guetzli VS libjpeg-turbo

JPEGLI = A small JPEG

The suffix -li is used in Swiss German dialects. It forms a diminutive of the root word, by adding -li to the end of the root word to convey the smallness of the object and to convey a sense of intimacy or endearment.

This obviously comes out of Google Zürich.

Other notable Google projects using Swiss German:

https://github.com/google/gipfeli high-speed compression

Gipfeli = Croissant

https://github.com/google/guetzli perceptual JPEG encoder

Guetzli = Cookie

https://github.com/weggli-rs/weggli semantic search tool

Weggli = Bread roll

https://github.com/google/brotli lossless compression

Brötli = Small bread

A JPEG encoder developed by Jyrki Alakujala, Robert Obryk, and Zoltán Szabadka, and released by Google in 2017. Guetzli specializes in high-end image quality where it is claimed to produce significantly smaller files than prior encoders at equivalent quality, albeit at very low speed. It is named after the Swiss German expression for biscuits, in line with the names of other compression technology from Google. github.com/google/guetzli

I'm a huge fan of AV1 for video, but for images JPEG-XL is simply the better codec than AVIF. If you've not actually looked closely at a comparison and are just on the side of AVIF in this debate because it's based on AV1 (and maybe you hate HEVC / HEIC), I'd urge you to look closer. Jpeg XL is pretty unrelated to Jpeg, Jpeg 2000 and Jpeg XR and instead a successor of Google Guetzli, FLIF and newer research.

I've never had much luck using jpegoptim. In most cases it's only removing the metadata, which isn't much on high-res files.

Guetzli is nice, if you don't have too many images to recompress (quite slow): https://github.com/google/guetzli

If you're going for full re-encoding, it might help to decode the current JPEG with https://github.com/google/knusperli ... but if you re-JPEG that you might have second-order artifacts. Give it a try. Then compress with https://github.com/google/guetzli

You know. I was actually quite annoyed ( to say the least ) with the post. For one the post is lacking a date, and you have to search yourself it was published in April 2017. And without a date the article is completely lacking context because Guetzli [1] hasn't been worked on for 5 years. And as [2] mentioned its work and derivative was ultimately merged into JPEG XL, which is a very decent image format. ( People should definitely check out JPEG XL if it is not on your radar yet )

But then I notice it was Dan luu who submitted it, which likely means there must be something much deeper than is what is shown on the surface. So what is the context here ?

[1] https://github.com/google/guetzli

[2] https://news.ycombinator.com/item?id=30622303

> all decoders will render the same pixels

Not true. Even just within libjpeg, there are three different IDCT implementations (jidctflt.c, jidctfst.c, jidctint.c) and they produce different pixels (it's a classic speed vs quality trade-off). It's spec-compliant to choose any of those.

A few years ago, in libjpeg-turbo, they changed the smoothing kernel used for decoding (incomplete) progressive JPEGs, from a 3x3 window to 5x5. This meant the decoder produced different pixels, but again, that's still valid:

https://github.com/libjpeg-turbo/libjpeg-turbo/commit/6d91e9...

Last vestiges of this fact AFAIK were libjpeg, which had a macro NEED_SHORT_EXTERNAL_NAMES that shortens all public identifiers to have unique 6-letter-long prefixes. Libjpeg-turbo nowadays has removed them though [1].

[1] https://github.com/libjpeg-turbo/libjpeg-turbo/commit/52ded8...

While I think the move to safer code through Rust and other alternatives is a nice breath of fresh air, I doubt you can get these kinds of optimization without using unsafe code in Rust. These optimized implementations often require some kind of safety-bypassing memory modifications to work as efficiently ad they do.

There's a reason https://github.com/libjpeg-turbo/libjpeg-turbo/tree/main/sim... is filled with assembly files with conditional loading.

Sure. You'll see it very often in codec implementations. From rav1e, a fast AV1 encoder mostly written in Rust: https://github.com/xiph/rav1e/tree/master/src/x86

Large portions of the algorithm have been translated into assembly for ARM and x86. Shaving even a couple percent off something like motion compensation search will add up to meaningful gains.

Or the current reference implementation of JPEG: https://github.com/libjpeg-turbo/libjpeg-turbo/tree/main/sim...

zune-jpeg is 1.5x to 2x faster than jpeg-decoder and is on par with libjpeg-turbo.

This is interesting from the standpoint of how new JVM features can be used to improve performance (what I presume the article's main purpose to have been), but the image processing improvement itself isn't head-turning. Also, we've found that libjpeg-turbo (https://libjpeg-turbo.org/) is ~5x (IIRC, can re-run my JMH benchmark if anyone wants me to) as fast for decoding JPEGs as ImageIO, so we wouldn't even benefit from this change in 21 much.

libjpeg-turbo: https://github.com/libjpeg-turbo/libjpeg-turbo/blob/main/simd/x86_64/jsimdcpu.asm

libjpeg-turbo

For all but the simplest formats (basically BMP), the difficulty of implementing encoding/decoding from scratch is significant - well beyond a beginner's ability, and challenging/time-consuming even for senior developers. So, libraries are used in practice - e.g. libpng and libjpeg-turbo.