It's really weird that the response to JPEG XL shipping is "yesss fuck WebP!"
Like, we've had AVIF for years, and it outperforms JPEG XL in almost all web use-cases.
It's really weird that the response to JPEG XL shipping is "yesss fuck WebP!"
Like, we've had AVIF for years, and it outperforms JPEG XL in almost all web use-cases.
@jaffathecake Wait, people dislike .webp? Why? Is it because of compatibility? (I do remember being annoyed when webp started to become prominent but yet wasn't supported by OSes, so opening .webp images was way more painful than .jpg/.gif/.svg. But that was just a knee-jerk reaction — it had nothing to do with the format itself.)
@filiph @jaffathecake yes
@jaffathecake normal users should have never had to care to about image formats at all. That was a mistake with webp that somehow was never taken seriously.
@evilpie @jaffathecake WebP solved the "use PNGs for graphics and transparencies, jpeg for photos and GIF for small animations"-problem.
The only problem with WebP was not doing enough outreach to Adobe to convince them to support it in Photoshop
@jaffathecake I guess this is just good old Google vs. Mozilla tribalism?
@data0 both AVIF and JPEG XL are primarily developed by Google teams.
@jaffathecake I believe major social media platforms are a key factor holding back AVIF adoption.
@jaffathecake If browser vendors began phasing out legacy image formats, it would accelerate the adoption of modern alternatives.
@individual8 this will never happen. Compatibility is way more important.
@jaffathecake If browser vendors aren’t the ones driving the impulse toward modern image formats, who actually takes the lead? Without clear direction and coordination, adoption just fragments in random directions for arbitrary reasons.
@individual8 @jaffathecake You see sites like YouTube fairly aggressively adopting new video codecs. The marginal bandwidth savings per-stream add up when you look at the whole user base.
The same kind of economics could make sense for a service that primarily serves images.
As a downside, web browsers will be adding ~ 10 bytes to the Accept header of image HTTP requests, no matter whether the website can serve the format. Hopefully that turns out to be worth it.
@jamesh @jaffathecake My guess is it comes down to compute cost vs. bandwidth. The platforms definitely have an interest in better formats, but AVIF is notoriously CPU-heavy to encode.
Saving a few dozen kilobytes on an image that might only get viewed a couple hundred times probably doesn't offset the extra server compute and ingest latency. Someone – especially at Meta – likely ran the numbers and figured that, at least for now, the trade-off simply isn't worth it.
@individual8 @jamesh I think the encode-cost info out there is out of date (and came from JPEG XL folks in the first place)
@jaffathecake It's because "XL" in Roman numerals is bigger than the "VI" that's in "AVIF". Simple psychology.
@brucelawson @jaffathecake JPEG XXX for porn images
@jaffathecake what does JPEG XL "shipping" mean? will it be supported by my OS and every image editor after shipping? because that was the whole problem with webp as far as I can see.
@kolya I'm talking about https://hacks.mozilla.org/2026/08/intent-to-ship-jpeg-xl/
@jaffathecake right, that's a technical description of the format. and I applaud the efforts to develop a new image format. but my point is that releasing the format and even integrating it in chrome and firefox isn't the biggest task. webp suffered from user acceptance and so will other new formats if there isn't a broad cross industry alliance to turn them into a fully supported standard. does such an alliance exist now?
@kolya sure, but you asked what I was talking about, and I'm talking about the announcement that it's shipping in Firefox and Chrome. That's what the "yesss fuck WebP!" has been in response to.
If they don't like WebP due to things other than browser support, then those things apply to JPEG XL too.
@jaffathecake When you say "outperform", I'm assuming you're talking about the one metric you mentioned on https://hacks.mozilla.org/2026/08/intent-to-ship-jpeg-xl/: quality per bytes.
But there are other interesting aspects:
- Compute/storage savings by storing one format that can be served in another (JPEG to JPEG XL and vice versa)
- JPEG XL encoding faster than AVIF for on the fly conversion use cases (what CDN providers tend to do)
@anthony do you have modern numbers on the encoding speed? Things have changed a lot there in my experience. Storage is cheap.
@jaffathecake I don't have anything besides https://cloudinary.com/blog/jpeg-xl-and-the-pareto-front#what_about_lossy_in_strong_jpeg_xl_compression_strong_ (and that's biased towards JPEG XL given the author).
I don't know if storage is still cheap given the current industry trend. Regardless, the compute costs are still a factor.
@anthony yeah, all the pro-JXL stuff seems to come from biased sources. My advice: Don't parrot the findings without verifying them yourself.
Some of the data is simply out of date too.
@anthony one of the JPEG XL leads also claimed the post-COVID layoffs wouldn't have happened if JPEG XL was adopted.
@jaffathecake Out of context, that reads like a joke to me.
@jaffathecake Wow… https://xcancel.com/jyzg/status/1592813739623219202#m is quite a claim as well…
@anthony discourse around JPEG XL has a habit of getting very weird very quickly. I was recently called a racist for enquiring about JPEG XL use-cases https://github.com/web-platform-dx/developer-signals/issues/215#issuecomment-4930663101
@jaffathecake @anthony yikes.
@jaffathecake @anthony The only sane way to solve this is to get JXL Hater Jake and these JXL Lovers to duke it out in cage fight.
@brucelawson @jaffathecake @anthony I am all in favour of this approach. A fight is the way forward. Enough talk! We need... a little less conversation, a little more action
@brucelawson @jaffathecake @anthony i would pay good money to watch this video for real, incidentally
@brucelawson @jaffathecake @ppk Hmm, I remember the last time you suggested something like that: https://vimeo.com/150121414
@jvhellemond @brucelawson @jaffathecake What was this about? I vaguely remember the occasion once you pointed it out, but I can't for the life of me remember what we argued about.
@jaffathecake @jvhellemond @brucelawson Ah, right, NOW I remember!
@ppk @brucelawson @jaffathecake I only remember sourcing the boxing kit and it mysteriously disappearing after the event...
@jvhellemond @ppk @brucelawson nuthin' to do with me!
@jaffathecake I mean it's mostly about adoption and compatibility, I believe and because of macOS JpegXL already has a lead over Avif in terms being supported in programs like Affinity of Photoshop.
Ease of adoption seems to be symmetrical: it's very easy to support Avif for a web browser, since it has to ship a decoder for video anyways, yet most publishing tools don't already do that and seem more comfortable including JpegXL instead
@jaffathecake As far as I know the main reason people disliked WebP is because it made it harder for them to download and use images because their OS/tools didn't have support yet. Do people think that JPEG XL is magically gonna skip that part because it has JPEG in the name?
@tojiro Some people definitely think that but mostly I think the sort of person that says they don't like WebP doesn't really know what it is and doesn't know what AVIF or JPEG XL is either.
@jaffathecake I will never get tired of the endless Holy Wars on Image Codecs. pure entertainment