WebP - Image Format for the Web

The loading speed has long been one of the most important factors when optimizing a website, with visitors as well as search engines regarding the time it takes to call up a page as an important evaluation criterion. A lean and compressed code or the use of caching mechanisms are part of the tried and tested solutions for a fast Web presence, as is the use of compressed images. The latter – as design and background elements – are included in many projects in large numbers, and their effect on the loading time is often underestimated.

The search engine giant Google among others, tackled this problem, and it presented its own license-free  image format for faster websites in 2010 - WebP. But what exactly is behind the web format that arose from the video codec VP8? And how does it do in direct comparison with established variables such as JPEG?

On 30 September 2010, Google announced the release of a new open standard for lossy compression of 24-bit graphics on the Web. The video format VP8 served as a model for the format with the name WebP, and this was developed by the company On2 Technologies, which Google had bought up in the same year. In the period following, Google extended the WebP format with features such as a lossless compression mode, transparency (alpha channel), and animations. Chrome has supported the compression format since the end of 2010, as has the browser Opera. But there are still a variety of browsers to date – such as Safari and Firefox – which have only supported WebP experimentally up to now, and do not offer support with their regular releases. According to its own information, Microsoft is currently working on support for its browser Edge.

All software for the processing and presentation of WebP is under the BSD license as standard. Google uses the format, for example, in the Chrome Web Store, and in the mobile Google+ apps.

WebP has always served the purpose of reducing images on the Web to as small a file size as possible. According to information from Google, images and graphics in WebP format are around 30 per cent smaller than PNG or JPEG files with the same image quality. While both these classic Web formats rely on different compression methods – PNG is lossless and JPEG lossy – WebP allows both options. Thanks to this flexibility, the format is suitable for photos as well as small images and graphics. The compression features as well as the other central characteristics of the WebP format can be summarized as follows:

• Compression (lossy): Lossy compression with WebP uses keyframe coding of VP8.
• Compression (lossless): Lossless compression with WebP is based on various techniques that transform the parameters and image data. The LZ77 algorithm is also used here.
• Transparency: The 8-bit alpha channel that the WebP format offers can also be used for lossy RGB compression.
• Metadata: WebP may include EXIF and XMP metadata, which is typically generated by cameras.
• Color profile: The Google format may contain embedded ICC profiles (data sets that describe the color spectrum).
• Animations: The WebP format enables image sequences to be saved.

Lossy compression with WebP uses the same method as the block-based VP8 codec, to predict frames – i.e. individual images. Each frame is therefore divided into smaller segments, which are also described as macro-blocks. Within these macro-blocks, the coder can predict redundant movement and color information and subtract these from the block. The result of this predictive compression consists of the remaining block information and the separate redundant information (typically including many zero values). It is quantified and entropy coded (presentation of characters through specific bit pattern). The quantification phase is the only process here in which the bits are actually discarded as lossy.

With lossless WebP compression, the first step involves transforming the image. For this, the following techniques can be used:

The transformed parameters and image data are entropy coded, whereby a variant of the LZ77 algorithm is used. This uses small values for spatially close pixels.

Thanks to support from animated images WebP is an appealing alternative to GIF or APNG. Strengths come into play here such as the color depth of 24 bit and the 8-bit alpha channel, as well as the high degree of compression. This enables the WebP format in contrast with the competitive formats in addition to the lossless as well as the lossy compression of the animations. Animated WebP is also convincing in terms of decoding: As WebP metadata saves whether each frame contains alpha values, the decoder does not have to convert each frame individually to receive this information.

The keyframe technology of the WebP format already mentioned shows its strengths, particularly in animations, where the key frames ensure high quality. However, for this reason, the decoding process is also more CPU-intensive than for GIF, for example.

JPEG or JPG is to the presentation of photos and large images on the World Wide Web what Google is to the world of search engines. Even a quarter of a century after its release, it is impossible to imagine the image format not being on the Web scene. The greatest advantage with regard to formats such as PNG – which is evenly distributed but predominantly used in smaller images and graphics such as logos, icons etc. – is without a doubt that JPEG ensures considerable memory saving thanks to lossy compression .

However, in this crucial detail WebP proves to be more efficient and more flexible. On the one hand, the Google format offers a lossless compression method as well as a lossy one, while on the other hand, the memory saving for images in the WebP format outperforms that for JPEG images of the same value. The fact that JPEG permits images with a maximum size of up to 65,535 x 65,535 pixels, while the limit of WebP is only 16,383 x 16,383 pixels, can be ignored in light of the field of use (Web).

Similar to the comparison between WebP and JPEG is the comparison between WebP and GIF, where the format from Google also demonstrates convincing advantaged with regard to quality and compression. However, the problem is with the lack of support for WebP from browsers such as Firefox or Microsoft Edge. To date, users of these web browsers do not see images and animations in WebP, which is why barely any website operators rely on the format.

Basic data for the comparison of “WebP and JPEG” in table form:

File in WebP format cannot yet currently be opened with the usual image viewing programs from Windows, Linux and macOS If you want to view animations or images which have been coded with WebP, there are various options.

If Chrome or Opera are installed on your system, you can simply drag the file in question into an open window on either browser. As these browsers support WebP natively, you can view the content you wish to open without requiring further software. However, this solution is not very convenient – particularly if you want to view several images or animations. If you do not want to have to work with various browser windows and switching between them, it's easier to use IrfanView, a freeware picture viewer which also offers WebP support after the installation of the official plug-in package. However, the program is only available for Windows.

Furthermore, in the official WebP developer area Google offers a range of its own solutions for working with WebP – above all the libwebp library, which you can use to implement the WebP coding and decoding in your own programs. Also, the following tools are included in the free downloadable collection:

• cwebp enables the compression of an image field with the help of WebP. Possible input formats include PNG, JPEG and TIFF. In the tools options, you can define whether to use lossless or lossy compression.
• dwebp is a tool with which you can transform existing WebP files into other formats such as PNG (standard option), BMP, TIFF or PGM.
• With vwebp, Google supplies its own solution for displaying WebP images and animations based on OpenGL.
• webpmux is the decisive tool for working with extended WebP. You can create animations in WebP format, for example, with the program, or extend image files with metadata and ICC profiles.
• With gif2webp you convert existing GIF files into WebP. As with regular image compression, lossless as well as lossy coding is possible.