio capture has started successfully, any frame listeners assigned to the VoiceProcessor will start receiving audio frames with the given frameLength and sampleRate.

Capturing with Multiple Listeners

Any number of listeners can be added to and removed from the VoiceProcessor instance. However, the instance can only record audio with a single audio configuration (frameLength and sampleRate), which all listeners will receive once a call to start() has been made. To add multiple listeners:

VoiceProcessorFrameListener listener1 = (frame) { }
VoiceProcessorFrameListener listener2 = (frame) { }
List<VoiceProcessorFrameListener> listeners = [listener1, listener2];
_voiceProcessor?.addFrameListeners(listeners);

_voiceProcessor?.removeFrameListeners(listeners);
// or
_voiceProcessor?.clearFrameListeners();

Example

The Flutter Voice Processor app demonstrates how to ask for user permissions and capture output from the VoiceProcessor.

Releases

v1.1.0 - August 4, 2023

  • Numerous API improvements
  • Error handling improvements
  • Allow for multiple listeners instead of a single callback function
  • Upgrades to testing infrastructure and example app

v1.0.0 - December 8, 2020

  • Initial public release.