Hacker News Reader

Comment by benchly | Hacker News Reader

I thought that too assumed my blog fit the criteria. I was wrong; weighed in at just over 100KB too heavy to get in the club.

My guess is the photos.

View on HN

Exactly, so it's not so much a demonstration of how nice a website fits in 512K as it's about _just not using any media_. Not very interesting imho.

Same here. You can join me in the https://1mb.club/

Hah! Maybe for now, but I'll probably pass that with a few more posts I have in the works. They're related to some projects I have going (that I will eventually finish), which means more photos.

What I really need to do is standardize my format and size, just to make things a bit more visually tidy. I'm not all that interested in keeping below a certain size, more so that I'm not doing anything I consider unnecessary for the function and feel of the site. Fairly certain whatever readership I have is either related to me or people I play D&D with, so it's really just a fun thing for me to do once in awhile.

I ran a pretty lightweight blog, but sometimes several .jpg would easily cost >512KB

maybe lower the resolution?

at least to test your guess

are you using webp photos?

webp is not much of an upgrade in my experience - jpegli pretty much matches it in quality/size while having better compatibility, and if you don't have the original photo and are working with old crusty jpegs it's often best to just leave them alone rather than re-encoding. jpeg-xl does make a noticeable difference, but it's not widely supported.

Webp is utter garbage. Awful quality loss, which they're still in denial of, for minimal size gains over a jpeg encoded with modern software (which is also way way more compatible - software from 35 years ago can open it). But I'm sure it scores well in whatever flawed perceptual benchmark they automated.

If only google didn't oppose jxl - but they'd have to implicitly admit that webp is garbage and they don't like doing that.

WebP produced good file size reductions because they recompressed JPEGs and ignored the loss in detail. I benchmarked it the day it launched and it was never once competitive enough to be worth the cost of using it. If browsers had supported JPEG-2000 in the 2000s, it’d have stomped WebP on every benchmark – even a tuned JPEG encoder did surprisingly well given the age of that format.

HEIC, AVIF, JXL, etc. are worth the trouble.

WebP isn't enough of an improvement to justify the overhead of supporting it for me. AVIF is twice as efficient as JPEG, and has 97+% support across browsers. Use <picture> for JPEG fallback.

https://caniuse.com/?search=avif

https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...

Reply on news.ycombinator.com