← Home ← Back to /g/

Thread 105701134

38 posts 8 images /g/
Anonymous No.105701134 [Report] >>105701943 >>105701951 >>105705166
https://www.phoronix.com/news/PNG-Specification-2025

PNG Spec Updated For HDR Images & Animated PNGs

PNG BREHS WE ARE SO BACK
Anonymous No.105701164 [Report] >>105702159 >>105702189
png is good now?
Anonymous No.105701943 [Report]
>>105701134 (OP)
Why does the WC3 focus so much on bloating everything it touches?
Anonymous No.105701951 [Report] >>105703003
>>105701134 (OP)
gif already did that
Anonymous No.105702027 [Report]
Worthless unless the spec says JUST USE JPEG-XL INSTEAD
Anonymous No.105702159 [Report] >>105702479 >>105702615 >>105703262 >>105703280 >>105703295 >>105703446
>>105701164
No
Anonymous No.105702189 [Report]
>>105701164
always has been (the best)
always will be (the best)
Anonymous No.105702479 [Report]
>>105702159
>cloudinary
>>>/trash/
Anonymous No.105702615 [Report] >>105702793 >>105704482
>>105702159
>lossless
>1 point for PNG
>2 points for WebP/HEIC/AVIF
>3 points for JPG2000
>4 points for JPGXL

This does not add up. A format is either lossless or not. I mean I guess you can get a bonus point for being -optionally- lossless, but what is with this arbitrary distinction over other formats?
Anonymous No.105702793 [Report] >>105702991
>>105702615
>A format is either lossless or not.
No.
Anonymous No.105702991 [Report] >>105703221 >>105703234
>>105702793
explain then.
Anonymous No.105703003 [Report] >>105703477
>>105701951
>animated gif in 2025
kys
Anonymous No.105703221 [Report]
>>105702991
...efficiency is a factor retard
png is the least space-efficient lossless format
but to be fair it should be 4 points to webp 2 points to jpgxl/heic/avif, 1 point to png, dunno about jpeg2000 but no way it's better than lossless webp
Anonymous No.105703234 [Report] >>105703248
>>105702991
A format is just a container for data, data can be encoded in different ways even for the same format, JPGXL for example can de lossy or lossless (I think even JPGXL lossless is kinda lossy) same for WEBP.
Anonymous No.105703248 [Report]
>>105703234
even png can be lossy if the bit depth is lower than the source
Anonymous No.105703262 [Report] >>105703292
>>105702159
Why is lossless jpeg compression a testing point?
Anonymous No.105703280 [Report]
>>105702159
FireWire > USB
DisplayPort > HDMI
JPEG XL > WebP

Guess what will come out on top.
Anonymous No.105703292 [Report] >>105703307 >>105703483
>>105703262
because jpeg xl can losslessly re-compress legacy jpegs further (20% space savings on average).
Anonymous No.105703295 [Report]
>>105702159
Forgot to add an "actually used outside of faggot ass web browsers" metric.
Anonymous No.105703307 [Report] >>105703380
>>105703292
7z can do so too doesn't mean it should be a point tested on all image formats
Anonymous No.105703380 [Report] >>105703396
>>105703307
just tried it on a bunch of jpegs previously optimized with mozjpg-tran
it can barely save like 2% most of the time, plus it requires decompressing the archive first and of course you cannot embed such an image anywhere
I don't get what your point is
>doesn't mean it should be a point tested on all image formats
exactly, IMAGE FORMATS, 7zip isn't an image format you fucking retard
>a very important and useful feature which could save a ton of bandwidth without further quality loss on the internet shouldn't be a point on such comparison
kys jew.
Anonymous No.105703396 [Report] >>105703423 >>105703432
>>105703380
hmm im gonna try again:
why is a format-specific feature tested on all other formats?
Anonymous No.105703423 [Report] >>105703447
>>105703396
Are you retarded? All formats can take JPEG, the de facto standard for lossy compression today, and try to compress it further.
Anonymous No.105703432 [Report]
>>105703396
>why is a format-specific feature tested on all other formats?
I literally answered this question already
because it's a huge fucking deal
Anonymous No.105703446 [Report] >>105703875
>>105702159
>cloudinary
I'm sure they're a neutral and unbiased source, benchmarking all those formats on a 0 to 5 scale rather than giving the actual numbers.
Anonymous No.105703447 [Report] >>105703467
>>105703423
we are talking about lossless re-compression here
>Are you retarded?
he is retarded yes but you clearly are an even bigger retard
Anonymous No.105703467 [Report] >>105703499 >>105703510
>>105703447
Here's a typical story of an image:
> It's 2005
> You take a photo with your digital camera
> The camera takes the raw data and converts it lossily to a jpeg
> It's 2025
> You want to reduce the size of the image you have on your disk
> You wonder which format offers the best compression without additional loss of information
Anonymous No.105703477 [Report] >>105703838
>>105703003
GIF is lossless, all it needs is an update to allow 3-channels.
GIFs "look bad" becuse it has a 256 palette, that's lossless greyscale or limited colors. Dithering is how they cope around and emulate full color, with artifacts, but it's actually lossless. People just use it wrong.

Animated GIF for pixel art is THE BEST format there is.
Anonymous No.105703483 [Report] >>105703519
>>105703292
>jpeg xl can losslessly re-compress legacy jpegs further (20% space savings on average).

so can jpegtran and pngcrush.
Anonymous No.105703499 [Report]
>>105703467
Or you can just buy a new drive for the first time ever since 2005?
Anonymous No.105703510 [Report] >>105703521
>>105703467
>You wonder which format offers the best compression without additional loss of information
...and jpeg-xl is literally the only format that can achieve that
??? what the fuck is your point? do you have severe brain damage
Anonymous No.105703519 [Report]
>>105703483
20% space savings on top of jpegtran obviously.
Anonymous No.105703521 [Report]
>>105703510
>...and jpeg-xl is literally the only format that can achieve that
What? No. You can convert your lossy JPEG into any lossless format and hope that its size goes down.
Anonymous No.105703838 [Report]
>>105703477
it's actually really good for screenshots like op
Anonymous No.105703875 [Report] >>105704165
>>105703446
Everything on the cloudinary chart anon posted checks out at the date it was published. The only change was AVIF getting progressive decoding that although it barely does anything should be flagged as supported with one or two dots. AVIF also got better at high fidelity with the SSIMULACRA2 tune now merged into the mainline encoder so I'd give it an extra half a dot, not full because it's still kind of experimental behind a flag.
Anonymous No.105704165 [Report]
>>105703875
Also 12-bit professional profile AVIF was missing at the time.
Anonymous No.105704482 [Report]
>>105702615
That field is labelled "compression," anon. The ratings are for how small it can compress an image losslessly.
Anonymous No.105705166 [Report]
>>105701134 (OP)
>jpg
>more pixels
>ligh than png
How?