Google Rolls Out AMP Lite, Turns to Machine Learning to Save Bandwidth on Google+

Image optimizations using Google AMP Cache, AMP Lite rolling out to bandwidth-constrained users in several countries. Google+ begun using machine learning and a new tech RAISR for images.

Share online:

Viewing high resolution images has also meant using lots of bandwidth, leading to slower loading speeds and higher data costs. Now to help everyone to see large images without worrying about the data, Google has turned to machine learning and a new technology called "RAISR", to display some of the beautiful photos on Google+.

Introduced in November, RAISR, uses machine learning to produce great quality versions of low-resolution images, and as Google says they've been "able to use up to 75 percent less bandwidth per image they've applied it to."

Google notes, "they've begun applying RAISR to more than 1 billion high-resolution images per week, appearing in the streams of a subset of Android devices," reducing users' total bandwidth by about a third.

Further, they says, the technology will roll out more boradly in the coming weeks.

Google+ RAISR Machine Learning Tech for Image Optimization

Accelerated Mobile Pages (AMP) format helps ensure that content reliably loads fast, but we can do even better. Today, the company highlights two improvements introduced recently include: "optimized image delivery," and "enabling content to be served more successfully in bandwidth-constrained conditions through a project called 'AMP Lite'."

says, that smart caching served through Google AMP Cache, is one of the key ingredients in the near instant AMP experiences users get in products like Google Search and Google News & Weather.

Google AMP Cache, which is available for free to anyone to use, employs PageSpeed Modules and Chrome Data Compression for image optimization and disregards "Cache-Control: no-transform" header.

The reduced image data is exhaustively compressed, and reduce bytes by 40%+ while not being noticeable to the user's eye. Here's the optimizations rundown:

  • Removing data which is invisible or difficult to see such as thumbnail and geolocation metadata.
  • For JPEG images, also reduced the quality and color samples "higher than necessary". To be exact, "JPEG quality is reduce to 85 and color samples to 4:2:0," explains Google — i.e., one color sample per four pixels.
  • Converted JPEG to WebP format leads to an additional 25%+ reduction in bytes "with no loss in quality."
  • Adding "srcset" applies to "amp-img" tags as well as resizing image to multiple dimensions, reduces the byte count further on devices with small screens.
  • Decrease JPEG images quality when indicated by a user or for very slow network conditions—for example, "JPEG image quality is reduced to 50 for Chrome users with Data Saver turned on—leads to another 40%+ byte reduction to JPEG images."

The following example shows the images before (left) and after (right) optimizations. Originally the image has 241,260 bytes, and after applying Optimizations 1, 2, & 4 it becomes 25,760 bytes. After the optimizations the image looks essentially the same, but 89% of the bytes have been saved.

Further optimizing for severely bandwidth constrained users, Google has launched AMP Lite, that apply all of the above optimizations to images to remove even more bytes from AMP pages.

Google notes, that a combined eduction of 45% in bytes across all optimizations is gained through optimiztions listed above.

In addition, external fonts is optimized as well by using the amp-font tag, involve "setting up font loading timeout to 0 seconds for immediate page display regardless of external font previously cached or not."

AMP Lite is rolling out for bandwidth-constrained users in countries such as Vietnam and Malaysia and for holders of low ram devices globally.