coast_audio

coast_audio is a high performance audio processing library written in Dart.
This package aims to provide low-level audio functionalities with no Flutter dependency.

Features

• Audio Buffer Management
• Ring Buffer Management
• Audio Node Graph System
  • Decoder Node
  • Mixer Node
  • Function Node
  • Volume Node
• Encoding and Decoding mechanism
  • Wav Audio Encoder/Decoder
• Wave Generation
  • Sine
  • Triangle
  • Square
  • Sawtooth
• Audio Device I/O
  • Playback
  • Capture
• Audio Format Management and Conversion

Supported Platforms

Android	iOS	macOS	Windows	Linux	Web
✅	✅	✅	⚠️	✅	❌

• Windows is not tested but it should work if the native library is compiled correctly.

How It Works

coast_audio is built on top of dart:ffi.
Thus, it can be used in any Dart environment that supports FFI.

Some of the functionalities are implemented in native code which uses miniaudio.
This repository contains pre-built binaries for each supported platform.

Installation

Add the following to your pubspec.yaml:

coast_audio: ^1.0.0

Android

1. Create a new directory android/src/main/jniLibs in your project.
2. Copy the {ABI}/libcoast_audio.so file from the native/prebuilt/android directory to the jniLibs directory.

macOS/iOS

Add the following to your Podfile:

target 'Runner' do
  ...
  pod 'CoastAudio', :git => 'https://github.com/SKKbySSK/coast_audio.git', :tag => '1.0.0'
end

Open the AppDelegate.swift file and add the following import and CoastAudioSymbolKeeper.keep() call:

import CoastAudio // 1. Add import

@UIApplicationMain
@objc class AppDelegate: FlutterAppDelegate {
  override func application(
    _ application: UIApplication,
    didFinishLaunchingWithOptions launchOptions: [UIApplication.LaunchOptionsKey: Any]?
  ) -> Bool {
    CoastAudioSymbolKeeper.keep() // 2. Add this line to prevent native symbols from being stripped (You can place this anywhere inside your iOS/macOS code)
    GeneratedPluginRegistrant.register(with: self)
    return super.application(application, didFinishLaunchingWithOptions: launchOptions)
  }
}

Linux

1. Copy the {ARCH}/libcoast_audio.so file from the native/prebuilt/linux directory to the linux/libs directory in your project.
2. Add the following to your linux/CMakeLists.txt:

install(FILES "linux/libs/${CMAKE_SYSTEM_PROCESSOR}/libcoast_audio.so" DESTINATION "${INSTALL_BUNDLE_LIB_DIR}" COMPONENT Runtime)

Windows

TODO

There is no pre-built native library for Windows yet.
You need to build it manually.

Usage

Generate Sine Wave

coast_audio provides various audio nodes to generate/process audio data easily.
This example code generates a 440hz sine wave by using FunctionNode.

// define the audio format with 48khz sample rate and stereo channels.
final format = AudioFormat(sampleRate: 48000, channels: 1, sampleFormat: SampleFormat.int16);

// create a sine wave function node with 440hz frequency.
final functionNode = FunctionNode(
  function: const SineFunction(),
  frequency: 440,
);

AllocatedAudioFrames(length: 1024, format: format).acquireBuffer((buffer) {
  // read the audio data from the function node to the buffer.
  functionNode.outputBus.read(buffer);

  // floatList contains sine wave audio data.
  final floatList = buffer.asFloat32ListView();
});

Generate Wav File

You can use WavAudioEncoder to encode audio data.

// define the audio format with 48khz sample rate and stereo channels.
final format = AudioFormat(sampleRate: 48000, channels: 1, sampleFormat: SampleFormat.int16);

// create a sine wave function node with 440hz frequency.
final functionNode = FunctionNode(
  function: const SineFunction(),
  frequency: 440,
);

final fileOutput = AudioFileDataSource(
  file: File('output.wav'),
  mode: FileMode.write,
);

// create a wav audio encoder.
final encoder = WavAudioEncoder(dataSource: fileOutput, inputFormat: format);

encoder.start();

final duration = AudioTime(10);
AllocatedAudioFrames(length: duration.computeFrames(format), format: format).acquireBuffer((buffer) {
  // read the audio data from the function node to the buffer.
  final result = functionNode.outputBus.read(buffer);

  // encode the audio data to the wav file.
  encoder.encode(buffer.limit(result.frameCount));
});

encoder.finalize();

Mixing Audio with Audio Node Graph System

AudioNode can be connected to other nodes to build an audio graph.
This example code demonstrates how to mix two sine waves and write to a wav file.

const format = AudioFormat(sampleRate: 48000, channels: 1);
final mixerNode = MixerNode(format: format);

// Initialize sine wave nodes and connect them to mixer's input
for (final freq in [264.0, 330.0, 396.0]) {
  final sineNode = FunctionNode(function: const SineFunction(), format: format, frequency: freq);
  final mixerInputBus = mixerNode.appendInputBus();
  sineNode.outputBus.connect(mixerInputBus);
}

AllocatedAudioFrames(length: 1024, format: format).bufferFrames.acquireBuffer((buffer) {
  // read the audio data from the function node to the buffer.
  functionNode.outputBus.read(buffer);

  // floatList contains mixed sine wave audio data.
  final floatList = buffer.asFloat32ListView();
});

Play, record and loopback

coast_audio provides AudioDevice class to handle audio device I/O.
Please see the following examples:

• Play audio from a file: player_isolate.dart
• Record audio to a file: recorder_isolate.dart
• Loopback audio from input to output: loopback_isolate.dart

Q&A

Can I use this package in Flutter?

Yes, you can use coast_audio in Flutter.

Most of coast_audio operation is synchronous and it may block the Flutter's main isolate.
So, it is recommended to use coast_audio in a separate isolate.

Please see the example app implementation for more details.

Can I use coast_audio in web?

In short, no,
You should use dart:web_audio instead.
But it may become available if the FFI is supported on the Web platform.