Hacker News new | ask | show | jobs
by GeneticGenesis 564 days ago
I know it's not public domain per-say, but for me, the thing that's most exciting is that in 2025, the last remaining patents on the h.264 (AVC) video codec will expire [1].

Now if only HEVC wasn't such a hot patent / licensing mess.

[1] https://meta.wikimedia.org/wiki/Have_the_patents_for_H.264_M...

6 comments

Just thought you might want to know - it's "per se" not "per say"/variations thereof.
Per se is latin for "for itself".
Maybe word for word, but "per se" means "as such".
I'd say "by itself".
If you translate it literally, "per" is closer to "for".

If you don't translate it literally, I'd vote for "in itself". "In itself" (viewed in its essential qualities; considered separately from other things[0]) has a different meaning than "by itself" (alone/unaided). And to me it's clear that "per se" pretty much universally means the former.

[0] https://www.google.com/search?q=in+itself

A less literal translation like "essentially" or "in essence" is deployed by master Latin translators like Robert Fagles. I've even seen "in a vacuum" which does a better job at communicating the original intent than a string of cryptic prepositions.
Sometimes I use it (maybe wrongly?) as a synonym to "technically".

"Maybe John is not the boss per se, but we follow his orders"

"Maybe John is not technically the boss, but we follow his orders"

That's another valid translation for the same preposition.

And there are many definitions of English "for" as well. This would fit the one used in the phrase "if not for this, ..." In other words, for itself = by virtue of itself, through the existence of itself.

Also note in terms of Indo European roots, per is a cognate with English for.

Prepositions are some of the least translatable bits of language. For that matter, even without translation they tend to get slippery within a language, especially over time (one that springs to mind is the whole “quarter of” referring to a time which I first encountered some 50 years ago and still don’t know if it’s quarter to or quarter after).¹

1. Cue some dude to tell me in 3…2…1²

2. And this knowledge will promptly disappear from my brain five minutes later, sort of like the guy I knew in my 20s whose name was either Jack or Chad and to this date, I still am not sure, but I do know that every single time I called him by name, I got it wrong and it totally wasn’t on purpose even though he didn’t believe me.

That's my cue!

I once had a Spanish teacher, who also had problems remembering what that kind of time specification stands for and I came up with maybe a trick to remember. We do the same thing in German, so I guess it translates:

Lets say you have 11:00. That's easy. But what about 11:15? We would say "quarter 12", so I guess the English version is "quarter of 12". How to memorize, that this is 11:15? Well, you can imagine a round clock and the minute pointer has moved _quarter of its way to 12_. So you only have a quarter of that hour "already done". 10:30? We say "half 11". So I guess English is "half of 11", meaning that the minute pointer has moved half the way to 11.

Maybe this will help.

(Actually I personally usually don't use those ways of specifying the time, neither in English nor in German. I just say the 24h format as it is written: "11:15" is "eleven fifteen", 13:35 is "thirteen thirtee five" not 1pm something.)

> quarter of

Still can't beat stuff like "bi-weekly" which can mean "every two weeks" or "twice a week" or probably some other thing as well.

Tok Pisin, a.k.a. New Guinea Pidgin, has exactly two prepositions: bilong, which means "of" or "from" in a possessive or attributive sense; and long, which means everything else.
> Now if only HEVC wasn't such a hot patent / licensing mess.

Somehow I suspect HEVC suddenly became a thing in the past few years precisely because AVC patents are expiring.

Yes, and in fact this is explicitly the business model[0] of ISO MPEG and ITU VCEG. They pay for their basic research by letting participants patent and license the resulting standards-essential inventions[1].

HEVC/H.265 has been in development since 2004, i.e. right after AVC/H.264 was published, and took almost a decade to actually be standardized. There's even an H.266, which started in 2017, a few years after H.265 was released. Though the primary concern of patent holders is not AVC patents expiring. Those patents actually aren't that valuable, because AVC is licensed way too cheap. MPEG-LA had negotiated a very generous free rate for online video[2], in response to MPEG-4 ASP (aka "DivX :-)") basically not getting much use online.

What patent owners want is to go back to the days of MPEG-2 where they were making money hand over fist just for owning a functional codec. They even sacked Leonardo Chiariglione, the founder and head of ISO MPEG, because he was trying to change ISO's patent policy to be more favorable to developing royalty-free codecs.

[0] ISO does not license patents and has no affiliation with MPEG-LA/Access Advance/etc, but Leonardo has gone on record saying this is their 'business model': https://blog.chiariglione.org/a-crisis-the-causes-and-a-solu...

[1] under FRAND licensing

[2] Which is why YouTube's allowed to use H.264 without paying $$$ for it. Before that, they used whatever codec was available in Flash Player. Adobe (and Macromedia before it) used On2 VP6 primarily because it had no patent licensing royalty; before that they'd used H.263.

Encoding efficiency for a given perceptual quality is very important when you pay for bandwidth or disk space.

Otherwise there would have been no effort to create vp9 and av1, as everyone on that side of the codec wars would've stuck with vp8.

That's incredible. With MP3 already completely patent-free as well, we have an extraordinary free set of audio and video codecs for the next couple of decade, at least until HEVC becomes free.
Let's not forget Opus. Not technically patent-free but it practically is. https://en.wikipedia.org/wiki/Opus_(audio_format)#Patent_cla...

Also Vorbis has always been patent free.

Opus and Vorbis are far from useable everywhere. On the other hand, there isn't a piece of software or hardware that won't accept MP3. MP3 is the lingua franca of sound, and should stay that way for a century.
One of the primary reasons why AV1 exists is because HEVC is such a hot mess.
There are already two patent pools for av1 that want rent.
So 1 less than the 3 HEVC pools, and in HEVC they were all part of the development process so have pretty iron clad claims.
Do you think AOM is going to start charging royalties?
No, Sisvel and Avanci are.
How does that work? They force AOM to pay up and then I guess either AOM passes the royalty burden onto AV1 users or they take the hit and pay it themselves?
They did this, paid then off and took the hit, for some VP9 patents.

They also snipe away at the shittier parents in court.

Plus they have the "if you sue us over AV1 patents, you no longer get to use AV1 for free, patent landmines in place”.

Patent pools want money from companies and end users.
A tale as old as video codecs.
In the link it seems the last patent in US go as long as 2027?

If the patents really expire in 2025, is there an already open source library written either in C or C++ one could use for reading h.264?

Cisco published their implementation under BSD license:

https://github.com/cisco/openh264/

They also make a reproducible build of it for firefox in order to shield mozilla from patent suits.
x264 has been around forever, and it's FOSS.
Dude, FFMPEG has open source decoders for pretty much every codec ever created.

As the sibling points out, for H.264 we even had a high-quality open source encoder for a long time.