[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / 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]


[Advertise on 4chan]


File: pepes.jpg (140 KB, 976x425)
140 KB
140 KB JPG
2.4KiB avif vs. 2.8KiB jpg

Why the difference in quality?
>>
>>106482961

people been reading jpeg spec for 33 years they have come up with bandwidth savings
>>
>>106482961
thats a 140kb jpg tho
>>
>>106483140
I uploaded it as a png but 4chan turned it into a jpg for some reason. Anyway it looks like the original avif and jpg so it illustrates the point fine.
>>
>>106483189
you didnt show any avif tho, thats a jpg
>>
>>106482961
Because unlike lossy webp, lossy avif doesn't have that retarded problem where it gives you a file bigger than the jpeg. Ironically lossless avif gives you a bigger filesize than lossless webp.

Jpeg xl is trying to fix this but there some images will still result in smaller file sizes via AVIF so it looks like we're practically speaking, stuck with both.
>>
only people that care about this are mega corpos and autists, if i got paid 200k a year to look into this i would but im not so i dont care
>>
>>106482961
right one looks more sovlful
>>
>>106482961
post jpeg xl pepe
>>
File: pepe.jpg (295 KB, 1464x425)
295 KB
295 KB JPG
>>106483529
jxl on the right, but even on the lowest quality setting in magick it would not go below 5KiB
>>
>>106483559
You are either lying or you produce lossy slop.
Its a cartoon frog with 4 colors, use lossless like every normal human person.
>>
>>106483189
you uploaded a jpg and the picture looks like a jpg
>>
>>106483559
i only see one jpg here
>>
>>106483559
that's just one image
>>
png or death
>>
>>106482961
>look guys, i took this image that is originally a lossless png, i encoded it in two lossy formats, put them next to each other, and encoded it again in a lossy format
>now let me explain why we need to use lossy google slop and reencode all our images and lose quality
>>
>>106483390
>[webp] gives you a file bigger than the jpeg
In 2.7% of cases...
>>
A jpeg reencoded into AVIF will always have worse quality than the original jpeg.
A lossy AVIF will always, in all cases, have worse quality than a PNG.

Therefor i don't see a usecase in avif.
At least webp got some acceptable lossless performance going for it.
>>
>>106483786
2.7% too many
>>
>>106482961
>Pepe fed edition
fuck off
>>
>>106483846
The point is that WebP is net positive, especially when you consider that images can be higher quality for less data use.
>>
>>106483872
>images can be higher quality
see >>106483840
>>
File: 1755380516252714.jpg (86 KB, 1179x951)
86 KB
86 KB JPG
>>106483786
>>106483846
>>106483872
When you attempt archival quality lossy it's more like 50%. That's because webp is limited to 4:2:0 chroma sub sampling which sacrifices color resolution in an image for better compression efficiency at low file sizes.

Essentially Webp ONLY works okay when the existing JPEG itself is 4:2:0 and not 4:4:4. Also it needs to be significantly huge in file size to not justify a lossless jpeg xl compression.

Lossy webp is fucking cursed.
>>
>>106483910
>archival quality lossy
that would be a lossy compressed DNG, which allows jpeg-xl, but doesn't allow avif, as it is recommended by the library of congress

If your mean your shitty meme folder on your external drive, just keep it lossless. Like everybody does.
>>
File: baseline.png (24 KB, 1200x742)
24 KB
24 KB PNG
>>106483952
No, many internet jpeg 4:2:0 images are encoded via a GPU instead of a CPU. Similar to video, an image GPU encoder like Nvjpeg will either produce unreasonably huge file sizes or piss poor quality. If it's the former then a webp encode can produce an archival quality via lossy compression and reduce the file size by 20-30% because libwebp is exclusively a CPU encoder.

The image codec situation on the web is an absolute fucking mess desu.

https://developer.nvidia.com/nvjpeg
>>
>>106484035
I am not reading your cope.
Lossy to lossy encoding is idiotic.
And lossy is always worse quality than lossless. That is the whole point of it.

And archival is about storing the highest quality. You archive music in flac so you archive images as lossless or even RAW if you get that.
Lossy-to-lossy reencoding for archival is an oxymoron.

The only usecase of lossy-to-lossy is fucking thumbnails that you cache and can delete at any point.
But WHO GIVES A FUCK ABOUT THUMBNAILS.
Are you digging through charts just to find out what format to use for temporary thumbnails?
>>
>>106484035
>The image codec situation on the web is an absolute fucking mess desu.
Most images on the web are lossless PNGs.
So the situation is much better than expected! Turns out that most people actually value quality.
>>
>>106482961
Using the fed pepe with the line the CSNBC article added, is ironic.

The only reason why this pepe spread is because it was the first result on the Google image search for pepe.
Back in the day, people actually used google image search and copied images from there to repost.
It all changed when google decided to serve webps. Even the most normiest normies stopped doing it, because they realized that google images saved funny and werent accepted anywhere.

And here you are, shilling for webp3
>>
>>106484103
????

If you can't wrap your head around the concept of lossy archival quality then I can't help you desu.

What we can all agree on is that when that a GPU encoded 4:2:0 JPG with dogshit quality gets converted to a webp you end up with a webp that is BIGGER in file size compared to the source JPG especially at high quality which is why Jpeg XL applying lossless transformations to the huffman tables or whatever resulting in smaller file sizes is pretty fucking cool and whatnot.

That's why I say webp is cursed because even IF you come across a correct 4:2:0 JPEG image you don't know if the file size is needlessly huge or not for webp to achieve that 20-30% filesize reduction with no noticeable quality loss. Websites DON'T FUCKING CARE if the net positive means less bandwidth used.

>>106484116
For mom and pop websites, sure. Not the globo-homo one serving millions of user uploads per day. Do you think they would be able to make any money if they had a backlog of images to CPU encode because they decided to focus on better compression efficiency? You think they're your friend?

>>106484195
50% lower filesize compared to both 4:2:0 and 4:4:4 JPGs encoded by a GPU is nice though it still assumes that the file size of the JPEG is huge. So call it what you will but by volume AVIF would still save us a lot of bandwidth but knowing globo-homo they'll just use GPU AVIF encoders because profit line must go up lol. But yeah, hopefully only CPU encoder for JXL exist like webp and the fucking imbeciles don't touch the default settings.



[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.