video-frames

Client-side video frames extraction as base64 encoded images.

Demo / CodePen

:warning: Doesn't work in Safari on iOS

From Apple Developer Documentation,

The image object can be an img element, a canvas element, or a video element. Use of the video element is not supported in Safari on iOS, however.

Installation

npm

npm install video-frames

CDN

<script src="https://cdn.jsdelivr.net/npm/video-frames@1/dist/videoframes.umd.min.js"></script>

or

<script src="https://unpkg.com/video-frames@1"></script>

Usage

const videoFrames = require('video-frames');

videoFrames({
	// Big Buck Bunny (1080p 60fps)
	url: 'http://distribution.bbb3d.renderfarming.net/video/mp4/bbb_sunflower_1080p_60fps_normal.mp4',
	// Extract 10 evenly spaced (time-wise) frames
	count: 10
}).then((frames) => {
	// frames is an array of objects
	// [
	//   {
	//     offset: (timestamp in seconds)
	//     image: (base64 encoded image)
	//   },
	//   ...
	// ]
});

API

videoFrames(options)

Returns a Promise for when all frames have been extracted. There are a few properties that can be set in options.

options

• url (required)

  Default value: empty

  The URL (self, remote [an Access-Control-Allow-Origin header must be present in case of a remote URL], or blob) of the source video from which the frames are to be extracted. Since the video element is used in the extraction process, the allowed formats are the ones that are playable in it. You can search for the supported formats on caniuse.com/?search=video%20format

• width

  Default value: 128

  Width of the extracted frames in pixels. If no value for width is set, but a value for height is set, then the width will be calculated using the video dimensions.

• height

  Default value: auto

  Height of the extracted frames in pixels. If not set, height is calculated automatically from the value of width using the video dimensions.

• format

  Default value: image/png

  MIME type of the extracted frames. Since the canvas element is used for drawing the frames and toDataURL(format) is used for reading them as base64 encoded images, the allowed MIME types are the ones that are supported by toDataURL.

  From MDN,

  toDataURL(type)

  ...

  type

  A string indicating the image format. The default type is image/png; this image format will be also used if the specified type is not supported.

  So, if a type is not supported, it will fall back to image/png.

• startTime

  Default value: 0

  Start timestamp (in seconds) of the range from where the frames are to be extracted. It will be ignored if a valid value for offsets is set.

• endTime

  Default value: Video Duration

  End timestamp (in seconds) of the range from where the frames are to be extracted. It will be ignored if a valid value for offsets is set.

• count

  Default value: 1

  Number of frames to be extracted from the range set by startTime and endTime. The frames are extracted from evenly spaced timestamps across the range. It will be ignored if a valid value for offsets is set.

• offsets

  Default value: []

  Array of timestamps (in seconds) to extract frames at. If a valid value for offsets is set, startTime, endTime, and count are ignored.

• onLoad

  Default value: false

  Function to be called when the source video has loaded and the extraction process has started.

  onLoad: () => { console.log('video loaded') }

• onProgress

  Default value: false

  Function to be called on every successful frame extraction.

  onProgress: (framesExtracted) => { console.log(`${framesExtracted} frames extracted`) }

  onProgress: (framesExtracted, totalFrames) => { console.log(`${framesExtracted} of ${totalFrames} frames extracted`) }

License

MIT © Utkarsh Verma