To the webm encoder guy,
Why are you using multiple frames for your webms? If you are making a two-second webm (good practice, as some browsers get confused by very short loops), you can set the framerate to 0.5fps and save even more bandwidth
this file was encoded with the invocationffmpeg -loop 1 -i maki\ is\ cute\ 548.png -c:v vp8 -qmin 4 -qmax 6 -minrate 40M -maxrate 40M -t 00:00:02 maki1.webm
See the next post for the same file, encoded with the addition of a framerate argument
ffmpeg -loop 1 -r 0.5 -i maki\ is\ cute\ 548.png -c:v vp8 -qmin 4 -qmax 6 -minrate 40M -maxrate 40M -t 00:00:02 maki2.webm
this is the original file btw
What is the advantage of using minrate and maxrate instead of crf when you have qmin and qmax set?
Makiposter, don't bother with an autist that has higher powerautism level than you
>>56816805
I like her 7 foot long arm.
>>56816958
shit
>>56816958
Fuck I'm dying hahahaha
>>56816958
Kek. was this on purpose?
maki has extremely long arms so that she can touch your heart, wherever you are
Why are you encoding still images as a video at any frame rate?
because its a more optimal method of lossy compression than jpeg, and 4chan doesn't support webp
>>56816780ffmpeg -loop 1 -r 0.5 -i $(find ~/makis/ -iname \*.png | shuf -n1) -c:v vp8 -crf 4 -b:v 9001k -deadline best -t 1 out.webm
Use -deadline best, senpai. It allows ffmpeg to spend longer figuring out the the best way to compress the frame
nice ty
also i didnt know about shuf that's super handy
testing if this one confuses my browser as well, >>56817438 and >>56816780 do
2 seconds seems to be the sweet spot, but it probably shouldn't matter how long it is so long as you set the framerate appropriately
>>56817622
More like -r 1 -t 2, just like THE autist adviced us... The size difference is not going to be that big in theory as the additional frames are just telling the decoder to redraw the first keyframe
>>56817731
>>56816771
Quit giving the shitposter tips
>>56817808
>>56816936
Web guy here: I encode a 2 second video with 1 FPS because it prevents unnecessary high cpu utilization.
What you just did (forcing 1 frame in 2 seconds) can cause high cpu utilization in so.e browsers.
tl;dr better safe than sorry and extram frame is like 1-2KB extra
>>56817867
>>56817890
whoops meant to quote >>56816771
>>56816936
CRF does jack shit in VP8. Changing the qmin/max settings is the best quality control possible in VP8.
>>56817916
>>56817923
You're late. We already started circlejerking without you. At least go get more lotion while we cool down for round 3.
>>56817958
now you're being silly
>>56817944
>>56817998
Original image: click to view it in full size
Webm: click to play/pause
WOW
Useful!
>>56817986
you seem upset
webm bro pls webm this and save me a kb or two if possible.
>>56818017
>>56818027
>>56818056
Okay lotion boy, unzip your pants.
>>56818041
>>56817926
>CRF does jack shit in VP8
Interesting, pic related now with -qmin 0 -qmax 0.
Problem however with these better quality options is that it doesn't really beat jpg in quality for the same file size, it's only in the autismo small files where some benefits could be had and that only benefits those living in US or in other 3rd world nations with trash tier internets
>>56818114
Not sure how many more bytes i can take off this
>>56818140
it looked like shit long ago. just stop
>>56818148
still looks better than the png desu
and that's the smallest I can goffmpeg -loop 1 -r 0.5 -i input.png -c:v vp8 -qmin 63 -maxrate 1k -deadline best -t 00:00:02 output.webm
than the jpg
its late
>taking shitposting this seriously
>>56818120
why use it though for these small files, a few kilobytes matter little compared to the inconveience of having a webm. And for very big images you can't zoom in(i. e. view parts in original resolution.
>>56818246
maximum fucking kek
i only have 2 tabs open but my browser is eating 30% cpu for some reason
did u just hack me
>>56818341
ChromeOS doesnt have this problem
>>56818341
>>56818348
>only 30%
>>56818411
now do one with webm. That's right: you cant because theyre arbitrarily limited to 2048x2048 here on 4chan and were not allowed to use vp9 either
>>56818417
>implying that i haven't tried already
It's really annoying isn't it. I tried arguing on /qa/ about it but as usual the people who hang out there just wanted to agree with current policy
>>56818431
webp isn't supported in firefox so no.
>>56818431
also, why bother with webp when VP9 webms would actually make a difference?
Why does ksthis happen?
>>56818431
Yeah, cuckzilla is probably paying 4chan to not adopt webp yet in hopes of the sjw jpg encoder becoming relevant. Though eventually as firecocks plummits to 1% marketshare they will probably add support for webp out of the blue.
Only a matter of time I guess.
>>56818485
Shitty video player. Use mx player or Clover.
>>56818485
pix format in that one is yuva480p which your media player can't hand, this is with yuv480p and should play normaly
>>56816771
do the world a favor and kys. you're too fucking autistic to live.
>>56818549
yuv[a]420, not 480
>>56818558
>no fun allowed
good luck with that
>>56818613
get a fucking blog.
>>56818627
What do I need a dog for? To bark and shit itself every 5 minutes? No thanks.
>>56818246
fixed
>>56817926
>CRF does jack shit in VP8. Changing the qmin/max settings is the best quality control possible in VP8.
Could you elaborate on this point?
I use "constant quality mode" (or maybe it is constrained quality) to encode webms for 4chan. A lot of times I will use two passes.
Changing the '-crf' parameter makes a difference. The difference is noticable to me when crf is changed in the range of 5 to 10 (or more).
When you say it doesn't do anything, are you saying that with respect to this thread (making static 2 second images)...or is there something more that I am missing?
>>56816771
>>56816780
I can't zoom in on these...
>>56816805
I can zoom in on this.
>>56818663
!!!!!!!!!!!!!!!!!!!!!!!!!!!
this thread is gold