Image Load Lab measures how image formats, compression settings and loading techniques behave on a documented test bench, then publishes the tables, the figures and the raw CSV behind every number. Twenty CC0 photographs, real screenshots, Chrome on a throttled phone profile, and this WordPress site itself are the test material.
At the same SSIMULACRA2 score, AVIF was smaller than MozJPEG for all 20 CC0 photos (median −29.4 %), while WebP was not smaller at the median. Screenshots told a different story.
WebP vs AVIF vs JPEG at Equal Quality: 20 Photos Measured
Formats & Compression Tested Sep 29, 2026
Three images without dimensions gave an article a CLS of 0.15 on mobile, above the 0.1 'good' limit. width and height attributes or CSS aspect-ratio brought it to 0.
Images Without width and height: Layout Shift Measured
Performance Lab Tested Oct 17, 2026
Lazy-loading the 12 photos below the fold cut initial image bytes from 1,652 KB to 372 KB and LCP by 39 % on mobile. Lazy-loading the hero too cost 160–244 ms.
Lazy Loading Measured: Bytes Saved and the Cost of a Lazy Hero
Performance Lab Tested Oct 17, 2026
Tests of what Chrome actually does with images on a page: which files it downloads when they are hidden, inside closed elements or offered in several formats, how it chooses between srcset…
Client Hints on Image Requests: What Chrome Sends
Chrome 155 sent DPR, width and network hints with image requests only after Accept-CH, and to another origin only with Permissions-Policy delegation. Width was in device pixels.
Browser Behaviour 4 min read
Video Poster Images and LCP: Same as an img?
A video poster was discovered as early as an img, got the same priority and counted as the LCP element with the same timing, 1,316 ms against 1,320 ms. It cannot use srcset.
Browser Behaviour 3 min read
Native Lazy Loading vs lazysizes vs IntersectionObserver
All three cut LCP from 1,464 ms to 944 to 1,116 ms. Native lazy loading fetched one image more up front and was slowest after a long jump; it needs no JavaScript.
Browser Behaviour 4 min read
Carousel Slides: fetchpriority=low Did Not Help, loading=lazy Did
Hidden slides with fetchpriority=low still downloaded and LCP stayed at 5.0 s. With loading=lazy they were never requested, and LCP fell to 1.3 s.
Browser Behaviour 4 min read
Measured comparisons of JPEG, WebP, AVIF, PNG and SVG on a fixed corpus of 20 CC0 photographs, real screenshots and synthetic graphics. Each test reports the file size needed for the same…
AVIF Encoder Effort: Speed vs File Size on 20 Photos
AVIF effort 9 took 13.8 times as long as the default effort 4 for files only 0.4 % smaller. Effort 0 encoded a photo in 34 ms at a 10 % size cost.
Formats & Compression 4 min read
Chroma Subsampling 4:2:0 vs 4:4:4: When It Matters and What It Costs
Full colour (4:4:4) made JPEG photos 28 % larger for under 2 points of quality, while WebP's sharp YUV gave screenshots 7–9 points for about 10 % more bytes.
Formats & Compression 4 min read
PNG, oxipng or Lossless WebP for Screenshots? 8 Captures Measured
Lossless WebP made our screenshots 59 % smaller than Chrome's PNGs with identical pixels. oxipng saved 45 %; its slow zopfli mode added only 1.4 %.
Formats & Compression 4 min read
Which WebP Quality to Use for Photos, Screenshots and Graphics
Across 28 images, the lowest WebP quality reaching SSIMULACRA2 70 ranged from 30 for flat graphics to 95 for a screenshot. One setting does not fit all.
Formats & Compression 4 min read
Experiments on the image audits in Lighthouse, the tool behind PageSpeed Insights: when ‘Properly size images’, ‘Serve images in modern formats’, ‘Efficiently encode images’ and the lazy-loading audits fire, how much the…
Lighthouse LCP Request Discovery: What Each Hero Setup Gets
The insight passed only heroes with fetchpriority=high and flagged lazy and script-inserted ones, but simulated throttling reported the lazy hero 600 ms faster, the opposite of applied throttling.
Lighthouse Audits 3 min read
Does Lighthouse Catch Bad Alt Text? Ten Cases Tested
Lighthouse 13.5 failed only the image with no alt attribute. File names, the word image and alt text starting with Photo of all passed; redundant alt text was noted without affecting the score.
Lighthouse Audits 4 min read
Lighthouse ‘Avoid Enormous Network Payloads’: Where the Warning Starts
Lighthouse 13.5 warned from about 2.7 MB of transfer, with no extra penalty for heavier pages, and the performance score stayed at 75 from 1.6 to 13 MB while LCP reached 26 seconds.
Lighthouse Audits 3 min read
Which JPEGs Does Lighthouse Still Flag for Format or Compression?
Lighthouse 13.5 passed JPEGs up to quality 85 and flagged them from quality 90, plus PNG photos. A quality-80 JPEG passed although AVIF would have saved 54 %.
Lighthouse Audits 3 min read
Loading experiments on controlled test pages: lazy loading and its distance threshold, width and height for layout stability, fetchpriority and preload, image decoding and request audits. Every page is loaded several times…
How Far Below the Fold Does Chrome Start Loading Lazy Images?
Chrome 154 requested lazy images up to about 1,200 pixels below the fold on phone, tablet and desktop, and DevTools throttling did not change that distance.
Performance Lab 4 min read
We Audited Every Image on This Site: What We Found and Fixed
All 188 image loads were WebP and close to the right size on a phone. On desktop, a stale sizes value and a WordPress quirk sent oversized files; fixing them cut the category archive from 127 to 77 KB.
Performance Lab 5 min read
A Base64 Image Inside Your CSS: What It Does to First Paint
Embedding a 54 KB photo in the stylesheet delayed the first paint of the whole page from 404 to 788 ms on a throttled phone. The image itself appeared only 85 ms sooner.
Performance Lab 3 min read
What a Page Cache Did to This Site’s Server Response Time
Enabling the page cache cut our HTML response time from 358 to 420 ms down to about 193 ms, as fast as a static image. Uncached search and API requests did not change.
Performance Lab 4 min read
Tests of srcset, sizes, device pixel ratio, art direction with the picture element and CSS background images, run in Chrome at phone, tablet and desktop viewports. Each guide records which file the…
object-fit: cover and srcset: Blurry Images and Hidden Bytes
Setting sizes to the box width made object-fit: cover enlarge a file a third too narrow. Files cropped to the box were sharp and 35 to 54 % smaller than the best srcset choice.
Responsive Delivery 4 min read
sizes for Grid Layouts: What 100vw Costs in a Card Grid
In a one- to four-column grid, sizes=100vw made Chrome download up to 556 KB for four cards where 92 KB was enough. A matching sizes list or sizes=auto picked the right files.
Responsive Delivery 4 min read
How Many srcset Widths Do You Need? Seven Candidate Lists Compared
Across 18 photos, candidates every 100 px wasted 6 % extra bytes on average and every 200 px 21 %. WordPress's default list wasted 56 % because nothing sits between 300 and 768 px.
Responsive Delivery 4 min read
Image Width vs File Size: What Halving the Dimensions Saves
Halving an image's width cut the file to about 31 % of its size, not 25 %: across 20 photos, bytes grew with pixel count to the power of 0.84 in JPEG and WebP and 0.82 in AVIF.
Responsive Delivery 3 min read
What WordPress really does with images, measured on this WordPress 7.1.2 site: the files created for each upload and their sizes, the srcset and sizes markup it writes, how AVIF uploads behave…
Uploading AVIF to WordPress: What Our Shared Host Actually Did
WordPress 7.1.2 accepted our AVIF upload but created no image sizes, and the server sent it as application/octet-stream. Chrome still showed it, by sniffing.
WordPress Workflow 4 min read
When Alt Text Breaks srcset: A WordPress sizes Quirk
WordPress checks for an existing sizes attribute by searching the whole img tag, alt text included. Five of our images lost sizes that way and downloaded full-size files on desktop.
WordPress Workflow 4 min read
Our Theme Lazy-Loaded the First Archive Image: What WordPress Does Instead
Forcing loading=lazy on archive cards overrode WordPress. Letting WordPress 7.1.2 decide gave the first card fetchpriority=high, and it was requested 360 to 500 ms sooner.
WordPress Workflow 5 min read
Cache Headers on a Shared WordPress Host: What Each File Type Gets
JPEG, PNG, WebP, SVG, font and script files on our host were cacheable for a year, but AVIF uploads came with the wrong content type and max-age=0. How to check and fix your own.
WordPress Workflow 3 min read
Every test runs on one Windows 11 desktop (Intel Core i5-13400F, 32 GB RAM) with Chrome driven by Playwright, sharp and libvips for encoding, and SSIMULACRA2, DSSIM and SSIM for quality. Loading tests use a throttled phone profile and report the median of several runs. When something did not work the first time, the guide says so.