God dam this is way more difficult than I thought it would be. The only other thing I can think of is converting the PNG to a Y4M file.

for %%f in (*.png) do (
ffmpeg -framerate 1 -stream_loop 1 -i "%%f" -pix_fmt yuv420p -y temp.y4m && ffmpeg -color_range tv -colorspace bt709 -color_primaries bt709 -color_trc bt709 -i temp.y4m -c:v libvpx-vp9 -quality best -crf 10 -qmin 10 -qmax 10 -pix_fmt yuv420p "%%~nf.webm"
)

Uploaded webm mediainfo reads:

Color range : Limited
Color primaries : BT.709
Transfer characteristics : BT.709
Matrix coefficients : BT.709

Hmmmm, I think 4chan is stripping one of these. The uploaded webm mediainfo reads:

Color range : Limited
Color primaries : BT.709
Transfer characteristics : BT.709
Matrix coefficients : BT.709

Here's 11 loops at 6 fps totaling 2 seconds. Could this be the holy grail?

Anons does this webm give you perfect playback on mobile/desktop web browsers and video players? No weird color glitching or CPU raping?

Okay I think I figured something out. Framerate seems to affect the VP9 decoder and 6 might be a magic number for that. So then all that we have to figure out is how many stream loops we need.

No loops = 166.666 milliseconds

5 loops = 1 second

Does this solve playback problems for some of you anons? I've increased the framerate and stream loop both to a value of 10. It takes 10X longer to encode but my hope is that this creates a more "correct" VP9 "video" stream...

>>106029228
>>106029280
I don't know man, this is just a dumb hack afterall. MPC/VLC both spaz out and rape my CPU in addition to the chrome web browser video player with your 40 millisecond video webm. I have no idea why though.

I'm just assuming that in order for this to work at all we have to trick the VP9 decoder into believing it's decoding real VP9 video somehow but maybe there's something else missing if my webms are causing playback/color shifting problems for other anons.

Then again this could be a software/browser issue?