I mostly just want one of the formats to become the new defacto standard.
I wouldn't be too put out if either "won". If for example Apple did their usual weird thing and vowed to only supported AVIF then that would be an unfortunate reality I could live with.
Having said that, the data put forward by the AVIF team to support their removal of the rival is somewhere between partisan and fraudulent. The current submission is very measured in its criticism.
Most worryingly, from a non partisan standpoint, poor benchmarks lock us all into inferior performance.
If for whatever reason AVIF becomes the standard, the team's crappy report immediately becomes a barrier to actually improving the performance of AVIF, because any honest attempt to improve it will show JPEG-XL in a better light and therefore cause embarrassment.
Since we already went through this with WebP Vs MozJPEG it's sad to see a rerun.
AVIF sort of makes sense in how easy it is to support with the support of AV1 in the first place, but it is one of a long line of still-image codecs based on video codecs (coming to mind immediately are WebP and BPG), and the domain of video is significantly different from that of still images. Individual frames don't need to be as optimal for viewing as the aggregate sequence, fine details often get lost. Progressive decoding basically doesn't exist. Presumably the upper pixel limit is drastically limited for what could be in a still image format.
JXL makes an entirely different set of tradeoffs that really should pit it as the winner for a better still image format. Better quality, wider colorspace support, smaller files, progressive decoding, et al.
Just look at some of the sample pages. To me, it's immediately obvious AVIF's downfalls in fine details versus JXL.
3 comments
[ 0.25 ms ] story [ 15.9 ms ] threadI wouldn't be too put out if either "won". If for example Apple did their usual weird thing and vowed to only supported AVIF then that would be an unfortunate reality I could live with.
Having said that, the data put forward by the AVIF team to support their removal of the rival is somewhere between partisan and fraudulent. The current submission is very measured in its criticism.
Most worryingly, from a non partisan standpoint, poor benchmarks lock us all into inferior performance.
If for whatever reason AVIF becomes the standard, the team's crappy report immediately becomes a barrier to actually improving the performance of AVIF, because any honest attempt to improve it will show JPEG-XL in a better light and therefore cause embarrassment.
Since we already went through this with WebP Vs MozJPEG it's sad to see a rerun.
AVIF sort of makes sense in how easy it is to support with the support of AV1 in the first place, but it is one of a long line of still-image codecs based on video codecs (coming to mind immediately are WebP and BPG), and the domain of video is significantly different from that of still images. Individual frames don't need to be as optimal for viewing as the aggregate sequence, fine details often get lost. Progressive decoding basically doesn't exist. Presumably the upper pixel limit is drastically limited for what could be in a still image format.
JXL makes an entirely different set of tradeoffs that really should pit it as the winner for a better still image format. Better quality, wider colorspace support, smaller files, progressive decoding, et al.
Just look at some of the sample pages. To me, it's immediately obvious AVIF's downfalls in fine details versus JXL.