[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • You may highlight syntax and preserve whitespace by using [code] tags.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


Janitor acceptance emails will be sent out over the coming weeks. Make sure to check your spam folder!


[Advertise on 4chan]


File: fuck_that_took_forever.png (135 KB, 1361x823)
135 KB PNG
Webpgoy from umi here: What does /g/ think of lossless Webp? To clarify I'm not the site owner, I'm just another image codec autist.

>Relative to lossless JXL filesize, how close in compression does a given lossless image codec get?
That's what I'm trying to answer here. A dozen common type 1920x1080 PNG images were used.
4chan still doesn't support lossless Webp but umi does, so I'll post the oxiPNG files here and lossless Webp ones in umi.

https://umigalaxy.com/explore/general/785-lossless-webp

JXL KB = 100% and BMP KB = 0%. Lossless JXL/Webp/AVIF/oxiPNG/QOI tested. Anything above 100% means JXL was defeated.

cjxl in.png out.jxl -d 0.0 -e 10 --brotli_effort=11
cwebp -lossless -m 6 -q 100 in.png -o out.webp
avifenc -l -s 0 --cicp 1/13/16 in.png out.avif
oxipng -o 6 in.png
qoi "...\in.png" "...\out.qoi"
>>
I dont like (((webp))), to easy to embed glownigger things inside.
>>
>>109108306
How so? Hidden images are found in PNG images all the time. I think you just need to be careful what you download so you don't end up unknowingly hosting cp.
>>
File: 1.png (3.09 MB, 1920x1080)
3.09 MB PNG
Picrel oxiPNG_: 3166 KB
Lossless Webp: 2305 KB
~27% smaller
>>
webp has too many security vulnerabilities.
And its not even good, it is already getting deprecated.
Why switch to something that is already scheduled to be removed?
>>
Just use jpg.
>but muh filespac-
Then tell youtube to make every video 240p
>>
File: gs3.png (1.59 MB, 1080x864)
1.59 MB PNG
>>109108369
I was actually going to address all of that on my second post but I forgot. I got a headache doing all of this testing desu. I can't write scripts so each of these codec formula results had to be manually input into windows calculator. Every fucking 50 or so.

My case for lossless Webp:
A) There are diminishing returns beyond lossless Webp. JXL lossless isn't worth the squeeze IMHO.
B) Lower RAM/CPU usage for both encoding and decoding means even poorfags can use it. Even on their phones.
C) Supported on nearly every Web browser, image viewer/editor on both desktop/mobile out there.
D) Nearly 3 years have gone by since the worst libwebp RCE security exploit was discovered and patched.
E) Lossless JXL/AVIF might have undiscovered 0-days. Adopting Lossless Webp now is relatively much safer.
>>
>>109108399
Some of these images are bigger as a JPG than a lossless Webp/AVIF/JXL/oxiPNG.
>>
File: 2.png (61 KB, 1920x1080)
61 KB PNG
Picrel oxiPNG_: 62 KB
Lossless Webp: 35 KB
~44% smaller
>>
>>109108331
Yeah im not talking about that, im talking about illegal tracking and tracing bs.
>>
>>109108439
But why would they even need to put that in the image? Most websites just embed it into links. But what I have seen is that certain PNG out there on social media harbor a second hidden image so my guess is that acts like a backup in case the user removes all the tracking stuff from the link.
>>
File: 3.png (3.07 MB, 1920x1080)
3.07 MB PNG
Picrel oxiPNG_: 3142 KB
Lossless Webp: 2338 KB
~26% smaller
>>
File: pixDAIZ is here.webm (94 KB, 360x270)
94 KB
94 KB WEBM
>>109108299
>another one
>>
>>109108561
Another what?
>>
>>109108299
stop trying to make webp happen. 0 fucking use case. stop it you autistic niggas, DAMN
>>
>>109108590
This is lossless webp. You might be confusing it with lossy webp.

On a related note this isn't just a problem for Webp. JXL and AVIF have the same problems. I've often thought how much better it would be if they used separate extensions for lossless and lossy but then again lossy PNG compression exists as well so maybe it wouldn't really clear up anything in the end.
>>
>>109108299
Why did you give up on AVIF?
Imagine spaming AVIF for years, just to then shill old and abandoned webp (based on VP8)?
>>
File: 4.png (307 KB, 1920x1080)
307 KB PNG
Picrel oxiPNG_: 307 KB
Lossless Webp: 195 KB
~37% smaller
>>
>>109108669
I've never shilled for AVIF, otherwise I would have never posted this thread showing how it's inferior to PNG sometimes. If you're curious about why lossless Webp (not lossy webp), my reasons are stated in >>109108410
>>
>>109108707
why are you lying?
you think everybody here is stupid, spamer?
>>
File: 5.png (491 KB, 1920x1080)
491 KB PNG
Picrel oxiPNG_: 492 KB
Lossless Webp: 372 KB
~24% smaller

>>109108713
Maybe you're confusing me with another codec autist here? There's a lot of us like that here, we're not all the same person. I've been posting as anonimouse for a while desu.
>>
>>109108739
Sure thing, retard.
webp is a thing for over a decade and hated by everyone and you just now suddenly discovered your love towards an abandoned Google project after yet another google project got wrecked...
>>
File: 6.png (58 KB, 1920x1080)
58 KB PNG
Picrel oxiPNG_: 58 KB
Lossless Webp: 34 KB
~41% smaller

>>109108747
Lossy Webp I get, that's why Iris-Webp exists. Most people literally don't even know that Lossless Webp exists though. Like be honest, did you even know that the latter existed at all before I made this thread?

https://halide.cx/iris/
>>
>>109108759
Everybody knows that lossless webp exists, idiot. It is still deprecated, since it is part of the vp8 video compression rip-out format.
If you want lossless, stick with png or wait for jpeg-xl.
Apparently you spent over a decade in complete ignorance, but wating a year or two is too much for you spaming retard?

Or you just want to bait some newfags again... like you always attempted to do.
>>
>>109108785
So you're not here to have any honest discussions about lossless image compression, are you? I mean my OP pic should make it clear why Lossless Webp shouldn't be brushed off so easily.

On the chance that I'm wrong, I really doubt most people know that Lossless Webp exists because the only exposure they've had to it is high traffic websites compressing a quality 10 JPG to a quality 10 lossy Webp. It's only been like a few years since it started showing up on GUIs, before that you had to be the top 1% of users who had the ability to use a CLI.
>>
>>109108299
I'd still prefer a png. It's compatible with far more websites and programs. And I can tell just looking at the file extension that it's lossless. Meanwhile I have to go looking for a program that can deal with a webp, much less tell me whether it's lossless or not.
>>
>>109108759
>did you even know that the latter existed at all
so you are saying that there are >>109108739 "lots of us different codec autists here", yet you think that this is new knowledge?

I added a lossless webp conversion to the contextmenu of png files almost 10 years ago. Because my vidya screenshot folders got large and i didn't want to cull them.
Nowadays i have a single "Convert to JpegXL" option, for all image types, that converts png to lossless jpeg-xl and losslessly transcodes jpg into jpeg-xl.

Now you show up in the year 2026, have never been here before (if we believe your words), with half-knowledge declaring yourself an knowing codec autist.... totally unrelated to a totally different person who also throws half-knowledge around of the exact same type and exact samestyle?
Kill yourself.
>>109108828
>but but you are dishonest
You literally admited that you are an unknowing retard
>Lossless Webp shouldn't be brushed off so easily.
it's brushed off because it is mandatorily paired in a format together with a fucking deprecated vp8 video frame compression!
That is a very good reason. Nobody wants to have this bloat (at best) or security vulnerability (at worst).
>>
>>109108299
Can you codecs autists make a general for this like "media codecs general" /mcg/ and rant about it all day.
>>
>>109108856
I've been here a long time man and if you use JXL for everything, good for you I guess. I just don't have the same positive outlook on that format because I consider other things like decode/encode performance and web browser support which seems kinda important for sharing stuff on the web and whatnot.

>>109108884
I doubt it would be constructive.
>>
>>109108905
jxl is part of the pdf standard.
webp is not.
Every single device that wants to be able to open PDFs will support jxl. It is inevitable and here to stay.

Meanwhile you totally legit year old oldfag discovered worse performing lossless webp 16 years after its introduction and decided that its a pressing issue?
>>
>>109108912
>jxl is part of the pdf standard
I don't think that's something worth bragging about desu. You don't think it's an issue that JXL has such poor overall web browser support?

https://blog.acer.com/en/discussion/2924/how-a-pdf-exploit-hack-took-down-4chan-in-2025
>>
>>109108905
>web browser support
Google bowed and promised to support jpeg-xl in Chrome the day the pdf association declared jpeg-xl as their way forward.

Again something you seemed to have missed? Despite being here for years and soooo interested in codecs, you missed that news?
>>
>>109108924
4chan got hacked thanks to postscript parsing, not pdf parsing.
Again that convenient half-knowledge...
>>
>>109108933
>April 14, 2025, the notoriously anonymous platform 4chan was hacked through a critical vulnerability involving PDF file uploads.

>involving PDF file uploads.

>Attackers exploited outdated and insecure software handling PDF files, allowing them to gain remote code execution and ultimately escalate privileges to obtain full administrative access to 4chan’s servers. This breach led to significant data exposure, including sensitive user and administrative information.

>software handling PDF files

https://blog.acer.com/en/discussion/2924/how-a-pdf-exploit-hack-took-down-4chan-in-2025
>>
>>109108949
a PostScript file got uploaded with a pdf extension, and it didn't get checked whether or not its even a pdf, it got fed into an 14 year old ghostscript version that executed the PostScript

Why do you have to link to a blog from acer.com? Weren't you here when it happened?
>>
So you be saying png -> lossless webp is smaller? With magick if I do lossless webp from png it becomes 4x the size. I will try with cwebp maybe but this difference makes no sense lol
>>
>>109108299
pnglet
>>
>>109108925
I'm still seeing 14% browser support.

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

Meanwhile Webp now has ~96% browser support.

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

So this JXL thing is going to be another "2 more weeks!" kind of deal, isn't it? If so why not use lossless webp until the "2 more weeks!" are over? Doesn't that seem like a good compromise to you?
>>
>>109108959
k well I'm uploading every PDF I ever come across to multiple online virus scanners from now on to know when it's time to do a OS re-install just in case.
>>
>>109108968
How dishonest.
Are you denying that google promised to add it?
>but but think about whats RIGHT NOW
webp is 16 years old, there is no need to rush. You waited those 16 years but waiting a single year more is not an option?
>>
>>109108961
Check if the PNG is lossy. It's hard to believe it but some PNG files on the web are not PRACTICALLY lossless. Otherwise lossless webp cuts down oxiPNG filesize by 20-30% on average as shown in my thread.

https://pngquant.org/
>>
>>109108988
>"2 more weeks!"
k, keep me posted when JXL reaches ~96% web browser support.
>>
>>109108974
No online virus scanner would have detected that malicious postscript.
The best prevention method is to not have bloat lying around.
Like GNOME had a 0-click RCE exploit because they automatically parsed old .cue files.

Which means:
>don't support obsolete PostScript files
>don't support obsolete webp
>don't support .cue sheets
>>
>>109108994
>>109108968
4chan doesn't support webp so i don't use it
>>
>>109108989
Oof it was actually a jpeg, not png.
https://unsplash.com/photos/aerial-photo-of-brown-moutains-JgOeRuGD_Y4
Does it get smaller with cwebp and those options if you do lossless? Doubt it accepts jpeg doe. And if you do jpeg -> png -> webp? Highly doubt this can be shrunk lowk
>>
File: 1723530290596701.png (76 KB, 997x822)
76 KB PNG
>>109108299
It's the only useful part of webp
>>
>>109108828
nigga none of your shit is honest at all, you arent sharing any real details just running your mouth like an LLM defending itself
>>
in cases where lossless jxl is very slightly bigger than lossless webp,
-q 100 -e 11 --allow_expert_options
 makes it smaller.

conclusion: lossless webp has no use.
>>
adding -g 3 or -g 0 also sometimes produces smaller jxl images without having to wait for maximum -e 11 effort.
>>
>>109108361
> 27% smaller and still looks like shit on my phone
>>
the fact that it's impossible to tell whether a webp is lossy or lossless without knowledge of the format and how to check makes it a retarded, awful option to pursue, and it should be abandoned immediately. At the bare minimum reskin/rename the lossless version, ensure it's lossless 100% of the time, and switch to that.
>>
>>109108420
> jpg bigger than lossless webp? that's a new level of cope.



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.