If you have a predictor/compressor LLM which was trained on all the movies in the world, would that not also be infringement?

We started CCCP because at the time, anime fansubs were predominantly traded on P2P filesharing services like Kazaa, Gnutella, eDonkey, Direct Connect, and later Bittorrent. The most popular codec pack at the time was K-Lite / Kazaa Codec Pack which was a complete and utter mess, and specifically for fansubbing, it was hard to get subtitles to work properly unless they were hard embedded. Soft-subbing allowed for improvements, and there were a lot of improvements to subtitling in the fansubbing community over the years, one of the biggest came when the Matroska (MKV) container format came about, that allowed arbitrarily different formats/encodings to share a single media container, and the community shifted almost entirely to ASS formatted subtitles, but because an MKV could contain many different encodings, any given MKV file may play correctly or not on any given system. CCCP was intended to provide an authoritative, canonical, single-source way to play fansubbed anime correctly on Windows, and we achieved that objective.

But let's be clear, nobody involved was under any illusions that the MPEG-LA or any other license holders of for instance h264 were fans of our community or what we're doing. Anime fansubbing at all came out of piracy of foreign-language media into the English market via the Internet and P2P filesharing. None of us gave a shit, and the use of Soviet imagery in the CCCP was exactly a nod to the somewhat communist ideal that knowledge and access to media should be free, and that patent encumbering codecs and patenting software isn't just stupid, it's morally wrong. I still strongly feel software patents are evil.

Nonetheless, at no point was CCCP through it's life fully legal/licenses appropriately for usage, and effectively nobody cared, not even the licensing authorities, because the existence of these things made their licenses for encoders more valuable for companies producing media, as it was easier for actual people to consume.

The GIF page on wikipedia has an entire section for the patent troubles https://en.wikipedia.org/wiki/GIF#Unisys_and_LZW_patent_enfo...

> At long last everybody realises that the old MPEG business model is now broke

And the entire post is about how dysfunctional MPEG is and how AOM rose to deal with it. It is tragic to waste so much time and money only to produce nothing. He's criticizing the MPEG group and their infighting. He's literally criticizing MPEG's licensing model and the leadership of the companies in MPEG. He's an MPEG member saying MPEG's business model is broken yet no one has a desire to fix it, so it will be beaten by a competitor. Would you not want to see your own organization reform rather than die?

Reminder AOM is a bunch of megacorps with profit motive too, which is why he thinks this ultimately leads to stalled innovation:

> My concerns are at a different level and have to do with the way industry at large will be able to access innovation. AOM will certainly give much needed stability to the video codec market but this will come at the cost of reduced if not entirely halted technical progress. There will simply be no incentive for companies to develop new video compression technologies, at very significant cost because of the sophistication of the field, knowing that their assets will be thankfully – and nothing more – accepted and used by AOM in their video codecs.

> Companies will slash their video compression technology investments, thousands of jobs will go and millions of USD of funding to universities will be cut. A successful “access technology at no cost” model will spread to other fields.

Money is the motivator. Figuring out how to reward investment in pushing the technology forward is his concern. It sounds like he is open to suggestions.

He first points out that a royalty-free format was actually better than the patent-pending alternative that he was responsible for pushing.

In the end, he concludes that the that the progress of video compression would stop if developers can't make money from patents, providing a comparison table on codec improvements that conveniently omits the aforementioned royalty-free code being better than the commercial alternatives pushed by his group.

Besides the above fallacy, the article is simply full of boasting about his own self-importance and religious connotations.

Such a huge catch that the companies that offer you a royalty-free license, only do so on the condition that you're not gonna turn around and abuse your own patents against them!

How exactly is that a bad thing?

How is it different from the (unwritten) social contracts of all humans and even of animals? How is it different from the primal instincts?

Unsurprisingly companies that are losing money because their rent-seeking on media codecs is now over will spread FUD [0] about royalty free codecs.

[0] https://en.wikipedia.org/wiki/Fear%2C_uncertainty_and_doubt

Av1 for 7

The problem is every platform wants to force their own codec, and get earn royalties from the rest of the world.

They literally sabotaging it. Jxl support even got removed from chrome.

Investment in adopting in software is next to 0.

In hardware it’s a different story, and I’m not sure to what extent which codec can be properly accelerated

Until the new codec comes to together all those small optimizations aren’t really worth much, so it’s a long term research project with potentially zero return on investement.

And yes, most of the small optimizations are patented, something that I’ve come to understand isnt’t viewed very favorably by most.

Daala was never meant to be widely adopted in its original form — its complexity alone made that unlikely. There’s a reason why all widely deployed codecs end up using similar coding tools and partitioning schemes: they’re proven, practical, and compatible with real-world hardware.

As for H.265, it’s the result of countless engineering trade-offs. I’m sure if you cherry-picked all the most experimental ideas proposed during its development, you could create a codec that far outperforms H.265 on paper. But that kind of design would never be viable in a real-world product — it wouldn’t meet the constraints of hardware, licensing, or industry adoption.

Now the following is a more general comment, not directed at you.

There’s often a dismissive attitude toward the work done in the H.26x space. You can sometimes see this even in technical meetings when someone proposes a novel but impractical idea and gets frustrated when others don’t immediately embrace it. But there’s a good reason for the conservative approach: codecs aren’t just judged by their theoretical performance; they have to be implementable, efficient, and compatible with real-world constraints. They also have to somehow make financial sense and cannot be given a way without some form of compensation.

H.264 was something like >90% of all video a few years ago and wasn't it free for streaming if the end user wasn't paying? IIRC someone also paid the fees for an open source version. There were pretty good licensing terms available and all the big players have used it extensively.

Anyway, my point was only that expecting Google to develop every piece of tech in the world and give it all away for free isn't a general model for tech development, whereas IP rights and patent pools are. The free ride ends the moment Google decide they need more profit, feel threatened in some way or get broken up by the government.

The top AI companies use very restrictive licenses.

I think it’s actually the other way around and AI industry will actually end up following the video coding industry when it comes to patents, royalties, licenses etc.

If we did that we would probably be stuck with low-bitrate 720p videos on YouTube.

Basically MBA drool material.

My expectation from experience when implementing something from a DSP paper is that the result will be unreproducable without contacting the authors for some undisclosed table of magic constants. After obtaining it, the results may match but only for the test images they reported results on. Results on anything else will be much worse.

Also it's normal for techniques from the literature to have computational/memory bandwidth costs two orders of magnitude greater than justified for even their (usually exaggerated) stated levels of performance.

And then their comparison points are almost always inevitably implemented so naively as to make the comparison useless.

It's always difficult because improvements in this domain (like many other engineering domains) are significantly about tradeoffs ... and tradeoffs are difficult to weigh in a pure research environment without the context of concrete applications. They're also difficult to weigh with implementation cleverness having such a big impact particularly since industry heavily drains academia of naturally skilled software engineers.

And as other comments have pointed out, academia is in some sense among the worst of the patent abusers. They'll often develop technology just far enough to lay patent mines around the field, but not far enough to produce something useful out of it. The risk that you spend the significant effort to turn a concept into something usable only to have some patent holder show up with a decade old patent to shake you down is a big incentive against investment.

Of course they have. Guess how MP3 was developed - an offshoot of the German Fraunhofer Institute and FAU Nürnberg-Erlangen, amongst others.

The fact that no one seems to even be able to imagine how funding anything from the government could even work (despite that era being just a few decades ago) is shocking.

[1] https://de.wikipedia.org/wiki/MP3

https://theonion.com/new-5-000-multimedia-computer-system-do...

I also remember when they went through and re-encoded all of the videos so they could play on the original model iPhone.

As long as IP law continues in the same form, the alternative to that is completely closed agreements among major companies that will push their own proprietary formats and aggressively enforce their patents.

The fair world where everyone is free to create a new thing, improve upon the frontier codecs, and get a fair reward for their efforts, is simply a fantasy without patent law reform. In the current geopolitical climate, it's very very unlikely for nations where these developments traditionally happened, such as US and western Europe, to weaken their IP laws.

The point is, if there had been no incentives to develop codecs, there would have been no MPEG. Other people would have stepped into the void and sometimes did, e.g. RealVideo, but without legal IP protection the codecs would just have been entirely undocumented and heavily obfuscated, relying on tamper-proofed ASICs much faster.

Our industry has come to take Google's enormous corporate generosity for granted, but there was zero need for it to be as helpful to open computing as it has been. It would have been just as successful with YouTube if Chrome was entirely closed source and they paid for video codec licensing, or if they developed entirely closed codecs just for their own use. In fact nearly all Google's codebase is closed source and it hasn't held them back at all.

Google did give a lot away though, and for that we should be very grateful. They not only released a ton of useful code and algorithms for free, they also inspired a culture where other companies also do that sometimes (e.g. Llama). But we should also recognize that relying on the benevolence of 2-3 idealistic billionaires with a browser fetish is a very time and place specific one-off, it's not a thing that can be demanded or generalized.

In general, R&D is costly and requires incentives. Patent pools aren't perfect, but they do work well enough to always be defining the state-of-the-art and establish global standards too (digital TV, DVDs, streaming.... all patent pool based mechanisms).

The question is more, "who would write the HTTP spec?" except instead of sending text back and forth you need experts in compression, visual perception, video formats, etc

You might be misunderstanding that almost all of Linux development is funded by the same kind of companies that fund MPEG development.

It's not "engineers in their basement", and never was

https://www.linuxfoundation.org/about/members

e.g. Red Hat, Intel, Oracle, Google, and now MICROSOFT itself (the competitive landscape changed)

This has LONG been the case, e.g. an article from 2008:

https://www.informationweek.com/it-sectors/linux-contributor...

2017 Linux Foundation Report: https://www.linuxfoundation.org/press/press-release/linux-fo...

Roughly 15,600 developers from more than 1,400 companies have contributed to the Linux kernel since the adoption of Git made detailed tracking possible

The Top 10 organizations sponsoring Linux kernel development since the last report include Intel, Red Hat, Linaro, IBM, Samsung, SUSE, Google, AMD, Renesas and Mellanox

---

curl does seem to be an outlier, but you still need to answer the question: "Who would develop video codecs?" You can't just say "Linux appeared out of thin air", because that's not what happened.

Linux has funding because it serves the interests of a large group of companies that themselves have a source of revenue.

(And to be clear, I do not think that is a bad thing! I prefer it when companies write open source software. But it does skew the design of what open source software is available.)

People don’t develop video codecs for fun like they do with software. And the reason is that it’s almost impossible to do without support from the industry.

It took a while (and a lot of pain) to get a lot of driver vendors to come fully into the project, yet Linux was already gaining a bunch of traction at that time (say last half of 90s).

I'll give you that Intel was always more or less a good actor though! But Google didn't exist when Linux already mattered. And when Google was created, they definitely benefited a lot from it, basing much of their infra on it.

Marketing needs (and laywer approval) can bring support faster than most things. Opus for audio is a good example of that too.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4693148

The entire internet, end to end, runs on FLOSS.

> This is still the argument for software copyright.

And open source licensing is based on and relies on copyright. Patents and copyright are different kinds of intellectual property protection and incentivize different things. Copyright in some sense encourages participation and collaboration because you retain ownership of your code. The way patents are used discourages participation and collaboration.

Hardware vendors don't benefit from the patent pools. They usually get nothing from them, and are burdened by having to pass per-unit licensing costs on to their customers.

It's true that designing an ASIC-friendly codec needs special considerations, and benefits from close collaboration with hardware vendors, but it's not magic. The general constraints are well-known to codec designers (in open-source too). The commercial incentives for collaboration are already there — HW vendors will profit from selling the chipsets or licensing the HW design.

The patent situation is completely broken. The commercial codecs "invent" coding features of dubious utility, mostly unnecessary tweaks on old stuff, because everyone wants to have their patent in the pool. It ends up being a political game, because the engineering goal is to make the simplest most effective codec, but the financial incentive is to approve everyone's patented add-ons regardless of whether they're worth the complexity or not.

Meanwhile everything that isn't explicitly covered by a patent needs to be proven to be 20 years old, and this limits MPEG too. Otherwise nobody can prove that there won't be any submarine patent that could be used to set up a competing patent pool and extort MPEG's customers.

So our latest-and-greatest codecs are built on 20-year-old ideas, with or without some bells and whistles added. The ASICs often don't use the bells and whistles anyway, because the extra coding features may not even be suitable for ASICs, and usually have diminishing returns (like 3x slower encode for 1% better quality/filesize ratio).

The only reason I can think of why you would say this is that nowadays we have good compiler infrastructure that works with many hardware architectures and it has become easy to create or modify compilers. But that's only due to the fact that it was so insanely complicated that it had to be redone from scratch to become generalizible, which led to LLVM and the subsequent direct and indirect benefits everywhere. That's the work of thousands of the smartest people over 30 years.

There is no way that a single company could develop a state of the art compiler without using an existing one. Intel had a good independent compiler and gave up because open source had become superior.

For what it's worth, look at the state of FPGA compilers. They are so difficult that every single one of them that exists is utter shit. I wish it were different.

I don't quite get your point though. Mine was only that it's common for single companies to develop multiple independent state of the art compilers, whereas after the 1990s video codecs tend to be collaborations between many companies. That's a piece of evidence that codecs are harder. But this is all quite subjective and I don't really care. Maybe compilers are actually harder and the trend to collaboration in video is just a cultural quirk of that subfield - doesn't really matter. The starting point of the thread was a belief that if MPEG didn't exist video codecs would have all been 100% free right from day one and I just don't see any evidence for that. The competition to MPEG in the 90s was mostly Sorensen and RealVideo if my fading memories aren't too garbled. Although the last version of Sorensen Spark was apparently a tweaked version of H.263 according to Wikipedia.