Guides
A guide here is a page that exists because something was measured against the tools themselves, or because an official source says something and we checked it on a stated date.
Each one opens with the finding, in a few sentences you can read on their own, and puts the method that produced it directly underneath — the sample, the command, and the numbers the prose is built from, so you can run it again and disagree with us.
Nothing goes up before its measurement exists. A guide with no working to show is a claim, and a claim is not worth a URL.
What happens when a large photo must reach 20 KB
By Sucheet BoppanaPublished Updated
A 1600×1067 synthetic test scene of 393,418 bytes came out of the 20 KB page at 20,111 bytes, and it got there by giving up pixels rather than quality: 524×350 at quality 68, in 1.83 s.
JPEG vs WebP at the same file size
By Sucheet BoppanaPublished Updated
On the photograph-like sample the two formats really did land on the same bytes, and WebP won: 41.22 dB PSNR and 0.9797 SSIM at the 100 KB ceiling against JPEG's 36.98 and 0.9568, and 36.92 and 0.9566 at 50 KB — within 0.06 dB of what the JPEG needed 96.97 KB to reach.