This is all it takes to upset the WebP shills

shilling samefag

>an image stored as WebP is still noticeably worse quality when re-compressed a smaller amount of times, which happens with most images shared frequently across the Internet
OH NO NO NO
IT'S A USELESS METRIC

Attached: webpiss 2.png (1920x1080, 1.57M)

The last time i bothered to educate myself about jpeg, it saves level of compression into header of the file so even if you happen to use JPEG as work format for drawing, repeated saving the pic wont shit up its quality since all normal jpeg libraries dont apply the compression again unless there was change in the 8x8 pixel area.

It looks extremly sketchy to me.

do a near-equal-filesize comparison

i don't see what your point is anyway. do you think you're some sort of codec insurgent? really cool

shill harder google employee

do a comparison with similar filesizes

>samefag
FLIF (the only format that produces above garbage-tier results) is, sadly, not supported by anything. Until that changes, websites should stop recrompressing everything thrown at them. Even when FLIF sees more adoption, recompressing is bad for security reasons, as it can trigger de-/encoder vulns.

Attached: samefag.png (367x136, 7K)

What a retarded faggot you are holy shit
>lets do a biased as fuck unfair comparison!
>hur dur stupid shill

SEETHING

>recompressing is bad for security reasons, as it can trigger de-/encoder vulns.
I would come to the opposite conclusion. Let's suppose someone uploads a maliciously crafted image. If the server recompresses it, only the server needs to be patched against the latest vulnerabilities. If it doesn't recompress, EVERY client viewing the image needs to be patched.

Surely the act of recompressing would trigger the vuln on the server anyway

Yeah if you're using a decoder from 2012. This hasn't been a thing for years anymore.
Have your (You), bored faggot.

encoder* derp

Oh, it's a FLIF shill who uses outdated information for the sake of false flagging.
Nevermind, that's just sad.

Yes, if the server isn't patched. But this way ONLY the server needs to be patched, instead of everyone who might see it.

Lossless webp is good.

Take your autism meds.

>FLIF
>not GLORIOUS FARFELD
i shiggy diggy

this comparison makes no sense, at least when looked from web development perspective. most important metric here (besides quality) is file size, which means you would bump up webp quality until file size approaches jpeg, and you would get image that's smaller and better looking than the one you would get by using jpeg

so deep fried memes were just a lucky accident?

Friendly reminder to not listen to WebP haters AND shills.
These comparisons showed a valid point against WebP, but you have to note that
a) the libwebp version used is severally outdated by now. Newer libwebp versions hold details incredible well for low resolution images (almost as if devs tweaked the encoder to prevent comparisons like this). For high resolution images it fucks up color over time.
b) FLIF's lossy quality setting has no lower limit, so if people really want to save space/bandwidth they can still ruin the quality over several encoding cycles.

The same person resaving his JPEGs isn't the problem. Websites compressing uploaded images to shit and people doing quick edits of JPEGs and recompressing it as JPEG are.

Given -Q0 FLIF starts introducing heavy noise over time. WebP compressed to reach roughly the same size starts smoothing over details and spilling predominant colors into other parts of the image.
Not sure if I still have my comparison pics saved. I'll look once I'm back on my main PC.

>q 49
>quarter of size
Neck yourself.

>if only the gays have sex, then only the gays need to use condom, therefore only the gays should have sex.

this is how stupid you are

I dunno bro, most internet memes dont see much edits / reuploads above number 10. This benchmark is completely joke. At this point i see most people saving everything as png anyway. It isnt rare to see 1mb jpeg photos saved as 30mb png files tjese days.

500 generations are unrealistic, but for WebP the biggest quality degradation happens during the first few generations. Once you hit the 100th, you see pretty much no further quality loss (the file size also stays pretty constant). That's something Google improved compared to these old comparisons.
>At this point i see most people saving everything as png anyway.
I don't. Normalfags don't know the difference between JPEG and PNG or think JPEG is always smaller, most pictures here get posted as JPEG (even if the footage would've been more suited for PNG) with an absolute shitty quality and big websites like imgur forcefully compress big files down to under 4MB.