Note, 1.8.24 – this DRAFT is new and public but won’t show up as a ‘new post’ because I have temporarily backdated it to 2014.
In this previous post I looked at scaling and cropping images within WordPress, and the effects on image (file) size and quality. Here I’m looking at compressing images via a WordPress plug-in (W3 Total Cache) which saves them in webp, rather than jpg, format.
These test images are from Green Path. One of each pair is compressed, typically by 80%, the other not; and I’m not saying whether the unpaired images are compressed or not.
All of them were uploaded to the media library at a width of 1200 pixels and have been placed on this page without any resizing, but WordPress itself resizes all images on the fly, according to screen size and browser settings (as described in my previous post) so they can appear on screen at anywhere between 150 and 1500 px wide. (How can you check? Screenshot it and open the screenshot in an image editor.)
And, of course, the lightbox will usually show you a larger version.
It’s complicated. Sigh.
1 
2a 
2b 
3 
4a 
4b 
5a 
5b 
6 
I picked the difference between the paired images on my desktop, but only by looking much harder at image quality than most people would. On my laptop (smaller but newer screen) I couldn’t pick the difference with any confidence.
The variation in quality (brightness, saturation, sharpness) from one image to the next (in their original (edited) forms) is far greater than the difference between webp and jpg versions of any single image.
Conclusion: the speed gain outweighs any slight loss of quality, so I should go ahead and use webp until/unless compatibility issues emerge.