Skip to content

Latest commit

 

History

History
188 lines (141 loc) · 6.43 KB

README.md

File metadata and controls

188 lines (141 loc) · 6.43 KB

RtAudio.js

RtAudio binding for Node

NPM

Node binding of the RtAudio C++ audio library. To see a detailed description of RtAudio visit it's homepage or Github repo.

An overview of the binding:

  • Access well-known audio APIs
    • Windows: WASAPI and DirectSound
    • Linux: PulseAudio
  • Probe available audio devices
  • Stream audio to output devices
  • Stream audio from input devices
  • Fully configurable audio streaming, allows configuring
    • sample rate
    • bit depth
    • frame size
    • number of channels
  • No additional library/software needed, besides an npm install

Installation

A single npm install or yarn add command is enough for Node between 14.x.x - 21.x.x and Electron between 11.x.x - 28.x.x on x64 Linux and x64 Windows. This should cover most of the installations but if you happen to use a different target/platform, you will have to build the binding yourself (see Building from source).

Note Only Windows and Linux are supported at the moment.

Installing for Node 14.x.x - 21.x.x

Install it using npm or yarn

npm install @hamitzor/rtaudio.js

or

yarn add @hamitzor/rtaudio.js

Installing for Electron 11.x.x - 28.x.x

If you'll be using the package with Electron, you'll have to set some environment variables before the installation command.

For example, for Electron v28.0.0

On bash:

export npm_config_runtime=electron
export npm_config_target=28.0.0

On powershell:

$env:npm_config_runtime = "electron"
$env:npm_config_target = "28.0.0"

On cmd:

set npm_config_runtime=electron
set npm_config_target=28.0.0

To see a complete list of Electron versions, see this registry.

These environment variables will help the installation command to pick the correct prebuilds. After setting these up, you can simply use npm or yarn

npm install @hamitzor/rtaudio.js

or

yarn add @hamitzor/rtaudio.js

As simple as that, no additional library/software required for installation. If you run into trouble during installation, don't hesitate to create an issue at Github.

Documentation

You can find the documentation here, also there is an RtAudio.js examples repo that demonstrates a few use cases.

Usage

This binding is pretty orthodox about staying loyal to the original RtAudio API, so using RtAudio's documentation and tutorials should mostly cover the usage of the binding as well. Almost all the method names and signatures are the same. There is a separate repository named RtAudio.js examples that demonstrates the usage of this binding with a couple of examples. Besides, here is a small example for a quick start:

const {
  RtAudio,
  RtAudioFormat,
  RtAudioErrorType,
  RtAudioStreamStatus,
  RtAudioApi,
} = require("@hamitzor/rtaudio.js");

const rtAudio = new RtAudio();

const defaultInputDevice = rtAudio.getDefaultInputDevice();
const defaultOutputDevice = rtAudio.getDefaultOutputDevice();

// Create a stream
rtAudio.openStream(
  { deviceId: defaultOutputDevice, nChannels: 1 }, // output stream parameters
  { deviceId: defaultInputDevice, nChannels: 1 }, // input stream parameters
  RtAudioFormat.RTAUDIO_SINT16, // PCM format
  48000, // sample rate
  1920, // buffer size
  null,
  // A callback that will be invoked when input is ready and/or output is needed.
  (output, input, nFrames, _time, status) => {
    // output and input are instances of Uint8Array of PCM samples.

    // To access individual PCM samples, you can convert Uint8Array
    // to the respective typed array. In this example it would be Int16Array
    // since we use RTAUDIO_SINT16:
    const data = Int16Array.from(input.buffer);
    data.at(0); // First samples
    data.at(1); // Second sample
    data.at(2); // Third sample
    data.at(3); // Fourth sample
    ...
    data.at(nFrames - 1); // Last sample

    // nFrames is most of the times the buffer size we provide
    // to the `openStream` function, i.e. 1920

    // If we've used 2 channels, the data would be twice in size and we would
    // access first and second channel's interleaved samples. For example:
    const data = Int16Array.from(input.buffer);
    data.at(0); // First sample of the first channel
    data.at(1); // First sample of the second channel
    data.at(2); // Second sample of the first channel
    data.at(3); // Second sample of the second channel
    ...
    data.at(nFrames * 2 - 2); // Last sample of the first channel
    data.at(nFrames * 2 - 1); // Last sample of the second channel

    // Note: RtAudio uses system's endianness.

    // For echoing, we can write input directly to the output buffer.
    output.set(input, 0);
  }
);

// Start the stream
rtAudio.start();

// On SIGINT, close the stream and exit
process.on("SIGINT", () => {
  rtAudio.closeStream();
  process.exit();
});

Building from source

If you don't use Node between 14.x.x - 21.x.x or Electron between 11.x.x - 28.x.x, or your platform is not x64 Linux or x64 Windows, the npm install command will attempt to build the binding from source. In that case you'll need to satisfy some prerequisites:

  • Your Node or Electron version should support N-API 4 and up (see this)
  • CMake
  • A proper C/C++ compiler toolchain
    • For Windows, MSVC should be enough
    • For Linux, GCC or Clang and make

Credits

This package uses the C++ library named RtAudio under the hood. To check it out, visit https://github.com/thestk/rtaudio.

License

MIT