Who is going to re-encode their images en masse? Content companies? Photo sites?
Facebook. Which is probably a large percentage of the images going over the web. Heck, they could start by just doing it on mobile.
Nobody will reencode anything in photo sites. Just newer content will be supported.
As for FB, their internal re-compression of jpegs is annoying as hell. Fine details taken with D750 become blocky almost cell phone quality images. Solution? Upload PNG.
It doesn’t need to be done by anyone en masse. It can easily be phased in gradually.
JPEG-2000 is an actual standard. I forget how supported it is. That said, not many images for the web showing up so encoded.
I don't think a format requiring JS to decode and negating the whole parallel download/decode thing has a chance regardless of the elegance.
It's not an imaging format discussion I this case.
JPEG 2000 is an actual standard but to get it working you need to license a decoder from someone or build a new one, you can’t just piggyback on some existing HEVC decoder. JPEG 2000 doesn’t currently work in every browser, and there have been at least a couple of security vulnerabilities in the decoder (Kakadu I think?) that OS X ships by default and enables for Safari. In many cases where there’s no native browser or direct plugin support you can get it working with a slow, buggy, resource-hogging Flash implementation, but I’ve had pretty poor experience as a user on sites that incorporate one. I don’t believe there are any pure-Javascript implementations.
As far as I can tell, this new format already has a better client compatibility story than JPEG 2000 right off the bat, as well as better quality for the same file sizes.
Prior HN thread - https://news.ycombinator.com/item?id=8704629
I can imagine Safari and IE building support for this format entirely as a counter to Google’s WebP, since they already need support for HEVC as a matter of course.
Many current websites send very heavily compressed JPEG images to save bandwidth, and just live with the big hit to image quality. If they had a similar-filesize alternative that they could use for some or all clients that didn’t degrade image quality, and still was guaranteed to work smoothly, it’s plausible they might adopt it.