That sounds interesting but it would be a whole lot more interesting if the page was itself an example of said effect!

I really like the fact that your "recipe" already includes support for prefers-reduced-motion. Parallax can really trigger motion sensitivity issues up to vomiting or several days of bed rest with severe migraine.

You did leave that part out of your recipe though. Would you mind adding it there, so that people who copy paste are more inclined to actually keep/use that part as well?

its beautiful! where can I track the compatibility?

and how do I know if I'm hitting the fallback?

updated it with info on what you're using and current baselines comp (Chrome/Edge 115+, Safari 26+, and Firefox behind a flag).

How does this compare to the classic css-native parallax effect? Before the scroll timeline APIs you'd use the "perspective" css property to create a container where the z plane is n pixels away from the screen, and then position each layer within it at a different z distance using transform: translateZ

That method is GPU accelerated too, so it is performant compared to some js solutions, and has worked well in every browser for around a decade

I like the idea of the scroll-timeline though, just keen to understand what the advantage is for this

OP here, thanks for asking. While the `perspective` technique works too, it has the downside of needing a careful combination of scroller elements and properties.

This approach adds a single class to the image container and that's it. Plus you can control many aspects of the animation such as entry/exit ranges, and make it control other properties like opacity or color, for example.

I know browser support is still lacking, but it will get there eventually. I'm not using this in production code yet, but I think it's useful to experiment with these new CSS APIs.

When using Z, if the element is close to the bottom of the page, or a very large Z, I found it to increase the length of the scroll bar unexpectedly. Or unexpectedly to me looking for parallax, it would make sense as a normal zoom or scale.

This method should still support GPU acceleration, as `transform` (or rotate/scale/etc.) is the only property being animated. The benefit of animation-timeline seems to be that it's much easier to set up than a CSS perspective context.

No doubt quite a few folk with the same question. Keen to understand performance tradeoffs.

Obvious comparison note would be that the "new" method currently enjoys somewhat limited browser support (no Firefox without a flag, and only since Safari 26)

I was wondering the same thing. That translateZ is a bit fiddly to get right, so I could believe this is a bit easier to use, maybe? And presumably this could be used for other properties besides position, like colors, opacity or blurs.

For people saying it's not working in any browser - do you have some kind of reduced motion preference setting turned on? I can imagine that would have an effect on something like this and it's definitely working in Chrome for me.

Yes... there's a media query in the codepen disabling animation for people with reduced motion enabled.

Enabling (layout.css.scroll-driven-animations.enabled) and refreshing the codepen gives a "we crashed this to prevent a crash from an infinite loop" clicking to allow the infinite loops allowed me to see the animation.

Fedora 44 Kernel: x86_64 Linux 7.0.10-201.fc44.x86_64 Firefox 151.0.2

The JS and bot check are for making additional functionality, beyond just showing the example, work easily. I.e. editing and sharing edits from a browser. If all you want is a static example, feel free to make it without these things.

firefox android does not support CSS animation-timeline, and firefox desktop needs layout.css.scroll-driven-animations.enabled. This probably should not be used for any critical features.

oh that's cool!

Can only see it on chrome though =/. I switched to Safari as the lesser of two data-harvesting evils. Or rather, with an iPhone I've already chosen my overlord. I also switched to Kagi. Trying to deGoogle myself.

I played around with this API some time ago. It’s simple and high-performance, but one feature I wish existed is damping. Scroll-driven animations are tied directly to the scroll timeline, so there’s no concept of the parallax object “catching up” to the scroll progress over, say, one second. From what I remember, `animation-timing-function` feels weird when you scroll, so it’s not the right solution. GSAP offers this, but it’s JS-only.

In a world where it's increasingly overlooked, I'm glad the author mentions disabling it respecting user settings. I do think it should be reversed and only enabled with the `@media (prefers-reduced-motion: no-preference)`, but that is the opinion of someone who gets negative value from animations and is bemused by how much dev and compute time is spent on them.

Great. I tried the Google examples a while ago and got nowhere with it, time for another go, within the netherworld of SVG, to map to several different layers.

Love the one-variable design where scale and translate stay in sync automatically, just wish Firefox would get off the flag already.

I was expecting a demo on the linked page itself. Interesting to let Codex or Claude Code do it :)

Using css perspective for parallax has been around for years and is much simpler code.

Slightly disorienting but mostly just gimmicky. Luckily the "scroll down and more stuff than just scroll i going on" trend has mostly faded.

Yes. I'm a little more sensitive than average, but not enough to turn off animations, and this is uncomfortable.

could this be combined with a sprite like image that shows a slightly different angle of the image with each step