/hls-server

Middleware for serving HTTP Live Streaming (HLS) compatible media streams.

Primary LanguageJavaScriptMIT LicenseMIT

hls-server

JavaScript Style Guide Travis

Simple HTTP middleware for serving HTTP Live Streaming (HLS) compatible media streams.

Usage

First you need a compatible media stream (see Producing Streams).

Fast way:

require('hls-server')(8000)

Detailed way:

var HLSServer = require('hls-server')
var http = require('http')

var server = http.createServer()
var hls = new HLSServer(server, {
  path: '/streams',     // Base URI to output HLS streams
  dir: 'public/videos'  // Directory that input files are stored
})
server.listen(8000)

Producing Streams

HLS can only stream files that have been properly encoded and segmented. FFMPEG is great for this.
Here is how to do it with node-fluent-ffmpeg.

var ffmpeg = require('fluent-ffmpeg')

function callback() { // do something when encoding is done }

// Below is FFMPEG converting MP4 to HLS with reasonable options.
// https://www.ffmpeg.org/ffmpeg-formats.html#hls-2
fmpeg('input.mp4', { timeout: 432000 }).addOptions([
    '-profile:v baseline', // baseline profile (level 3.0) for H264 video codec
    '-level 3.0', 
    '-s 640x360',          // 640px width, 360px height output video dimensions
    '-start_number 0',     // start the first .ts segment at index 0
    '-hls_time 10',        // 10 second segment duration
    '-hls_list_size 0',    // Maxmimum number of playlist entries (0 means all entries/infinite)
    '-f hls'               // HLS format
  ]).output('public/videos/output.m3u8').on('end', callback).run()

To create segments from an existing RTMP stream, use the following node-fluent-ffmpeg command. You can expect several seconds of latency, depending on hardware.

var ffmpeg = require('fluent-ffmpeg')

// host, port and path to the RTMP stream
var host = 'localhost'
var port = '1935'
var path = '/live/test'

function callback() { // do something when stream ends and encoding finshes }

fmpeg('rtmp://'+host+':'+port+path, { timeout: 432000 }).addOptions([
    '-c:v libx264',
    '-c:a aac',
    '-ac 1',
    '-strict -2',
    '-crf 18',
    '-profile:v baseline',
    '-maxrate 400k',
    '-bufsize 1835k',
    '-pix_fmt yuv420p',
    '-hls_time 10',
    '-hls_list_size 6',
    '-hls_wrap 10',
    '-start_number 1'
  ]).output('public/videos/output.m3u8').on('end', callback).run()

Using In-Memory Streams

By default, this module assumes files are kept in a directory on the local filesystem. If you want to stream files from another source (or don't want to relate URL paths to filesystem paths), you can specify a provider in the options like so:

var hls = new HLSServer(server, {
  provider: {
    exists: function (req, callback) { // check if a file exists (always called before the below methods)
      callback(null, true)                 // File exists and is ready to start streaming
      callback(new Error("Server Error!")) // 500 error
      callback(null, false)                // 404 error
    },
    getManifestStream: function (req, callback) { // return the correct .m3u8 file
      // "req" is the http request
      // "callback" must be called with error-first arguments
      callback(null, myNodeStream)
      // or
      callback(new Error("Server error!"), null)
    },
    getSegmentStream: function (req, callback) { // return the correct .ts file
      callback(null, myNodeStream)
    }
  }
})

See src/fsProvider.js for the default provider using the local filesystem.

CLI Tool

This package includes a CLI tool that can be installed globally with npm install -g hls-server.

To use, navigate to the directory where your .ts files are stored and run hlsserver in a command prompt. This will start a server on port 8000. (Use hlsserver --help to see additional options.)

The CLI tool will efficiently make use of multiple processors via the cluster module and can be used as an example of how to use the base module in the same way.

Notes

To publish from an RTMP client like OBS, use a RTMP server like rtmp-server-nodejs to echo the stream (direct streaming from that module is being worked on).

NOTE: Transcoding live streams is very CPU-intensive. Most consumer hardware won't be able to handle transcoding more than a few streams.