Create a gist now

Instantly share code, notes, and snippets.

What would you like to do?
Encode an ArrayBuffer as a base64 string
// Converts an ArrayBuffer directly to base64, without any intermediate 'convert to string then
// use window.btoa' step. According to my tests, this appears to be a faster approach:
// http://jsperf.com/encoding-xhr-image-data/5
/*
MIT LICENSE
Copyright 2011 Jon Leighton
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
function base64ArrayBuffer(arrayBuffer) {
var base64 = ''
var encodings = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/'
var bytes = new Uint8Array(arrayBuffer)
var byteLength = bytes.byteLength
var byteRemainder = byteLength % 3
var mainLength = byteLength - byteRemainder
var a, b, c, d
var chunk
// Main loop deals with bytes in chunks of 3
for (var i = 0; i < mainLength; i = i + 3) {
// Combine the three bytes into a single integer
chunk = (bytes[i] << 16) | (bytes[i + 1] << 8) | bytes[i + 2]
// Use bitmasks to extract 6-bit segments from the triplet
a = (chunk & 16515072) >> 18 // 16515072 = (2^6 - 1) << 18
b = (chunk & 258048) >> 12 // 258048 = (2^6 - 1) << 12
c = (chunk & 4032) >> 6 // 4032 = (2^6 - 1) << 6
d = chunk & 63 // 63 = 2^6 - 1
// Convert the raw binary segments to the appropriate ASCII encoding
base64 += encodings[a] + encodings[b] + encodings[c] + encodings[d]
}
// Deal with the remaining bytes and padding
if (byteRemainder == 1) {
chunk = bytes[mainLength]
a = (chunk & 252) >> 2 // 252 = (2^6 - 1) << 2
// Set the 4 least significant bits to zero
b = (chunk & 3) << 4 // 3 = 2^2 - 1
base64 += encodings[a] + encodings[b] + '=='
} else if (byteRemainder == 2) {
chunk = (bytes[mainLength] << 8) | bytes[mainLength + 1]
a = (chunk & 64512) >> 10 // 64512 = (2^6 - 1) << 10
b = (chunk & 1008) >> 4 // 1008 = (2^6 - 1) << 4
// Set the 2 least significant bits to zero
c = (chunk & 15) << 2 // 15 = 2^4 - 1
base64 += encodings[a] + encodings[b] + encodings[c] + '='
}
return base64
}
@tsenart

This comment has been minimized.

Show comment Hide comment
@tsenart

tsenart Jun 7, 2011

I kind of need the reverse. A base64 encoded utf8 string to an ArrayBuffer. How can I do that?

tsenart commented Jun 7, 2011

I kind of need the reverse. A base64 encoded utf8 string to an ArrayBuffer. How can I do that?

@jonleighton

This comment has been minimized.

Show comment Hide comment
@jonleighton

jonleighton Jun 7, 2011

I guess you would need to write this function sort of in reverse. So iterate over each 4 characters, converting them to three bytes which you append to the ArrayBuffer, and then with special handling for the padding at the end. I don't have code for that though :)

Owner

jonleighton commented Jun 7, 2011

I guess you would need to write this function sort of in reverse. So iterate over each 4 characters, converting them to three bytes which you append to the ArrayBuffer, and then with special handling for the padding at the end. I don't have code for that though :)

@n1k0

This comment has been minimized.

Show comment Hide comment
@n1k0

n1k0 Sep 10, 2011

As you just saved my life, I just wanted to say THANK YOU

n1k0 commented Sep 10, 2011

As you just saved my life, I just wanted to say THANK YOU

@joscha

This comment has been minimized.

Show comment Hide comment
@joscha

joscha Jan 18, 2013

kinda late, but here is the reverse

joscha commented Jan 18, 2013

kinda late, but here is the reverse

@joscha

This comment has been minimized.

Show comment Hide comment
@joscha

joscha Jan 18, 2013

And CryptoJS has a method to transform to a WordArray as well: CryptoJS.enc.Base64.parse

joscha commented Jan 18, 2013

And CryptoJS has a method to transform to a WordArray as well: CryptoJS.enc.Base64.parse

@tonikitoo

This comment has been minimized.

Show comment Hide comment
@tonikitoo

tonikitoo Jul 22, 2013

@joscha, the "reverse" (see comment https://gist.github.com/jonleighton/958841/#comment-733469) seems to be 3x slower than atob in my tests.

@joscha, the "reverse" (see comment https://gist.github.com/jonleighton/958841/#comment-733469) seems to be 3x slower than atob in my tests.

@sparrowprince

This comment has been minimized.

Show comment Hide comment
@sparrowprince

sparrowprince Aug 22, 2013

This is a pretty nifty method! Is there any chance you might add a license to it? I would really like to include it in a project of mine but am not allowed to do so without a proper license ;-)

This is a pretty nifty method! Is there any chance you might add a license to it? I would really like to include it in a project of mine but am not allowed to do so without a proper license ;-)

@lapsio

This comment has been minimized.

Show comment Hide comment
@lapsio

lapsio Oct 20, 2013

It's amazing, I'm working on >10 000 000 arrays (audio files) and performance was serious bottleneck, thanks

lapsio commented Oct 20, 2013

It's amazing, I'm working on >10 000 000 arrays (audio files) and performance was serious bottleneck, thanks

@ketting00

This comment has been minimized.

Show comment Hide comment
@ketting00

ketting00 Mar 19, 2014

This function is flaw. Everything turns out as "AAAAAAAAA" for small number like [0.000354254885, -0.521365849254, 0.002453687924].

This function is flaw. Everything turns out as "AAAAAAAAA" for small number like [0.000354254885, -0.521365849254, 0.002453687924].

@gfranko

This comment has been minimized.

Show comment Hide comment
@gfranko

gfranko May 28, 2014

Thank you very much, works like a charm! Would you mind adding a license to this?

gfranko commented May 28, 2014

Thank you very much, works like a charm! Would you mind adding a license to this?

@hogetsuyoshi

This comment has been minimized.

Show comment Hide comment
@hogetsuyoshi

hogetsuyoshi Oct 23, 2014

Thank you for this useful script. Which license is this? GPL? BSD?

Thank you for this useful script. Which license is this? GPL? BSD?

@andreagulino

This comment has been minimized.

Show comment Hide comment
@andreagulino

andreagulino Oct 23, 2014

You made my day.

You made my day.

@yamaneko1212

This comment has been minimized.

Show comment Hide comment
@yamaneko1212

yamaneko1212 Nov 24, 2014

You are my hero!

You are my hero!

@GeorgeGardiner

This comment has been minimized.

Show comment Hide comment
@GeorgeGardiner

GeorgeGardiner Feb 26, 2015

Boom, thanks.

Boom, thanks.

@mrmaffen

This comment has been minimized.

Show comment Hide comment
@mrmaffen

mrmaffen May 31, 2015

Thanks a lot! Would be awesome if you could state under which license you published this.

Thanks a lot! Would be awesome if you could state under which license you published this.

@thomthom

This comment has been minimized.

Show comment Hide comment
@thomthom

thomthom Aug 13, 2015

+1 for a license statement.

+1 for a license statement.

@AndyJ0127

This comment has been minimized.

Show comment Hide comment
@AndyJ0127

AndyJ0127 Mar 7, 2016

Thank you. It's works.

A question.
How max limit about arrayBuffer length ? Thanks again.

Thank you. It's works.

A question.
How max limit about arrayBuffer length ? Thanks again.

@wmaciel

This comment has been minimized.

Show comment Hide comment
@wmaciel

wmaciel Aug 5, 2016

Soooo, license?

wmaciel commented Aug 5, 2016

Soooo, license?

@hookenz

This comment has been minimized.

Show comment Hide comment
@hookenz

hookenz Aug 24, 2016

I tried this this just now on Chrome 52.0 thinking your approach would be faster. It's not, but it may once have been. Interestingly it's about 40ms slower than other approaches I tried. The most simplistic implementation is now very fast. Perhaps modern browsers have optimised certain routines more.

This is actually faster.

function uint8ToBase64(buffer) {
     var binary = '';
     var len = buffer.byteLength;
     for (var i = 0; i < len; i++) {
         binary += String.fromCharCode(buffer[i]);
     }
     return window.btoa( binary );
}

hookenz commented Aug 24, 2016

I tried this this just now on Chrome 52.0 thinking your approach would be faster. It's not, but it may once have been. Interestingly it's about 40ms slower than other approaches I tried. The most simplistic implementation is now very fast. Perhaps modern browsers have optimised certain routines more.

This is actually faster.

function uint8ToBase64(buffer) {
     var binary = '';
     var len = buffer.byteLength;
     for (var i = 0; i < len; i++) {
         binary += String.fromCharCode(buffer[i]);
     }
     return window.btoa( binary );
}
@Grantlyk

This comment has been minimized.

Show comment Hide comment
@Grantlyk

Grantlyk Dec 20, 2016

For those of you looking for an ES6 version

export function base64ArrayBuffer(arrayBuffer) {
  let base64 = '';
  const encodings = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/';

  const bytes = new Uint8Array(arrayBuffer);
  const byteLength = bytes.byteLength;
  const byteRemainder = byteLength % 3;
  const mainLength = byteLength - byteRemainder;

  let a;
  let b;
  let c;
  let d;
  let chunk;

  // Main loop deals with bytes in chunks of 3
  for (let i = 0; i < mainLength; i += 3) {
    // Combine the three bytes into a single integer
    chunk = (bytes[i] << 16) | (bytes[i + 1] << 8) | bytes[i + 2];

    // Use bitmasks to extract 6-bit segments from the triplet
    a = (chunk & 16515072) >> 18; // 16515072 = (2^6 - 1) << 18
    b = (chunk & 258048) >> 12; // 258048   = (2^6 - 1) << 12
    c = (chunk & 4032) >> 6; // 4032     = (2^6 - 1) << 6
    d = chunk & 63;        // 63       = 2^6 - 1

    // Convert the raw binary segments to the appropriate ASCII encoding
    base64 += encodings[a] + encodings[b] + encodings[c] + encodings[d];
  }

  // Deal with the remaining bytes and padding
  if (byteRemainder === 1) {
    chunk = bytes[mainLength];

    a = (chunk & 252) >> 2; // 252 = (2^6 - 1) << 2

    // Set the 4 least significant bits to zero
    b = (chunk & 3) << 4; // 3   = 2^2 - 1

    base64 += `${encodings[a]}${encodings[b]}==`;
  } else if (byteRemainder === 2) {
    chunk = (bytes[mainLength] << 8) | bytes[mainLength + 1];

    a = (chunk & 64512) >> 10; // 64512 = (2^6 - 1) << 10
    b = (chunk & 1008) >> 4; // 1008  = (2^6 - 1) << 4

    // Set the 2 least significant bits to zero
    c = (chunk & 15) << 2; // 15    = 2^4 - 1

    base64 += `${encodings[a]}${encodings[b]}${encodings[c]}=`;
  }

  return base64;
}

For those of you looking for an ES6 version

export function base64ArrayBuffer(arrayBuffer) {
  let base64 = '';
  const encodings = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/';

  const bytes = new Uint8Array(arrayBuffer);
  const byteLength = bytes.byteLength;
  const byteRemainder = byteLength % 3;
  const mainLength = byteLength - byteRemainder;

  let a;
  let b;
  let c;
  let d;
  let chunk;

  // Main loop deals with bytes in chunks of 3
  for (let i = 0; i < mainLength; i += 3) {
    // Combine the three bytes into a single integer
    chunk = (bytes[i] << 16) | (bytes[i + 1] << 8) | bytes[i + 2];

    // Use bitmasks to extract 6-bit segments from the triplet
    a = (chunk & 16515072) >> 18; // 16515072 = (2^6 - 1) << 18
    b = (chunk & 258048) >> 12; // 258048   = (2^6 - 1) << 12
    c = (chunk & 4032) >> 6; // 4032     = (2^6 - 1) << 6
    d = chunk & 63;        // 63       = 2^6 - 1

    // Convert the raw binary segments to the appropriate ASCII encoding
    base64 += encodings[a] + encodings[b] + encodings[c] + encodings[d];
  }

  // Deal with the remaining bytes and padding
  if (byteRemainder === 1) {
    chunk = bytes[mainLength];

    a = (chunk & 252) >> 2; // 252 = (2^6 - 1) << 2

    // Set the 4 least significant bits to zero
    b = (chunk & 3) << 4; // 3   = 2^2 - 1

    base64 += `${encodings[a]}${encodings[b]}==`;
  } else if (byteRemainder === 2) {
    chunk = (bytes[mainLength] << 8) | bytes[mainLength + 1];

    a = (chunk & 64512) >> 10; // 64512 = (2^6 - 1) << 10
    b = (chunk & 1008) >> 4; // 1008  = (2^6 - 1) << 4

    // Set the 2 least significant bits to zero
    c = (chunk & 15) << 2; // 15    = 2^4 - 1

    base64 += `${encodings[a]}${encodings[b]}${encodings[c]}=`;
  }

  return base64;
}
@ilaysener

This comment has been minimized.

Show comment Hide comment
@ilaysener

ilaysener Feb 14, 2017

I was looking to pure conversion from BufferArray to base64 (non-browser usage) and this has hit the spot perfectly. Much appreciated.

Cheers,
Ilay

I was looking to pure conversion from BufferArray to base64 (non-browser usage) and this has hit the spot perfectly. Much appreciated.

Cheers,
Ilay

@hookenz

This comment has been minimized.

Show comment Hide comment
@hookenz

hookenz Mar 2, 2017

@Grantlyk - that's better but still slower than the code I posted above.

hookenz commented Mar 2, 2017

@Grantlyk - that's better but still slower than the code I posted above.

@gertjanal

This comment has been minimized.

Show comment Hide comment
@gertjanal

gertjanal Aug 16, 2017

Thank you very much, this helped me get around the btoa() error "String contains an invalid character"!

Thank you very much, this helped me get around the btoa() error "String contains an invalid character"!

@otaroo

This comment has been minimized.

Show comment Hide comment
@otaroo

otaroo Sep 26, 2017

Thank you very much

otaroo commented Sep 26, 2017

Thank you very much

@stewartsims

This comment has been minimized.

Show comment Hide comment
@stewartsims

stewartsims Dec 14, 2017

Very handy. ES6 version works well. 👍

Very handy. ES6 version works well. 👍

@Activesite

This comment has been minimized.

Show comment Hide comment
@Activesite

Activesite Jan 25, 2018

Thank you very much. You save my day 👍

Thank you very much. You save my day 👍

@proton5000

This comment has been minimized.

Show comment Hide comment
@proton5000

proton5000 Feb 6, 2018

This is not work in EDGE and Internet Explorer 11. Please, help me :'(

This is not work in EDGE and Internet Explorer 11. Please, help me :'(

@TFrascaroli

This comment has been minimized.

Show comment Hide comment
@TFrascaroli

TFrascaroli Feb 7, 2018

@hookenz - I know this is quite a long way down the road (almost a year after your last coment), but I don't think you see what's going on here and I fear that other people may fall into the same trap. You see, OP is giving us a function to convert ANY arbitrary ArrayBuffer to a base64 string in javascript. Your solution works great aswell, except that it ONLY works when the underlying array of bytes comes from a UTF-8 encoded string, meaning that it can not be used to convert all byte arrays, only those that don't have "invalid" bytes in it, namely the infamous '\0' byte. So it doesn't really matter if it's faster if it doesn't work for all intended cases.

In case anyone is wondering why this is so, it's because of the way String.fromCharCode() works.
For a really small and homemade test, you can do this:

  • Assume hookenzBase64 is your function
  • Assume jonleightonBase64 is OP function
var b = new Uint16Array(25);
b.set([1,2,3,4,5], 3);
hookenzBase64(b.buffer); //[WRONG], gives "AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA="
jonleightonBase64(b.buffer); //[RIGHT] gives "AAAAAAAAAQACAAMABAAFAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA="

I hope this clarifies things for anyone else looking into these two functions ;)

TFrascaroli commented Feb 7, 2018

@hookenz - I know this is quite a long way down the road (almost a year after your last coment), but I don't think you see what's going on here and I fear that other people may fall into the same trap. You see, OP is giving us a function to convert ANY arbitrary ArrayBuffer to a base64 string in javascript. Your solution works great aswell, except that it ONLY works when the underlying array of bytes comes from a UTF-8 encoded string, meaning that it can not be used to convert all byte arrays, only those that don't have "invalid" bytes in it, namely the infamous '\0' byte. So it doesn't really matter if it's faster if it doesn't work for all intended cases.

In case anyone is wondering why this is so, it's because of the way String.fromCharCode() works.
For a really small and homemade test, you can do this:

  • Assume hookenzBase64 is your function
  • Assume jonleightonBase64 is OP function
var b = new Uint16Array(25);
b.set([1,2,3,4,5], 3);
hookenzBase64(b.buffer); //[WRONG], gives "AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA="
jonleightonBase64(b.buffer); //[RIGHT] gives "AAAAAAAAAQACAAMABAAFAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA="

I hope this clarifies things for anyone else looking into these two functions ;)

@TFrascaroli

This comment has been minimized.

Show comment Hide comment
@TFrascaroli

TFrascaroli Feb 7, 2018

@proton5000 - Tested on Edge/15.15063 and it works for me (with the really barebones test that I posted just now).
Also, it should work, according to CanIUse

@proton5000 - Tested on Edge/15.15063 and it works for me (with the really barebones test that I posted just now).
Also, it should work, according to CanIUse

@carpiediem

This comment has been minimized.

Show comment Hide comment
@carpiediem

carpiediem Feb 23, 2018

I'd like to validate the output string by rendering the image again. What do I need to add to the front to set it as the src attribute of an img tag? I tried this, but just got a broken image. (Failed to load resource: net::ERR_FILE_NOT_FOUND)

<img src="data:image/png;base64,[function result]">

Thanks.

I'd like to validate the output string by rendering the image again. What do I need to add to the front to set it as the src attribute of an img tag? I tried this, but just got a broken image. (Failed to load resource: net::ERR_FILE_NOT_FOUND)

<img src="data:image/png;base64,[function result]">

Thanks.

@eladkarako

This comment has been minimized.

Show comment Hide comment
@eladkarako

eladkarako Mar 9, 2018

alternative:

  //get ArrayBuffer from somewhere, for example:
  //FileReader's readAsArrayBuffer ".result"
  //and convert them to a BASE64-text.

function simpler_way(buffer){
  var tmp;

  tmp = (new TextDecoder("utf-8")).decode(buffer); //to UTF-8 text.
  tmp = unescape(encodeURIComponent(tmp));         //to binary-string.
  tmp = btoa(tmp);                                 //BASE64.
  return tmp;

//you can stop here.
//but if you need to return an ArrayBuffer of the BASE64 result,
//for example to be passed from a Worker back to a client, 
//by using 'postMessage' + Transferable object (much faster see: https://developer.mozilla.org/en-US/docs/Web/API/Transferable),
//uncomment lines below:
  
//tmp = (new TextEncoder("utf-8")).encode(tmp);    //to Uint8Array.
//tmp = tmp.buffer;                                //to ArrayBuffer.
//return tmp;
}

used in this repository,
as a Worker-FileReader example
(a no-upload file to Base64 converter )

eladkarako commented Mar 9, 2018

alternative:

  //get ArrayBuffer from somewhere, for example:
  //FileReader's readAsArrayBuffer ".result"
  //and convert them to a BASE64-text.

function simpler_way(buffer){
  var tmp;

  tmp = (new TextDecoder("utf-8")).decode(buffer); //to UTF-8 text.
  tmp = unescape(encodeURIComponent(tmp));         //to binary-string.
  tmp = btoa(tmp);                                 //BASE64.
  return tmp;

//you can stop here.
//but if you need to return an ArrayBuffer of the BASE64 result,
//for example to be passed from a Worker back to a client, 
//by using 'postMessage' + Transferable object (much faster see: https://developer.mozilla.org/en-US/docs/Web/API/Transferable),
//uncomment lines below:
  
//tmp = (new TextEncoder("utf-8")).encode(tmp);    //to Uint8Array.
//tmp = tmp.buffer;                                //to ArrayBuffer.
//return tmp;
}

used in this repository,
as a Worker-FileReader example
(a no-upload file to Base64 converter )

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment