Guetzli is an image encoder by Google currently in developmental stage, that aims to reduce file size by over 20 per cent, with significantly improved image quality. Guetzli-encoded images are typically 20-30% smaller than images generated by libjpeg, and unlike Google's other image compression projects (WebP, WebM), the new JPEG encoder is compatible with existing browsers, devices, photo editing apps, and the current JPEG standard.

The project codename, Guetzli is Swiss German for cookie; and was born out of Google Research office, Zurich.

While there are numerous ways of tweaking JPEG image quality and file size, Guetzli focuses mainly on the input of values stage, which allows for reduction of disordered data, thus resulting easily compressed data.

With normal JPEG encoding, this process usually reduces image colour gradients to single blocks of colour and often obliterates some details, but Guetzli uses a new model, Butteraugli, to determine which colours and details to implement.

Butteraugli is derived within hundreds of high-precision constants, which produces "approximates colour perception and visual masking in a more thorough and detailed way" than other encoders.

Albeit, encoding images with Guetzli, is slower than libjpeg due to the more involved input of value processes.

Guetzli has be open-sourced and freely made available to download on Github, so webmasters, graphic designers, and photographers can give it a try.

How Google's image encoder, Guetzli makes better compressed images

Guetzli is an image encoder by Google currently in developmental stage, that aims to reduce file size by over 20 per cent, with significantly improved image quality. Guetzli-encoded images are typically 20-30% smaller than images generated by libjpeg, and unlike Google's other image compression projects (WebP, WebM), the new JPEG encoder is compatible with existing browsers, devices, photo editing apps, and the current JPEG standard.

The project codename, Guetzli is Swiss German for cookie; and was born out of Google Research office, Zurich.

While there are numerous ways of tweaking JPEG image quality and file size, Guetzli focuses mainly on the input of values stage, which allows for reduction of disordered data, thus resulting easily compressed data.

With normal JPEG encoding, this process usually reduces image colour gradients to single blocks of colour and often obliterates some details, but Guetzli uses a new model, Butteraugli, to determine which colours and details to implement.

Butteraugli is derived within hundreds of high-precision constants, which produces "approximates colour perception and visual masking in a more thorough and detailed way" than other encoders.

Albeit, encoding images with Guetzli, is slower than libjpeg due to the more involved input of value processes.

Guetzli has be open-sourced and freely made available to download on Github, so webmasters, graphic designers, and photographers can give it a try.