Open DonaldTsang opened 4 years ago
https://github.com/google/brunsli claims that Python is already supported.
It does look that way! https://github.com/google/brunsli/blob/master/contrib/py/jxl_library_patches/jxl_pillow.py
JPEG XL format was recently frozen
@varnav so it has been standarized and passed through, correct?
ISO/IEC DIS 18181-1 is not completely done with being a standard. But this means format will not change anymore. I think it's a good time to start implementing it.
Maybe this code could be used: https://github.com/cgohlke/imagecodecs/blob/master/imagecodecs/_jpegxl.pyx
@olokelo Great news!
Hello! Is there anything missing to move this from the icebox to the backlog?
@brunoais Looks like we can now tentatively declare support via @olokelo 's plugin? (Thank you @olokelo !)
the plugin had no release since over a year. There are multiple issues from people that are unable to install it
I would appreciate built-in support
JXL is probably dead: https://bugs.chromium.org/p/chromium/issues/detail?id=1178058#c84
If everyone supports it Googles arguments get even more invalid, and maybe they will re-add support. We can't let Google dictate what becomes standard. There's already lots of software with JpegXL support.
JXL is probably dead: https://bugs.chromium.org/p/chromium/issues/detail?id=1178058#c84
You say that but most images on my PC are already .jxl
and just not all of them because some python software I use is waiting for Pillow to support JXL so they also support it.
JXL is probably dead: https://bugs.chromium.org/p/chromium/issues/detail?id=1178058#c84
https://en.wikipedia.org/wiki/JPEG_XL#Official_support says otherwise. libjxl and KImageFormats help a lot to grant support to basically any modern GTK and Qt application displaying images.
@radarhere @aclark4life is there anything really holding back built-in support? Or rephrased: What would you ideally want so that this issue can be resolved with a positive outcome for JXL?
The same thing that is holding back every new feature - time. While Tidelift is generous, Pillow is worked on in our spare time, and complex functionality is often sidelined while dealing with the deluge of other issues and PRs that are created.
Ideally, a PR with clear reference to documentation, and dependencies that fit within our license.
Small update: iOS 17 has JPEG XL support. Works in Safari and Photos app and reportedly (some?) third party apps. I hope this serves as a lesson for everybody to just keep on promoting the sane standards until they get adoption and to not give up just because a major player doesn't want to support something (yet). Especially if their proposed alternatives are just worse in most areas.
By the way, imlib2 is another library used by a lot of programs that has read and write JXL support (got the author to add the latter by opening an issue).
If I weren't already busy with a bunch of other projects I would gladly try to help on the software side of widely used imaging libraries to make adoption happen quicker. However, maybe this comment can help at making an enthusiastic developer work on JPEG XL support for pillow.
iOS 17 has JPEG XL support
macOS 14 as well.
Is there any progress on this feature? I noticed this library popped up in pypi
@Isotr0py is here on GitHub as well as it seems, as well as their project: https://github.com/Isotr0py/pillow-jpegxl-plugin However, it looks like this uses Rust bindings and it doesn't look like the Pillow repo has Rust code as of now.
Maybe they still can do this
PR with clear reference to documentation, and dependencies that fit within [Pillow's] license.
Hello again, for anyone interested please observe PR #7848 that enables read only support of JPEG XL images.
See https://jpeg.org/jpegxl/ since JPEG-XL is coming out in a few months Also https://google.github.io/brunsli has a sample app Co-referencing https://ai.google/research/pubs/pub48554 and https://www.iso.org/standard/77977.html