Why We Don’t Need Google’s WebP
Much cheer, Google has announced yet another open source project that is designed to move the web forward and make browsing a faster, better experience. The new project is called WebP (pronounced “weppy” according to Wikipedia) and it’s designed to be a new image format for the Internet that makes file sizes smaller and thus quicker to download.
The highlights of this new image format are:
- Lossy compression (like JPEGs)
- Uses block prediction to ‘guess’ the missing colours
- Google claims that WebP can compress a standard JPEG image by 39% and retain the same picture quality
- Still under development and no browser (or image editing software) currently supports this format
- Alpha support is planned (enables transparency)
If you want to see some examples of how much WebP can compress an image without affecting the quality, Google have put up a WebP comparison page here.
However, feedback from the community hasn’t been particularly great so far with many professional photographers claiming that the new format produces more blurry images, has colour loss and saturation problems. But who knows, since this is an open source project maybe these problems will be fixed before the WebP format is adopted en-masse.
Now the question that web designers and web developers need to ask ourselves is: Do we really need a new file format? Let’s be fair, if we properly optimized all our images before uploading then speed would be less of a problem. What’s the difference between saving the image as a WebP file if it’s just going to be the same size?
Another reason why so many people are perplexed about Google’s decision to introduce a new image format is because we already have the perfect alternative to JPEG in the form of JPEG-2000 (.jp2). While this is not very well known about outside the professional photography industry, JPEG-2000 provides superior compression while maintaining a high quality – effectively you can continue to compress a JPEG-2000 image well beyond the point where a normal JPEG image starts to take on a blocky effect.
JPEG vs JPEG-2000
So the WebP format claims to be able to offer a 39% compression while maintaining image quality, so in theory a 100kb JPEG could be replaced by a 60kb WebP image. How does that compare to using JPEG-2000 images?
On the left we have the standard JPEG image and on the right we have the corresponding JPEG-2000 image. Please note that since most browsers still do not support the JPEG-2000 format we have saved the image as a PNG file as this is a lossless format – it preserves the image EXACTLY as it looks as a JPEG-2000 image.
As you can see the lossy compression (which to be fair was designed in the late 80’s!) in the original JPEG images starts to distort very quickly. In technical circles the ‘blocky’ effect you start to see around the edges is known as ‘artifacts’, left over from the sampling algorithm as it tries to remove pixels and guess the colours instead.
On the other hand the JPEG-2000 images retain much of their quality even when the file size has been compressed by over 75% of the original image!
We would even argue that the quality of a 30.2kb JPEG-2000 image is virtually the same as the 61.2 kb JPEG image – which is just over 50% compression. What do you think? Can you tell much difference between the two?
Agriya does appreciate the initiative Google is taking to try and get web developers to use images with smaller file sizes, but when we already have the existing technology that can deliver 50% compression with the same image quality, shouldn’t they be focusing their resources on promoting it rather than trying to reinvent the wheel?
Add your thoughts below…