Hello /g/ you may have seen my posts compressing large images of OPs or other anons in threads to static Webms, that was me.
Anyway the point was to show how good VP8 compression is and how we missed out big time not adopting WebP. Sadly 4chan likes staying in the past and catering to icucks so Webp support will probably never arrive here.
I want to share with you all the scripts I have been using, just save them as .BAT, put them in the directory of images, and run them. You must have FFMpeg to run them ofc.
PNG:for %%f IN (*.png) do (
ffmpeg -loop 1 -i "%%~nf.png" -an -c:v libvpx -qmin 16 -qmax 16 -quality best -threads 4 -t 2 -r 1 "%%~nf.webm"
)
JPG:for %%f IN (*.jpg) do (
ffmpeg -loop 1 -i "%%~nf.jpg" -an -c:v libvpx -qmin 16 -qmax 16 -quality best -threads 4 -t 2 -r 1 "%%~nf.webm"
)
These scripts will generate a static 2 second Webm which will have reduced file sizes and nearly the same visual quality. VP8 is a lossy codec so you will lose some quality but it won't be much. You can change the -qmin and -qmax values to 4 if you want as close to lossless as possible (not recommended).
pic related is first static Webm example.
Now here is the JPG from which the static Webm was extracted from. This JPG was saved in 100% quality
>>54928193
And i was thinking why always one retard would repost the fucking pictures.
But actually thanks OP for showing me how shit 4chan is. That Webm loaded really fast even tough im in eu. It's really quick compared to jpg and png.
>>54928193
>using windows
Fuck off newfag
I believe that pics on 4chan shouldn't have to x-box huge, there's no need to upload 2-3MB JPG/PNG images when you can upload a static Webm with a file size of 1MB or less.
VP8 seems to compress images of real people/places the most with file size reduction ranging from 10-15X. It also does an amazing job of compressing PNGs of chinese cartoons/vectors as well with file size reduction ranging from 5-10X.
I will keep posting more Static Webms and the source pics from which they were extracted from thereafter. All JPGs were saved with a quality of 100% and PNGs compressed to level 9 in Gimp.
>>54928328
>Am I fitting in now?
>>54928346
Let's just skip to the conclusion of these threads.
>>54928193
>.BAT
How cucked are you?
>>54928328
I apologize but I do not know how to write the scripts in linux. I actually dual-boot ubuntu Linux and windows 7 but I am still new to Linux in general.
>>54928373
haha yeah these kinds of Webms are pretty bad. icucks who have less than 2GB of RAM on their pissphone may suffer crashes or severe lag playing these.
>>54928346
Looks like shit compared to
>>54928372
Please stop trying to ruin the images on the imageboard.
>>54928193
What's the point in a webm that doesn't move?
>>54928439
>>54928464
Reduced file size basically see >>54928346
>>54928423
I don't see any RAM hogging but I see high CPU usage. What does that webm do, 1000 fps or something?
>>54928519
>>54928346
>>54928372
Some of the leaf spots are missing
>>54928193
Why didn't you just tell us what you were doing as you were doing it? Was about to report you for spam, lucky my VPN blocked it.
>>54928519
>those conditions
lol what a gay
>>54928346
There's no clover support (at least not on 1.12) for zooming in on webm, that's an argument against.
And as smaller file sizes are especially useful for mobile users (at least in Europe) Floens got some work to do.
>>54928451
They look very similar actually. Yes VP8 is lossy, water is wet, what else is new? The point is despite the large resolutions VP8 will give you a small file size image that looks pretty much the same as a JPG 10-15X the file size. Yes, you will notice the artifacts if you zoom in at 100% on both images and closely inspect them otherwise you won't.
>>54928537
seems like it lol.
>>54928555
I was bored I guess.
>>54928591
>And as smaller file sizes are especially useful for mobile users (at least in Europe) Floens got some work to do.
Indeed. Maybe if enough people participate in creating and sharing static Webms floens might add support for zooming in and panning static Webms. That would be pretty nice. You can already do this in Opera.
>>54928537
>What does that webm do, 1000 fps or something?
65536fps
Here is an excellent example of how aggressive VP8 compression can get
This static Webm is ~23X smaller than the PNG
>>54928808
>>54928808
Lossy vs. lossless.
Apples and oranges.
>>54928833
>>54928545
>>54928484
You should consider optimizing your source files instead of converting them to videos.
Too bad FLIF beats webp for quality/size.
>>54928808
Obviously a lossless PNG will be bigger.
You should be comparing VP8 to a JPG at the same file size.
>>54928915
Okay family then try optimizing >>54928833
good fucking luck.
>>54928917
WebM transparency support WHEN
>>54928917
In case you all have not noticed, VP8 supports transparency as well.
>>54929060
>>54929061
Well then
>>54929076
There is a catch (for now). Mobile browsers including 4chan clients like Clover do not support displaying alpha channels on Webms. Hopefully this will change soon.
>>54929022
>I just had a thought: you can program pause-delays into Webms, right?
I'm not sure. I haven't tinkered enough with FFMpeg or VP8 parameters to know that.
>Would Webm be better for presentation slide-shows than picture format?
If the above was possible then I think so yeah. Better to show the slides on a video that pauses itself than to have to open up microshaft office praying it won't crash or glitch.
>>54928193
WebP had some troubles in support when I implemented it on the chan I ran. It was great for thumbnails when people ran Chrome (I had dynamic support in the nginx config to select format from the headers), but strangely would not render when people opened them directly - which is why I only used them for thumbnails.
Incidentally, FLIF seems to be quite seriously superior to WebP, beats JPEG2000 as well - it's right up there with BPG but without the dreadful H.265 patent problem that has.
Maybe we'll see in two decades if browsers ever want to implement it. Seriously, Mozilla would rather die than ship new image formats, ever. Google only did because of the YouTube team.
can anyone give me a good CPU burner webm?, preferable not anime picture, because of reasons.
>>54929111
I have noticed mobile users posting screenshots of transparent PNGs that have unreadable text because of the applications image viewer overlaying it over the text posts. If the dev just put it on a white or the boards color, it would solve that issue.
>>54928373
100% on all 8 cores
11/10 thanks anon
>>54929243
Huh, I'm not getting to 100% on my 2 core potato
VP8 also loves line art for some reason.
Static Webm is ~13X smaller than the PNG
>>54929388
>>54928346
>>54928372
You don't have to transform them into webms.
This is a 180kb jpg that looks better than your 125 wemb and way closer to the 1.21 MB original, without the pants on head retarded method of turning it into a 0 length video.
>>54928519
This one looks pretty crap desu
>>54929452
oh wait 150 kb lmao
and looks considerably better than the 125 kb webm
Also webms are utter shit for smooth patterns, see the pic on the OP.
>>54929388
Main problem is the edges of the lines are blurred slightly. If the edges were sharp png could compress that image down to basically nothing with zero quality loss
>>54929388
That picture is not a problem for JPG either.
>>54928193
Thanks, Is there a way to convert it back?
>>54928193
>CPU spike.
>>54929452
>>54929561
But JPG makes them look twice as shitty as VP8. You can clearly tell the JPG versions apart by how blocky and blurry they lock along with all the artifacts without having to look at them much.
>>54928591
>zooming in on webm
I doubt most video players even have.
>>54928193
Stop this cancer
>>54929577
>>54929608
It handles lines quite nicely but it eats up more details than jpg and it's even worse for smooth gradients
How autistic do you need to be to make static webms to make smaller pictures, or care about picture size in 2016 anyway?
>>54929608
Go look at the leaves, the jpg one has a few spots the VP8 one doesn't
Aaaalsoo
Even according to their developers, WebP is about 25-34% smaller than a comparable jpg. You're either a retard, delusional, a faggot or all the former if you believe you can get pictures 1000% (or more) smaller like you're claiming.
>>54929750
Go away, go post more 2-3MB images of frogs with your pissphone
Here have this one, it's on the house.
>>54929545
>>54929780
>Go look at the leaves, the jpg one has a few spots the VP8 one doesn't
Fine, I'll do a comparison. Be prepared to be disappointed by JPG.
>>54929750
Given how shit 4chan servers are, the difference between 3 MB PNG and 150 KB is noticeable even if you're not on mobile, so people who post video screenshots as PNGs are cancer.
But using CPU-hogging and inconvenient WebMs to save a couple dozen kilobytes versus a JPG is autism.
>>54928391
Well then, have a shell version:
JPG+PNG:for i in *jpg *png; do ffmpeg -loop 1 -i "$i" -an -c:v libvpx -qmin 16 -qmax 16 -quality best -threads 4 -t 2 -r 1 "${i}.webm"; done
>>54928346
>>54928372
>>54929452
>>54929904
what's with the black bars?
It's a nice idea but the execution is poor, you're degraining images and introducing color errors for no reason at all.
>>54929545
>>54929870
>>54929964
This. It's also stupidly clumsy to view and interact with them because the browser will treat them as video clips instead of images.
You can shill WebP as much as you want, but for the love of god please just actually use WebP, not WebM.
Why 2 seconds?
>>54929870
>>54928439
>>54928519
>>54928808
These all have clearly visible artifacts, though.
You shouldn't use webm or jpeg on pictures that have not much high-frequency details or simple color patches next to each other with high contrast.
Also get a better connection, poorfags.
VP9 when?
>>54929916
Thanks.
>>54929933
I'm not sure. The static Webm shows as transparent when I open it.
>>54928956
Lossless.
>>54930022
>>54930022
that the real question
>>54930022
Because converting a single image into a webm will cause unnecessary cpu utilization especially on phones as the Webm infinitely loops a single frame. Thus the 2-second duration prevents that from happening.
>>54929735
I meant as in a mass way?
I'm thinking of utilizing this to compress volumes of manga.
And see how this goes.
If it's feasible who knows, maybe /g/ would be regarded for creating manga standard format that /a/ will use.
VP8 vs JPG comparison: http://screenshotcomparison.com/comparison/174956
As you can tell at the same file size JPG looks more noise and blocky, it only gets worse as JPG tries to compete with smaller and smaller file sizes. Admittedly it's a little hard to tell the difference.
>>54930056 is an excellent example of when JPG tries to compete with VP8 at small file sizes.
>>54930086
Clever.
>>54929906
>people who post video screenshots as PNGs are cancer
That's mpv default.
What do you use then?
Im still trying to do it. I made a folder, put there a png and a bat with the code you said, run it but got nothing.
I'm sure is pretty easy to use, but I'm retarded and can't into basics.
>>54930139
JPG version looks better there
>>54929906
>not wanting lossless picture quality
back to imgur faggot
>>54929780
delet this
>>54930148
JPG, obviously.
The source is already lossy; nobody's going to see a difference.
>>54930175
>screenshot from a video
>lossless
coalgirls pls go
>>54928193
>>54928328
Wow you're so cool. Are you a hacker?
>>54930139
Look at the details and noise, JPEG keeps a lot more details.
>>54929178
The choice of picture is reserved for those with the ability to make them
>>54929379
You're probably using the libvpx decoder.
>>54929408
>unoptimized lossess PNG vs lossy
lad..
>>54930139
H.265 at 125 KB
(PNG was too large to attach, this is JPG at 95% quality)
>>54930139
>comparing 34kb to 125
>>54930228
your PNG is full of artifacts, fuck off
>>54930253
>artifacts
I don't think you know what that word means.
>lossy webm is smaller than lossless png
>>54930228
>sharpening out the blurry edges
that's cheating
>>54930236
>(PNG was too large to attach, this is JPG at 95% quality)
This is worthless, then.
>>54930240
You're retarded
>>54930253
protip: png is lossless
>>54930139
>>54930171
I agree: JPEG looks better.
Specifically the JPEG is sharper.
There is a trade-off between sharpness and noise, so it's only natural that it looks noisier.
>>54930227
Now that you mention it does appear that way. Still I think the VP8 version looks better, the lack of blockiness or blurryness is easier on the eyes despite some lost details.
>>54928373
My t420 only got 50% load on each core.
>chromium
>xubuntu
>i7-2620m w/ 8gb ram
>>54929933
>>54930041
Its your web browser I think. Firefox, Chrome/Chromium, and Clover all show the bars. Opera does the transparency.
>>54930284
>protip: png is lossless
Okay.jpg.png
>>54930341
>go from loseless to lossy to loseless
>expect a loseless result
is this b8?
>>54930379
Tell that to >>54930284
>>54928451
Shit eater detected.
>>54930274
>that's cheating
Nope, as we're comparing to a lossy format; I can be lossy too.
>>54930314
libvpx decoder
>>54930321
Fuck you, we're comparing to a lossy format; I'm just being lossy as well. You shouldn't need 678 colours to draw a black and white image with aliased edges; that's just being wasteful.
Here's some middle ground.
>>54930341
Yes we all know when you compress an image using a lossy codec that's not lossless.
JPEG is outdated and hopelessly inefficient by today's standards.
But those webm's are too way blurry for static images.
>>54930417
How about this
I think it's fair that we compare both codecs at extreme compression.
Webm related is encoded with the -qmin and -qmax parameters of 63 and JPG that follows is set to 0% quality. Webm is slightly smaller than the JPG.
>>54930552
>>54930326
obviously we should all switch to opera
>>54930652
b-b-but botnet anon ;_;
>>54930552
>I think it's fair that we compare both codecs at extreme compression.
Why?
Nobody uses extreme compression.
Real world results are what matter, not extreme lab tests.
>>54930499
If he'd set his quality to 0-0 or 0-1 they'd look fine, but only be 1/4 - 1/2 the size of the originals.
>>54930517
Both of ours are about equally bad in terms of absolute error to the original (which I hold is itself bad).
Your diff is the left one, with a mean absolute error or 118, mine on the right with a mean absolute error of 89.
>>54930700
Whoah, how did you do that? Is it a Gimp plugin?
>>54930552
fun stuff but it doesn't matter too much to me
Have you tried a large VP8 file size and compared quality?
>>54930725
>>54928193
>click to open "image"
>it's a video file so it fucks your CPU
>try to click it to close it
>lol nope, you just paused your fucking image
You know what, just fucking kill yourself.
>>54930788
Is your computer 10 years old?
>>54930788
>>try to click it to close it
>>lol nope, you just paused your fucking image
This.
Fuck that.
>>54930809
>Is your computer 10 years old?
The faggot asks this when he's trying to save 100 fucking kb on a fucking image, are you fucking serious?
>>54930700
>>54930725
For fun, here's the webm (a frame at 1 second)'s diff to the original, with a mean absolute error of 195.
>>54930769
Why is his question bait?
>>54930824
kek well you don't have to participate in this thread anon. That cpu spike in usage most likely means you browser sucks. All I get is a 10% cpu usage spike for like 1 second and then it goes away completely.
>>54930824
better then +100kb :^)
>>54928193
simply ebin
>>54930864
>>54930867
>well you don't have to participate in this thread anon
Because I want a bunch of fa/g/g/ots posting pseudo videos for images everywhere. Fuck off, hard drives are cheaper than they have ever been.
>>54928193
Post the webm for fools please
you know, that one... for fools
>>54930885
But anon, I need everything to be minimal just like my 100MB linux installation.
>>54930912
>Hmmm do you apple products?
Not only do I not them, I don't own any either.
>>54928193
>Sadly 4chan likes staying in the past and catering to icucks so Webp support will probably never arrive here.
But VLC can read webm on IOS, senpai.
>>54928346
The only way to have everyone conform is if its done server-side.
JPGs and PNGs are always converted to static webms before upload.
And that would be utterly retarded.
>>54930941
haha yea, I was about to correct that. Well if it makes you sleep better at night apple users will have difficulty opening let alone saving these static Webms.
>>54930959
Or they could just add support for FLIF when it's mature in a few years and call it done.
how?
>>54930949
>But VLC can read webm on IOS, senpai.
You can't save them though since tim cuck wants you all inside that walled garden as much as possible.
>>54930994
Install FFMpeg and save my scripts as a .bat
Trust me, much easier and simpler to use.
>>54930959
>JPGs and PNGs are always converted to static webms before upload
RIP /po/
>>54930992
Either way /gd/ will be pissed.
>>54928193
try thesedwebp --helpcwebp --help
>>54928808
>>54928833
>>54928954
i wonder why same-size webm's load faster for me than jpeg's
>>54930725
Oh, and it's Imagemagick's compare program.
>>54931079
>same-size
the jpg is 2.71MB, the webm is 121KB
>>54931079
OPs idea would help images load faster across all connections. 4chin has limited bandwidth it must share amongst users. It's not like 4chin can feed 1GB/s of bandwith to every user here.
Funfact: Opera's compression servers render jpg's as webp on the fly. (turbo mode)
>>54931117
How do they even pay for these servers. Converting millions maybe billions of files to webp on the fly doesn't sound cheap.
>>54930700
>mfw
>>54928808
wew
>>54931139
with the money they make from selling your personal information.
>>54931100
you've misread something
>>54931079
The JPEG loads progressively while the Webm needs to load the whole frame before decoding and showing it, so it looks like it is loading faster.
>>54931139
Opera is owned by chinese now. Interest on the US federal debt pays for it.
>>54931151
Stop being poor and get a 2GB RAM. They're cheap as fuck too, I saw the warp elite drop to $90 on amazon last week.
>>54931156
It still baffles me that you're choosing png as the comparison format. Start posting comparisons with lossy formats and stop posting bait.
>>54930121
Idiocy at its finest.
>>54931218
It's not a ram problem.
>>54930121
>compress manga to webm
wew lad
Not this retarded shit again.
>>54931300
>all of One Piece in a webm at 100 fps
This needs to be done.
>>54931279
Actually it is. That high res image has to be loaded into RAM. It may be small as a webp but once decoded and loaded into RAM it's big as fuck, probably like 50MB+
jpegoptim, optipng, and jpegmini can shrink image files without noticeably quality loss breddy gud, too.
jpegmini.com online.
png compression works by using the same bits for all pixels with exact color values. Imagemagick 'convert homer_simpson.png -colors 5 homie.png' can shrink a 5 color png if it's not optimized already.
>>54931374
staph.... I like the idea but it would create a surge in manga piracy on 4chan. Don't want anons getting raided by the feds for magoes.
>>54931261
Those are the source PNGs being used anon. I've even gone 1 step further toward PNG's favor and compressed them to level 9 in gimp. This usually shaves off a good 50-100KB in most cases.
>>54931390
No it's not, the problem is the shitty codec they implemented on this shitty phone, some webm doesn't work properly when playing on hardware mode, but with some app like MX Player with software playback they work fine.
>>54931300
>>54930121
It has already been done and it fucking sucks.
>>54931374
Compression would make it unreadable at 100fps.
>>54931439
Converting lossless images to lossy saves space? No shit sherlock.
But guess what? Jpeg can do that too! And it does it without the retardation of using a video format for a single frame.
VP8 can't even compress without chroma subsampling. As long WebP uses VP8 for it's lossy compression it's worthless as a format.
>>54931540
>As long WebP uses VP8 for it's lossy compression it's worthless as a format.
That's like your opinion mang. Also Webp has more support than all the other special snowflake image formats too.
>>54931532
>Compression would make it unreadable at 100fps.
Since when does compression depend on frame rate? You can play the exact same file at any fps if you don't care about audio.
>>54931532
I just read half of that before I realized it.
>>54931691
He's being dumb. It's all about bitrate.
>>54931422
>Don't want anons getting raided by the feds for magoes.
We have survived the sink threads, I doubt DMCA will affect us.
>>54931532
>Compression would make it unreadable at 100fps.
Disregard that, I suck cocks.
>>54931691
Yeah I thought VP8 had some temporal perceptual compression tricks but I must have been thinking of something else.
Post something that will burn my 4690K
>>54928193
>any type of image compression in a 1GB/s internet speeds world with terabytes of free storage available per person
Why the fuck would anyone think they need image compression in [currentyear] unless they lived in a third world shithole?
>>54931891
For the same reason why PNG images take longer to load than a JPG file with the same file size.
>>54928193
Jesus Christ you even went as far as puttin on a trip while proceeding in your inane crusade
Crafting webbums accurately engineered to burst an i5 into flames isn't a good thing
Doing it on windows is even more devilish
Find a hobby more suitable for you
>>54931891
Because 4chan bandwidth is limited and divided between users you retard. Smaller image sizes improve loading times for everyone even if they have a 1tb/s connection to the internet.
>>54931849 see >>54928373
Make sure you're not using the libvpx decoder (ie Chrome, Opera, Safari).
>>54931891
I know right? Also, why optimize your code when we live in the age of ultrapowerful CPUs and a gazillion RAM amirite?
>>54931938
Well I'm running Opera.
Also the Chaika webm uses more CPU than that one on my PC.
>>54931891
not that I'm defending the use of webms, but 4chan servers are slow as molasses.
>>54931532
How to convert this to regular PNG, for easier reading?
>>54931925
>4chan bandwidth is
irrelevant and not "divided between the users", given that any and all image request is served by Cloudflare.
>>54932042
>Cloudflare has infinite bandwidth for all users
k
>>54931891
Why are you such a Luddite
>>54932060
>Cloudflare's infrastructure is constantly saturated and that's why so many sites are using it
k
>>54930700
>Both of ours are about equally bad in terms of absolute error to the original (which I hold is itself bad).
Mine is literally just a grayscale version. It's otherwise lossless
>>54932088
I don't see this as a step forwards but a step back, if the webm method was lossless I'd be all for it. Lossy compression ruins the original image, yes it may be handy now but in the long run it ruins the quality of the images and site overall.
Haven't you experienced the crazy amount of resized images which you had to reverse image search just to get the full resolution which didn't look like shit?
>>54932248
yes, let's use lossless PNG for everything, even videos
>>54932248
I agree webm compression is tardy. I'm only suggesting that shunning any kind of advancements in compression is stupid
>>54932292
I'm all for it, lossless video > x256 any day
>>54932349
retard
>>54932372
nah, I agree that compression is useful in loads of cases, most of them actually, a 50% compression rate is almost unnoticeable when compared to the lossless counterpart with the techniques used today, I just don't think that lossy should be the standard, you don't see companies putting compressed video on blu-rays either.
>>54932451
Blurays are compressed.
They might be badly compressed, but they are definitely compressed.
>>54932451
Are you retarded? Blu-rays use lossy x264 with 4:2:0 subsampled color. They just have fuckhuge bitrate so you can't really see the difference. Truly lossless video would be many times larger.
>>54932451
>you don't see companies putting compressed video on blu-rays either.
They literally do you mongoloid.
Just 2 hours of raw 1080p24fps video is ~1,000GB. A blu-ray disc is only 25-50GB. Movie companies have to compress that raw ~1TB of video into those discs.
Luckily video compression has become so advanced that even a ~10GB 1080p h264 movie rip will look exactly the same to a viewer comparing it against the raw 1TB of video.
>>54932451
cont.
Lately everyone around me sees YIFY encodes as the best "quality", "because it's 1080p right!" while the quality is actually not that great, bad even. Standards are getting lower and lower while they should be getting higher and higher just like what happened with the music industry and all their weird mediums.
>>54932512
>>54932522
>>54932562
I'm probably retarded, it's kinda late here. still, the new uhd blu-rays are also getting a higher bitrate next to a higher resolution improving the color depth and quality. At least they are trying to improve instead of compressing it further and further.
>>54932578
You're getting a $20-60 movie for FREE you ungrateful nigger. You have no right to complain about yify quality. Go circlejerk with your private tracker fags if you want your 50GB blu-ray muxes.
>>54932626
I wasn't talking about piracy being good, I'm merely pointing out that piracy is actually degrading standards, if people were actually aware about quality differences they wouldn't want YIFY and would probably be fine buying the blu-rays instead of downloading a 20 to 130GB file just to get some decent quality.
>>54932451
>you don't see companies putting compressed video on blu-rays
What? An uncompressed 2-hour movie in 1080p takes up around 1 terabyte. Blu-Rays are 25-50 GB, and typically have extras.
>>54928373
Whats happening here?
Why is this webm in particular so demanding?
I have a 5820k and it uses 20% of all six cores.
>>54932451
>you don't see companies putting compressed video on blu-rays either.
retard
>>54932773
yes yes, we already came to that conclusion.
>>54932756
high bitrate
>>54932451
but they do compress videos to fit them on blu ray
retard
>>54928372
This loads instantly for me. Instead of fucking with image quality and making a worse experience for everyone, maybe you should join us in modern times with decent Internet connections.
It's 2016, you shouldn't have to rape pictures with lossy compression.
>>54932834
>implying jpg isn't lossy compression.
>>54932834
The bandwitdh of the user has nothing to do with the bandwidth 4chan has you micro cuck. You can't just "buy" a 1TB/s 4chan connection.
>>54932853
You just wait, soon they'll have to start a 4chan patreon just to keep the site afloat.
>>54932877
They already have. If you don't have a pass 4chan will cuck you to death with solving captchas.
>>54932756
extremely high frame rate
>>54932756
>>54932818
no, high framerate
More importantly, FLIF when?
http://flif.info/
>>54932847
Some lossy compression is acceptable, but the less, the better. Every one of these VP8 images has incredibly noticeable artifacting. For example, look here >>54928439 at the several giant blue bands it threw into the picture. It also warped all of the line art. That's most noticeable if you look at the edge of the brown edge of the mat. But all of the lines are blurrier.
And what do we get out of switching to this? Some Africans have a better experience browsing 4chan on their shitty internet?
>>54932853
4chan serves shit over CloudFlare, which has some of the biggest pipes available. If images are taking a long time to load, it's your fault.
>>54932909
Never it's a meme and DOA. Nobody fucking supports it.
>>54928225
The webm has some fairly obvious horizontal lines across the black background compared to the jpg for me. Other than that it's pretty good
Great
>>54933013
Use opera and/or check your INTERNET connection.
>>54932942
No fucking shit, the alpha release was only a few months ago.
>>54933044
Yeah because I should switch to a proprietary web browser to use your retarded static WebM shit. No thanks, I'll just watch my fucking picture buffer.
>>54932451
A raw video is fucking huge. Just take some recording program that doesn't encode right afterwards on your computer and see how it goes
>>54932451
>you don't see companies putting compressed video on blu-rays either.
breh
>>54932451
OF COURSE they use compression
raw video is huge you fucking retard
>>54933044
The webm is 40ms long; it's in an endless cycle of starting over. Whoever made that webm is retarded.
>make one mistake
>admit said mistake
>still 7 other people pointing out same mistake
I love you /g/, never change
>>54928193
>autism the post
>>54933082
It's not hard to just calculate the size of raw video
(frames per second) * (resolution) * (bits per pixel) * (video length in seconds) = (video size in bits)
So a 60 minute 1080p video at 24fps and 8bpp would be 166.86 GB.
>>54932451
>they dont use compression
no, they just magically fit 49TB of raw video onto a 50GB blu ray disc
>>54933142
You're right. I'm sorry I doubted you anon-kun.
Here is the fixed version.
>>54933197
>49TB
You are off by a few orders of magnitude. You should probably read the post above yours, since an uncompressed feature length movie would be 350 - 400GB at most.
>>54933197
Yes, the big corporations already have highly advanced lossless compression technologies which enables them to do this.
>>54928193
Is web or a static webm really that much better than jpg and png?
Complete noob here btw
>>54933217
as i said
49TB
>>54933252
No, they look like shit.
>>54933252
> anonymous
>a
>Complete noob here btw
Checks out
>>54932451
>how the fuck do blu ray discs work: the post
>>54933252
yes, if you had compressed that image using the webm method, the scales would actually still reflect light like they did in the original book.
>>54933252
at least you didnt put your jpg on a blu ray
>>54933259
>I was pretending to be retarded
Even if you were using hyperbole, it doesn't fit here. BDs come in sizes of 25 GB, 50 GB, 100 GB, and 128 GB. You can easily fit a feature length movie onto a BDXL using lossless compression.
>>54933164
>>54933217
Wrong. Everyone knows 1 hour of uncompressed 1080p24fps video = 334GB. 2 hours would be like 668GB, add in all that extra gay shit they cram in blu-rays and you have now either reached 1TB of RAW uncompressed video or exceeded it.
https://en.wikipedia.org/wiki/Uncompressed_video
>>54933164
Your video is a gif.
>>54933309
>You can easily fit a feature length movie onto a BDXL using lossless compression.
Stop posting.
>>54933309
but anon
it was a hyperbole
raw video is huge compared to a blu ray disc
49TB is also huge for a file
my hyperbole is in place
>>54932451
name one company that doesn't compress their video for blu ray faggot
