Linked by David Adams on Mon 26th Jul 2010 00:30 UTC, submitted by boulabiar
Multimedia, AV Just after 3 weeks of the binary compatible vp8 decoder release, the FFMpeg team still impressing us but this time with a new benchmark of their own vp8 decoder. The new ffvp8 decoder written independently using pre-existent FFMpeg code-base is now the fastest vp8 decoder with margins going more than 30% faster than Google's official codec specially on 64bit machines.
Order by: Score:
Congratulations, OSS Devs!
by organgtool on Mon 26th Jul 2010 13:53 UTC
organgtool
Member since:
2010-02-25

This isn't the least bit surprising since the VP8 decoding algorithm is supposed to be quite similar to the h.264 decoding algorithm and the FFMPEG implementation of h.264 decoding (x264) is recognized as one of the fastest implementations of the standard. Hopefully news like this continues to achieve better adoption of open source software as well as open standards.

Finally, I want to give a big "thank you" to all of the developers that contributed to this accomplishment and any companies that supported this (especially Google)! I know there's still more work to be done, but this is a great start to a project that will make video decoding more accessible to many users.

Reply Score: 5

RE: Congratulations, OSS Devs!
by Gusar on Mon 26th Jul 2010 15:56 UTC in reply to "Congratulations, OSS Devs!"
Gusar Member since:
2010-07-16

x264 is an encoder and is actually a project independent of ffmpeg, though they do share developers.

Reply Score: 1

RE: Congratulations, OSS Devs!
by Zifre on Mon 26th Jul 2010 16:22 UTC in reply to "Congratulations, OSS Devs!"
Zifre Member since:
2009-10-04

the FFMPEG implementation of h.264 decoding (x264) is recognized as one of the fastest implementations of the standard.

x264 is not a decoder and nor is it part of FFMPEG.

Reply Score: 4

of course its faster
by FunkyELF on Mon 26th Jul 2010 14:51 UTC
FunkyELF
Member since:
2006-07-26

Its using patented algorithms.

I'll use a 30% slower implementation if its really free.

Whatever... choice is a good thing. Keep them off the payroll.

Reply Score: 1

RE: of course its faster
by Zifre on Mon 26th Jul 2010 15:17 UTC in reply to "of course its faster"
Zifre Member since:
2009-10-04

Its using patented algorithms.

Where did you see this? I see no mention of it in the article.

I'll use a 30% slower implementation if its really free.

Just about every piece of software has patents on it. I'm sure that Google's implementation infringes on some patents too, as well as nearly every piece of software you have ever used. It is unavoidable (which just proves how utterly idiotic and incompetent the patent system and all the people who created it are).

Whatever... choice is a good thing. Keep them off the payroll.

So you want Google's official implementation to remain bad?

Reply Score: 2

RE[2]: of course its faster
by FunkyELF on Mon 26th Jul 2010 17:00 UTC in reply to "RE: of course its faster"
FunkyELF Member since:
2006-07-26

Just about every piece of software has patents on it. I'm sure that Google's implementation infringes on some patents too, as well as nearly every piece of software you have ever used.


The difference being that Google's VP8, even if it does infringe on patents, has its own patents to fight back. If Google's VP8 and H.264 infringe on each other they are both probably safe. If ffmpeg's x.264 doesn't hold any patents it is a huge target.

I'm not saying I agree with it at all.... I don't.

Reply Score: 2

RE[3]: of course its faster
by lemur2 on Tue 27th Jul 2010 01:52 UTC in reply to "RE[2]: of course its faster"
lemur2 Member since:
2007-02-17

"Just about every piece of software has patents on it. I'm sure that Google's implementation infringes on some patents too, as well as nearly every piece of software you have ever used.
The difference being that Google's VP8, even if it does infringe on patents, has its own patents to fight back. If Google's VP8 and H.264 infringe on each other they are both probably safe. If ffmpeg's x.264 doesn't hold any patents it is a huge target. I'm not saying I agree with it at all.... I don't. "

In actual fact, the USPTO is not supposed to issue the same patent twice. The subject matter of patents is "methods of doing <something>".

Both MPEG LA members and On2 were in the business of designing codecs and applying for patents on those designs. There is absolutely no reason why any new method any party useed they should not apply for a patent. If the USPTO did not grant a given patent application because there was prior art, then all parties can use that method (no-one can have a patent).

The most likely outcome is that where codecs use the same methods, these methods cannot be patented because of prior art, and where they use different methods, each party holds a patent for that method.

The most likely outcome is that they don't actually infringe on one anothers patents at all.

Reply Score: 2

RE: of course its faster
by sicofante on Mon 26th Jul 2010 15:20 UTC in reply to "of course its faster"
sicofante Member since:
2009-07-08

Software patents apply to a tiny portion of the world. Please see my post below.

Reply Score: 2

RE: of course its faster
by Gusar on Mon 26th Jul 2010 16:00 UTC in reply to "of course its faster"
Gusar Member since:
2010-07-16

Its using patented algorithms.

To quote Dark Shikari: "Yes, some is shared with the H.264 decoder, but that’s because those parts do exactly the same thing. If that counts as “shipping H.264 code”, shipping libvpx is “shipping H.264 code”. Source: http://x264dev.multimedia.cx/?p=499#comment-5776
So ffvp8 is not different from libvpx regarding patents.

Reply Score: 1

How about a better encoder?
by Zifre on Mon 26th Jul 2010 15:12 UTC
Zifre
Member since:
2009-10-04

A new faster decoder is great, but I think the effort would be much better spent on a new encoder.

Right now, H.264 is quite a bit better than VP8, but I suspect that with a VP8 encoder that is as good as x264, the gap would nearly vanish.

The problem is, decoders are much simpler. I could probably write my own VP8 decoder given the spec and enough free time. But good encoders take teams of experts. Right now, even Google doesn't have a very good encoder for VP8.

Reply Score: 4

The world is not the US
by sicofante on Mon 26th Jul 2010 15:17 UTC
sicofante
Member since:
2009-07-08

Just a polite reminder to all of you worried about software patents and how they apply to the VP8 algorithms. The world is bigger than the US. We can use, buy, sell and distribute FFMpeg's software in Europe and many other areas in the world. It's you, Americans (and Japanese and maybe someone else) who have a problem, not the rest. You Americans might as well start some civil fight instead of asking developers to comply with your country's twisted IP law.

A big thank you to the FFMpeg devs and hope they keep developing their great software.

Reply Score: 10

RE: The world is not the US
by Zifre on Mon 26th Jul 2010 16:19 UTC in reply to "The world is not the US"
Zifre Member since:
2009-10-04

I understand your point, but it is kind of arrogant.

I would love to be the person living in Europe, laughing at all those silly Americans and their software patents. However, I live in the US, and software patents are a real concern here.

I'd love to be able to do something about it that didn't require me to move to another country, but realistically, there is nothing I can do.

The problem is, people are just too stupid.

You should never underestimate human stupidity, and for that reason, you should care about software patents. Someday, they may come to your country, and then you won't be so happy.

Reply Score: 6

RE: The world is not the US
by ferrels on Mon 26th Jul 2010 16:56 UTC in reply to "The world is not the US"
ferrels Member since:
2006-08-15

Hey, don't blame the average American for the screwed up IP laws in the US. Blame corrupt courts that are swayed by big money from big corporations. Unfortunately everything in the US is for sale for a price, even justice.

Reply Score: 2

RE[2]: The world is not the US
by Boldie on Mon 26th Jul 2010 17:50 UTC in reply to "RE: The world is not the US"
Boldie Member since:
2007-03-26

__start troll__ I blame the average american. The average (voting) american voted for Bush jr, twice. How did that happened? Did the average citizen mix it up even worse than Bush's "There's an old saying in Tennessee — I know it's in Texas, probably in Tennessee — that says, fool me once, shame on — shame on you. Fool me — you can't get fooled again." ;) __end troll__


But to stay on topic. I also understand that if we are going to have a thriving FLOSS community we have to involve all those bright american developers (even republicans). To do that we have to have software not under threat from patent litigation. I still hope that VP8 will be cleared in court because it can give us a more free internet.

Reply Score: 3

RE[2]: The world is not the US
by Damnshock on Mon 26th Jul 2010 19:09 UTC in reply to "RE: The world is not the US"
Damnshock Member since:
2006-09-15

Unfortunately everything in the US is for sale for a price, even justice.


That was precisely my impression when I was in the States for some months. The only important thing was *money*. Obviously not everyone thinks like that but it was my feeling in all the states I was in and the couple of universities I attended to.

It was something that made me change my mind in ever thinking going to live to the US. (never say never though).

Sorry for the offtopic, just felt the need to give my two cents.

Regards

Reply Score: 2

Time for a new round of benchmarks?
by jgfenix on Mon 26th Jul 2010 16:13 UTC
jgfenix
Member since:
2006-05-25

Now i would like to see a comparation against various h264 and mp4 decoders.

Reply Score: 1