[Boards: 3 / a / aco / adv / an / asp / b / bant / biz / c / can / cgl / ck / cm / co / cock / d / diy / e / fa / fap / fit / fitlit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mlpol / mo / mtv / mu / n / news / o / out / outsoc / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / spa / t / tg / toy / trash / trv / tv / u / v / vg / vint / vip / vp / vr / w / wg / wsg / wsr / x / y ] [Search | Extra juicy! | Home]

>encoding WebMs in VP9 on ffmpeg is slow as fuck compared to VP8

This is a blue board which means that it's for everybody (Safe For Work content only). If you see any adult content, please report it.

Thread replies: 17
Thread images: 1

File: 1446511993155.png (553KB, 678x718px) Image search: [iqdb] [SauceNao] [Google]
1446511993155.png
553KB, 678x718px
>encoding WebMs in VP9 on ffmpeg is slow as fuck compared to VP8
>>
>>51488363
literally no point in using VP9 (literally)
>>
>>51488363
Send me the video you're trying to encode and the settings on libvpx, because the last time someone claimed that vp9 is impossible to encode in I tried a random video and got 4 fps single threaded on a laptop CPU. That's definitely slow, but it's not that bad.
>>
>>51488400
I'm using WebM for Retards, though, which is basically a frontend for ffmpeg plus AviSynth.

https://gitgud.io/nixx/WebMConverter

The source is a few MB-large AVI that's not even 1 minute long. It took a few minutes to convert.
>>
>>51488438
I literally used the ffmpeg guide to do it:
https://trac.ffmpeg.org/wiki/Encode/VP9

And then I specified
-threads 1
for benchmarking.
>>
>>51488536
Did you use a constant or variable bit rate?
>>
>>51488363
The VP8 encoder is also slow as fuck.
>>
>>51488553
I think the exact line I used was
ffmpeg -i input.mp4 -c:v libvpx-vp9 -b:v 1M -na -threads 1 output.webm


With the files obviously changed. Let me try it again with a short video and see what happens.
>>
>>51488555
It's still far faster, though.
>>
>>51488591
libvpx encoders are just slow in general compared to H.264 & H.265. Of course VP9 is slower due to being a more complex & efficient codec, but building a codec without violating patents is kinda difficult.
>>
>>51488609
Will VP10 be even slower?
>>
>>51488624
Due to rising complexity, properly.
>>
>>51488553
>>51488579
Interesting, single threaded was around ~5 fps, but doing -threads 3 seemed to only spike my CPU to like 60%, and still it said 6 fps, which seems too low for using at least double the resources.

I'm not sure how the stdout works with ffmpeg but the 3 threaded test seemed to finish much quicker. I don't know if it was 3x quicker though.
>>
>>51488663
Properly?
>>
>>51488910
*probably
Dunno how that happened.
>>
>>51488609
It's slow compared to x264. x265 is basically equally slow, though.

You can always try a higher speed setting to trade off some efficiency for framerate.
>>
Oh fuck yes finally a Zeranoe binary for Windows which includes libvpx 1.5.0.

http://ffmpeg.zeranoe.com/builds/

I remember getting 0.7 frames a second encoding on practically the highest possible quality two pass encodes with VP9. I'll try those placebo-level settings with this new library. It's supposed to hold some performance and quality improvements but I have no idea if they will make any difference here.
Thread posts: 17
Thread images: 1


[Boards: 3 / a / aco / adv / an / asp / b / bant / biz / c / can / cgl / ck / cm / co / cock / d / diy / e / fa / fap / fit / fitlit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mlpol / mo / mtv / mu / n / news / o / out / outsoc / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / spa / t / tg / toy / trash / trv / tv / u / v / vg / vint / vip / vp / vr / w / wg / wsg / wsr / x / y] [Search | Top | Home]
Please support this website by donating Bitcoins to 16mKtbZiwW52BLkibtCr8jUg2KVUMTxVQ5
If a post contains copyrighted or illegal content, please click on that post's [Report] button and fill out a post removal request
All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.
This is a 4chan archive - all of the content originated from that site. This means that 4Archive shows an archive of their content. If you need information for a Poster - contact them.