Streaming GIF encoder
This is built as part of the gifsockets project. It is forked from gif.js to allow for a streaming API and performance optimization.
Install the module with: npm install gif-encoder
// Create a 10 x 10 gif
var GifEncoder = require('gif-encoder');
var gif = new GifEncoder(10, 10);
// using an rgba array of pixels [r, g, b, a, ... continues on for every pixel]
// This can be collected from a <canvas> via context.getImageData(0, 0, width, height).data
var pixels = [0, 0, 0, 255/*, ...*/];
// Collect output
var file = require('fs').createWriteStream('img.gif');
gif.pipe(file);
// Write out the image into memory
gif.writeHeader();
gif.addFrame(pixels);
// gif.addFrame(pixels); // Write subsequent rgba arrays for more frames
gif.finish();
gif-encoder
exports GifEncoder
, a constructor function which extends readable-stream@~1.1.9
. This means you can use any streams1
/streams2
functionality. I will re-iterate what this means below.
// streams1
var gif = new GifEncoder(10, 10);
gif.on('data', console.log);
gif.on('end', process.exit);
// streams2
var gif = new GifEncoder(10, 10);
gif.on('readable', function () {
console.log(gif.read());
});
Constructor for a new GifEncoder
- width
Number
- Width, in pixels, of theGIF
to output - height
Number
- Height, in pixels, of theGIF
to output - options
Object
- Optional container for any options- highWaterMark
Number
- Number, in bytes, to store in internal buffer. Defaults to 64kB.
- highWaterMark
NEVER CALL .removeAllListeners()
. NO DATA EVENTS WILL BE ABLE TO EMIT.
We implement the GIF89a specification which can be found at
function (buffer) {}
Emits a Buffer
containing either header bytes, frame bytes, or footer bytes.
function () {}
Signifies end of the encoding has been reached. This will be emitted once .finish()
is called.
function (error) {}
Emits an Error
when internal buffer is exceeded. This occurs when you do not read
(either via .on('data')
or .read()
) and we cannot flush prepared data.
If you have a very large GIF, you can update
options.highWaterMark
via the Constructor.
function () {}
Emits when the stream is ready to be .read()
from.
function () {}
Emits when at the start and end of .writeHeader()
.
function () {}
Emits when at the start and end of .addFrame()
function () {}
Emits when at the start and end of .finish()
Set milliseconds to wait between frames
- ms
Number
- Amount of milliseconds to delay between frames
Set delay based on amount of frames per second. Cannot be used with gif.setDelay
.
- framesPerSecond
Number
- Amount of frames per second
Set the disposal code
- disposalCode
Number
- Alters behavior of how to render between frames- If no transparent color has been set, defaults to 0.
- Otherwise, defaults to 2.
Values : 0 - No disposal specified. The decoder is
not required to take any action.
1 - Do not dispose. The graphic is to be left
in place.
2 - Restore to background color. The area used by the
graphic must be restored to the background color.
3 - Restore to previous. The decoder is required to
restore the area overwritten by the graphic with
what was there prior to rendering the graphic.
4-7 - To be defined.
Taken from http://www.w3.org/Graphics/GIF/spec-gif89a.txt
Sets amount of times to repeat GIF
- n
Number
- If
n
is -1, play once. - If
n
is 0, loop indefinitely. - If
n
is a positive number, loopn
times.
- If
Define the color which represents transparency in the GIF
.
- color
Hexadecimal Number
- Color to represent transparent background- Example:
0x00FF00
- Example:
Set the quality (computational/performance trade-off).
- quality
Positive number
- 1 is best colors, worst performance.
- 20 is suggested maximum but there is no limit.
- 10 is the default, provided an even trade-off.
Read out size
bytes or until the end of the buffer. This is implemented by readable-stream
.
- size
Number
- Optional number of bytes to read out
Write out header bytes. We are following GIF89a
specification.
Write out a new frame to the GIF.
- imageData
Array
- Array of pixels for the new frame. It should follow the sequence ofr, g, b, a
and be4 * height * width
in length.- If used with the options
palette
andindexedPixels
, then this becomes the index in the palette (e.g.0
forcolor #0
)
- If used with the options
- options
Object
- Optional container for options- palette
Array
- Array of pixels to use as palette for the frame. It should follow the sequence ofr, g, b, a
- At the moment, this must be used with
options.indexedPixels
- At the moment, this must be used with
- indexedPixels
Boolean
- Indicator to treatimageData
as RGBA values (false
) or indicies inpalette
(true
)
- palette
Write out footer bytes.
For performance in gifsockets, we needed to open up some lower level methods for fancy tricks.
Don't use these unless you know what you are doing.
We have a secondary internal buffer that collects each byte from writeByte
. This is to prevent create a new Buffer
and data
event for every byte of data.
This method empties the internal buffer and pushes it out to the stream
buffer for reading.
Internal store for imageData
passed in by addFrame
.
First part of addFrame
; runs setImagePixels(removeAlphaChannel(imageData))
and runs analyzePixels()
.
- imageData
Array
- Same as that inaddFrame
- options
Object
- Optional container for options- indexedPixels
Boolean
- Indicator to treatimageData
as RGBA values (false
) or indicies inpalette
(true
)
- indexedPixels
Reduces imageData
into a Uint8Array
of length 3 * width * height
containing sequences of r, g, b
; removing the alpha channel.
- imageData
Array
- Same as that inaddFrame
; array containingr, g, b, a
sequences.
Save pixels
as this.pixels
for image analysis.
- pixels
Array
- Same asimageData
fromaddFrame
GifEncoder
will mutate the original data.
Save palette
as this.userPalette
for frame writing.
- palette
Array
- Same asoptions.palette
fromaddFrame
Second part of addFrame
; behavior varies on if it is the first frame or following frame.
In either case, it writes out a bunch of bytes about the image (e.g. palette, color tables).
Third part of addFrame
; encodes the analyzed/indexed pixels for the GIF format.
Support this project and others by twolfson via donations.
http://twolfson.com/support-me
In lieu of a formal styleguide, take care to maintain the existing coding style. Add unit tests for any new or changed functionality. Lint via grunt and test via npm test
.
As of Nov 11 2013, Todd Wolfson has released all code differences since initial fork from gif.js to the public domain.
These differences have been released under the UNLICENSE.
At the gif.js time of forking, gif.js was using the MIT license.