Hacker News Reader

Comment by bla3 | Hacker News Reader

I kind of doubt it. JPEG is good enough, and is pervasive. JPEG XL is better, but removing an incumbent file format requires more than just being better. We'll see.

View on HN

JPEG XL has a party trick where you can losslessly transcode JPEG files to JPEG XL for a ~20% size savings. I imagine that every CDN out there will enable transparent conversion of JPEG to JPEG XL and that will drive adoption just by itself.

Dropbox quietly replaces the insides of JPGs with a better compression system (lossless change) to save their storage costs, and then returns it to normal when the file is read. (It may be they do a JPEG XL transform, for all I can remember about the article which explained it).

I wouldn't be too surprised if every big CDN already does something like that internally as a competitive advantage, and if so then the incentives switch around and there's no benefit publically switching to JPEG XL - when everyone has it, there's no easy way to do better than everyone else.

Aren't the biggest costs from bandwidth, not storage? I'm sure they would love to reduce data transmission by 20% for browsers that support it.

CDNs would love to bill customers 20% less for bandwidth used? Customers will be able to move to storing JPEG XL files on CDNs and reduce their own storage and bandwidth costs, but I was responding to the claim that it would be in CDN's interest to hurry the transition, and I think it will be a loss for CDNs.

Unlike preceding formats that tried to replace JPEG, this one was designed with the replacement challenges in mind.

A great blog post on this by one of the creators of the format, Jon Sneyers:

https://cloudinary.com/blog/time_for_next_gen_codecs_to_deth...

JPEG XL supports a recompression and bit-perfect reconstruction of JPEG(1) files, so once it got universal support it does have the biggest potential to completely displace JPEG.

Reply on news.ycombinator.com