sdl-speaker

0.4.1 • Public • Published

node-sdl-speaker

Output PCM stream data to speaker with SDL2. Especially for realtime audio stream.

Build Status NPM

Features

  • Stream Support. You can pipe a pcm stream to audio channel.
  • Multi Channels Support. You can open multi channels, it will mix all channels via SDL_MixAudioFormat.
  • Low Latency.

Installing / Getting started

$ npm install sdl-speaker

Example

const Speaker = require('sdl-speaker');
// Init SDL Speaker.
Speaker.init();
 
// Register a channel for write buf.
const testChannel = Speaker.register('test');
 
// pipe a pcm data to stream;
fs.createReadStream('test.pcm').pipe(testChannel);
 
// When you are ready..
// Start play audio
Speaker.start();

Api Reference

Speaker.init([ option ])

Init SDL. option is optional.

  • channels: Channel count. default: 2.
  • samplesRate: Sample rate for every channel. default: 44100.
  • samplesPerFrame: Samples per frame. default: 1024.

Speaker.start()

Start play audio. Read audio data from a ring buffer. If data in ring buffer less than samplesPerFrame, fill 0.

Speaker.stop()

Close and Destroy SDL player.

Speaker.pause()

Pause SDL player.

Speaker.resume()

Resume SDL player.

Speaker.detach(name)

Detach audio source from SDL.

Speaker.cleanAll()

Clean all audio sources's buffer.

Speaker.register(name) -> audioChannel

Register a audio source for write buffer. The name must unique.

audioChannel.write(buffer)

Write audio buffer to ring buffer. You can write before speaker start.

audioChannel.clean()

Clean this audio source's buffer.

TODO

  • More Test.
  • Fix Travis-CI build failed.
  • Pipe stream to audioSource.
  • Customise rbuf capability.
  • Self contained SDL2.

Package Sidebar

Install

npm i sdl-speaker

Weekly Downloads

5

Version

0.4.1

License

MIT

Unpacked Size

8.5 MB

Total Files

98

Last publish

Collaborators

  • 0nese7en
  • sharpbai