npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

@mmomtchev/ffmpeg

v1.0.0

Published

Full native ffmpeg bindings for Node.js with Streams support

Downloads

44

Readme

ffmpeg (w/avcpp) bindings for Node.js

GitHub License Node.js CI codecov npm (scoped) Test npm package

node-ffmpeg is a JavaScript wrapper around avcpp which is a C++ wrapper around the low-level C API of ffmpeg.

Overview

Unlike the myriad of other npm packages, this project does not launch the ffmpeg binary in a separate process - it loads it as a shared library and it uses its C/C++ API. It allows to do (mostly) everything that can be accomplished through the C API such as processing individual video frames and audio samples and interacting with multiple streams at the same time. The project relies on nobind17 for interfacing with the C/C++ API. It supports asynchronous operations that can run independently in different threads on multiple cores. Additionally, it takes advantage of the built-in multithreading of most included codecs.

Currently the project has a first beta release on npm.

Low-level C++ API

node-ffmpeg exports directly most of the C++ interface of avcpp which is a C++ wrapper for the low-level C API of ffmpeg. You should be aware that ffmpeg has a notoriously poorly documented and difficult to use C API. avcpp is a layer on top of it, which renders is somewhat more intuitive, but you should still have a very good understanding of its API in order to take full advantage of it. Unless you have previous experience with the C API, it is recommended to use the higher-level streams API which is much safer and easier to use and adds a minimal overhead.

Streams API

node-ffmpeg includes a Node.js Readable/Writable-compatible API that allows to work very naturally with video and audio data for everyone who has used Node.js streams. In this mode, a Demuxer is an object that provides a number of Readables that send compressed packet data. Each one of them can be fed into Node.js Transforms such as AudioDecoder or VideoDecoder to obtain a stream of uncompressed video frames or audio samples. These can be processed by other Transforms such as Filter, VideoTransform or AudioTransform that expect uncompressed data - or their underlying Buffers can be accessed. New streams can be created by creating video frames from scratch. There are various examples that show how the raw pixel data can be imported and exported from and to ImageMagick. Finally, raw uncompressed video and audio frames can be sent to a VideoEncoder or AudioEncoder for compression and then to a Muxer for multiplexing and time-based interleaving.

Memory safety

You should be aware that ffmpeg is a low-level C API and it is very unsafe to use. There are many cases in which a video frame will simply be identified by a raw C pointer - in this case trying to interpret 720p as 1080p will always end up with a segfault. avcpp adds a semi-safe layer on top of it, but mismatching stream parameters will still lead to a segfault. Producing a completely safe wrapper that never segfaults, no matter what the user does, is a gargantuan task that is currently not planned.

These bindings should never segfault if all the parameters are correctly checked and set up - but may easily segfault if these are mismatched - or if the asynchronous methods are reentered.

The current goal of node-ffmpeg is to simply be able to guarantee that a correct JavaScript code will never segfault on any input file.

The goal of the streams API is to make writing such correct code trivially simple and intuitive.

Performance

All the underlying heavy-lifting is performed by the ffmpeg C code - which means that unless you access and process the raw video and audio data, the performance will be nearly identical to that of ffmpeg when used from the command-line. This includes rescaling and resampling via the provided tools and using built-in filters. Background processing is provided via the libuv thread pool of Node.js - which means that when processing a stream, it is possible to automatically run each stage of the pipeline - demuxing, video decoding, audio decoding, filtering, video encoding, audio encoding and muxing on a separate physical processor core independently of V8/JavaScript.

If you need to access the actual pixel data or audio samples, then, depending on the processing, performance may be an order of magnitude lower.

Usage

Install

ffmpeg comes with prebuilt binaries for Windows, Linux and macOS on x64 platforms.

npm i @mmomtchev/ffmpeg

You can rebuild it from source using node-pre-gyp which will be automatically called by npm install. This will pull and build ffmpeg using conan which will leave a very large directory ${HOME}/.conan which can be safely deleted.

The package is a CommonJS package that should have identical interface across CJS, ES6 and TS environments:

const ffmpeg = require('@mmomtchev/ffmpeg');
const { Demuxer } = require('@mmomtchev/ffmpeg/stream');
import ffmpeg from '@mmomtchev/ffmpeg';
import { Demuxer } from '@mmomtchev/ffmpeg/stream';

Quickstart

A quick example for generalized video transcoding using the streams API.

import {
  Muxer,
  Demuxer,
  VideoDecoder, 
  VideoEncoder,
  Discarder,
  VideoTransform,
  VideoStreamDefinition
} from '@mmomtchev/ffmpeg/stream';

// Create a Demuxer - a Demuxer is an object that exports
// multiple Readable,
// it decodes the input container format and emits compressed data
const input = new Demuxer({ inputFile:'launch.mp4' });

// Wait for the Demuxer to read the file headers and to
// identify the various streams
input.on('ready', () => {
    // Once the input Demuxer is ready,
    // it will contain two arrays of Readable:
    // input.video[]
    // input.audio[]

    // We will be discarding the audio stream
    // (unless you discard these will keep piling in memory
    // until you destroy the demuxer object)
    const audioDiscard = new Discarder();
    // A VideoDecoder is a Transform that reads compressed video data
    // and sends raw video frames (this is the decoding codec)
    const videoInput = new VideoDecoder(input.video[0]);
    // A VideoDefinition is an object with all the properties of the stream
    const videoInputDefinition = videoInput.definition();

    // such as codec, bitrate, framerate, frame size, pixel format
    const videoOutputDefinition = {
      type: 'Video',
      codec: ffmpeg.AV_CODEC_H264,
      bitRate: 2.5e6,
      width: 320,
      height: 200,
      frameRate: new ffmpeg.Rational(25, 1),
      pixelFormat: new ffmpeg.PixelFormat(ffmpeg.AV_PIX_FMT_YUV422P)
    } as VideoStreamDefinition;

    // A video encoder is a Transform that reads raw video frames
    // and sends compressed video data (this is the encoding codec)
    const videoOutput = new VideoEncoder(videoOutputDefinition);

    // A VideoTransform is a Transform that reads raw video frames
    // and sends raw video frames - with different frame size or pixel format
    const videoRescaler = new VideoTransform({
      input: videoInputDefinition,
      output: videoOutputDefinition,
      interpolation: ffmpeg.SWS_BILINEAR
    });

    // A Muxer is an object that contains multiple Writable
    // It multiplexes those streams, handling interleaving by buffering,
    // and writes the to the output format
    const output = new Muxer({
      outputFile: 'video.mkv',
      outputFormat: 'mkv',
      streams: [videoOutput]
    });

    // The transcoding operation is completely asynchronous, it is finished
    // when all output streams are finished
    output.on('finish', () => {
      console.log('we are done!');
    });

    // These are the error handlers (w/o them the process will stop on error)
    input.video[0].on('error', (err) => console.error(err));
    input.audio[0].on('error', (err) => console.error(err));
    output.video[0].on('error', (err) => console.error(err));

    // This launches the transcoding
    // Demuxer -> Decoder -> Rescaler -> Encoder -> Muxer
    input.video[0].pipe(videoInput).pipe(videoRescaler)
        .pipe(videoOutput).pipe(output.video[0]);
    input.audio[0].pipe(audioDiscard);
});

More examples

You should start by looking at the unit tests which are also meant to be used as examples:

  • transcode.test.ts contains simple examples for transcoding audio and video
  • extract.test.ts contains a simple example for extracting a still from a video and importing it in ImageMagick
  • encode.test.ts contains a simple example for producing a video from stills using ImageMagick
  • rescale.test.ts contains a simple example for rescaling/resampling a video using ffmpeg's built-in libswscale
  • resample.test.ts contains a simple example for resampling audio using ffmpeg's built-in libswresample
  • streaming.test.ts contains examples for transcoding to various formats and sending the resulting data to a Node.js WriteStream
  • filtering.test.ts contains several examples for using ffmpeg's filters including overlaying text or Picture-in-Picture that are fast enough to be used in real-time

  • transcodeVideo.js is a low-level C/C++ style example that transcodes the video stream obtained from a container file without using the streams API
  • transcodeAudio.js is a low-level C/C++ style example that transcodes the audio stream obtained from a container file without using the streams API

  • data-is-beautiful/orbital-launches/ is a real data visualization generated with @mmomtchev/ffmpeg and magickwand.js
  • orbitron is a tool for making short animations of objects in the Solar System using the NASA/JPL Horizons API - it can produce mp4 and gif videos using @mmomtchev/ffmpeg

Supported pixel and audio formats

License

ISC License

Copyright (c) 2023, Momtchil Momtchev [email protected]

Permission to use, copy, modify, and/or distribute this software for any purpose with or without fee is hereby granted, provided that the above copyright notice and this permission notice appear in all copies.

THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.