2 ms·
> After applying our standard image optimizations (stripping EXIF data, running smush.it, etc), the average size drops to 49k (91%). After converting to WebP th
by pornel 16y ago
> After applying our standard image optimizations (stripping EXIF data, running smush.it, etc), the average size drops to 49k (91%). After converting to WebP the average drops to 22.8kb (42%)
This sounds like they've been converting JPEG to WebP. Conversion of one lossy format to another lossy format is going to lose even more information, so the final size is not really fair — it's a size of lower-quality image.
A fair comparison would use high-quality source (uncompressed or high-resolution image that is downsampled) for both algorithms, rather than comparing original-to-JPEG with JPEG-to-WebP.
- joshfraser 16y agoGreat point, except we've compared hundreds of images are haven't been able to detect a noticeable drop in quality.
- pornel 16y agoIt is possible that JPEGs could have been even smaller with no noticeable drop in quality either. You can't fairly compare it with recompression of JPEG to lower-quality JPEG, as this JPEG encoding its own artifacts will amplify them. However, compression from original to same file size or quality as WebP could narrow gap between JPEG and WebP.
- dchest 16y agohttp://code.google.com/speed/webp/docs/c_study.html http://code.google.com/speed/webp/docs/c_study.html "Three compression methods, WebP, JPEG 2000 and Re-JPEG, were applied to the 900,000 JPEG images contained in the data set. JPEG images were re-compressed with Re-JPEG so that each was as close as possible to a target peak signal-to-noise ratio (PSNR) value."
- jacobolus 16y agoThat assumes that PSNR is a good measure of perceived image quality. I don’t buy it. (Which isn’t to say that WebP isn’t better, but only that that analysis is IMO severely flawed.)