Skip to main content
Topic: AAC-LC patent expiry? (Read 15607 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

AAC-LC patent expiry?

Has anyone tried to assemble a list of AAC patents and when they expire, as several people did for MP3? Given that AAC-LC was published in 1997, isn't it likely that all AAC-LC related patents have expired and binaries for LC encoders may now be published without legal qualms?

And I suppose HE-AAC (v1) will follow in ~3 years.


Re: AAC-LC patent expiry?

Reply #2
Additionally, it depends very much on the location.
Some patents expire at different times in different countries. With MP3 the last place for the last patent to expire, was in Japan, iirc.

It expired a week after it expired in most of the rest of the world, including the US.

Apparently, calculating the term of the patents after the debacle with MP3, became much more reliable.
This page explains how term calculation works, and gives a couple examples: https://www.osnews.com/story/24954/us-patent-expiration-for-mp3-mpeg-2-h264/
This "patent calculator" I find particularly interesting: https://en.wikipedia.org/wiki/User:Jrincayc/Patent_utils

I haven't looked into by what patents AAC is covered by. If you happen to create a list of patents, it'd be rather easy to calculate their terms of expiration.
I remember someone telling me on HA, that AAC patents are actually much easier to "track" than those of MP3, due to some legal changes in the mean time (I forgot the technicalities, but it was more or less a reaction to the U-Boat patents associated with MP3 that lead to those changes. Someone on HA explained it in a very old thread).

However, all this pertains only to patents filed in the US. How things work in different countries, depends on patent laws in those countries. It is very likely, that expiration terms are different. How much this is impacting a general availability of encoders and decoders, is questionable, though. The impact of this is visible in certain Linux distributions, which have a very strict reliance on free licenses, etc. like Fedora. In fact I use the official repos of Fedora as means of tracking what has expired patents what hasn't. When the MP3 patents expired, not even an hour after midnight in Pacific time, an update for Fedora was available, with things like MP3 playback capabilities for all sorts of media players, and things like binary LAME packages became available.

On a personal note, I think no one compiled a patents list, because no one really cares. AAC is as I see it, the last "hurrah" when it comes to patented codecs like this. With the recent release of AV1 for video, and things slike Opus and Vorbis for audio before that, there's simply no need to use codecs like AAC. The last product in the AAC family, xHE-AAC is next to impossible to find in-the-wild. I remember a Norwegian Icecast instance, which listed one xHE-AAC stream (I'm not sure that Icecast web-interface was supposed to be exposed to the internet), and I believe it was more of an experiment if anything. From a developer's perspective, using libraries for AAC is incredibly difficult, when you're not developing for Windows or mobile (I don't have experience developing for consoles, like the PS4 or X-Box). On the other hand, using Opus or Vorbis, is and was plain and simple, and there was no legal baggage (or at least, not as much).
The only remaining realm of audio codecs that is still kinda securely in the grips of proprietary codecs and heavy patented ones, is telephony and ultra-low-bandwidth codecs, however things are changing there, too, with things like Codec2.

From a perspective of pure knowlege-seeking, I'd very much hope, there will be eventually all sorts of nice AAC codecs available, perhaps as part of FFmpeg, etc.

Re: AAC-LC patent expiry?

Reply #3
With the recent release of AV1 for video, and things slike Opus and Vorbis for audio before that, there's simply no need to use codecs like AAC.
AV1 is still way too slow to be practically usable by ordinary users (except google or something). Are you really using it?
H264+AAC in video is already something like JPEG in picuture or MP3 in audio, so it shouldn't be dead soon as you say.

Re: AAC-LC patent expiry?

Reply #4
Once there are enough lossy files "with no lossless original", is very hard to get rid of the format. MP3 won't just disappear. And because Apple (the music vendor) settled for AAC, it is hard to get rid of that as well. (IIRC, Apple/iTunes accepted AAC uploads from musicians? Then they don't necessarily have any "better" to replace it with?)
There will be decoding support, and that makes it a "safer" choice for decoding.

If we look at (lossy) formats/codecs that once had some momentum (sorry, Musepack fans) and that we are maybe closing in on getting rid of, then what?
Real. A stellar example of how not to do it.
MP1 on DCC, and ATRAC.
Vorbis served the purpose of being a threat. Kept MPEG-formats cheap to costless. But I have only a few hours' worth of .ogg music. And video? (Can the Theora users please raise a hand? Both of you, please.)
WMA ... I have more .wma than .ogg, and even more when we count video files. Most downloaded from careless musicians who posted on their website in what their Windows computers would default to. But despite all the hardware players that support it, I don't see much .wma anymore. Abandonware it is as well. (I wish I could blame WMA on Steve Ballmer.)
High Voltage socket-nose-avatar

Re: AAC-LC patent expiry?

Reply #5
AV1 is still way too slow to be practically usable by ordinary users (except google or something). Are you really using it?
H264+AAC in video is already something like JPEG in picuture or MP3 in audio, so it shouldn't be dead soon as you say.
Yes, AV1 is slow as hell, the reference implementation, that is. On the other hand, things like the SVT encoder are very promising, and the codec has only been out for a couple months, it's gonna take some time till fast encoders show up.

The point really is, people mostly only care about one thing: Whether it works or not. Most people are completely unaware, what sort of encoding their Youtube videos has, or what they're actually looking at, when they're watching a Twitch stream, or when they're listening to something on Spotify.

This used to be quite different not too long ago.

The awareness what format or codec is used has gone way down. This is one reason, how Opus (and to some extent, Vorbis before that, see Spotify), snuck into every day usage, without anyone noticing. In a similar way, many pictures used on websites, are actually PNGs, not JPEGs. Nobody really cares what sort of image format is used, it just has to work, and PNG is fine for almost all cases. WebM is so common these days, it has supplanted many other video container formats.

So, since people really don't care, I'm not saying MP4(H.264, AAC) (this is how I often declare formats/codecs in writing, btw.) is gonna go away anytime soon, just as MP3 is still lingering around. However it got dropped for other things without people really noticing, because they shouldn't, they don't have to care at all. For media used on mobile devices, desktops etc., things that are easily updated, the changes will be more gradual, but also in faster succession. I think it's safe to say, that MP3 almost outlived its patent status, it kinda did, but also didn't. However with AAC, I don't think this will be the case, except some legacy applications. I think we'll see AAC being dropped for something else, way before the patents are expired.

Notwithstanding, are things like DAB+ streaming, etc. Those things are locked-in as hell, of course. Once you get a DAB+ radio, you gotta use that. When they change the format and/or codec, you usually need to get a new radio, similar things happened here with DVB-T and DVB-T2, obsolescence never felt in a more rapid succession...
Those things aren't really comparable though. By those standards, audio cassette tapes aren't "dead", because many people still use them in their car stereo systems.

Re: AAC-LC patent expiry?

Reply #6
Sorry, I didn't see your post when I wrote my previous one.
Once there are enough lossy files "with no lossless original", is very hard to get rid of the format. MP3 won't just disappear. And because Apple (the music vendor) settled for AAC, it is hard to get rid of that as well. (IIRC, Apple/iTunes accepted AAC uploads from musicians? Then they don't necessarily have any "better" to replace it with?)
There will be decoding support, and that makes it a "safer" choice for decoding.
OK, a couple things to unpack here.

* Lossless Original missing.
Oh boy, that's a sad. And I'm assuming the original medium, be it CD, Vinyl, magnetic tape, etc. is missing, too?
Sure, if a lossy file is just that, the best way is to just not touch it, just keep it as it is, and that's the "original" now.

* Apples choice to use AAC.
They might as well change it at some point down the line. As I said in my previous post, users generally don't really care what sort of format or codec their data is, as long as it works reliably. Also, iTunes is kinda on the declining leg of the industry. Streaming media is just so much more convenient for the average user. So on one hand, iTunes might as well switch to a different codec and/or format, and most people simply wouldn't even notice the switch. When it comes to streaming services, this problem doesn't even manifest itself.

* Musicians.
OK, I don't know how much faster the internet has yet to become. Seriously though, when you're a musician, you'll take care your music reaches the distributor in the best format it can. If you're essentially uploading a master, you might as well upload a FLAC or ALAC along with lossily encoded versions (if a distributing service cannot transcode to a lossy target on-the-fly, I don't even know what to say in this day and age). Musicians generally care about this (contrary to most users), and if it isn't the musicians, the producers and labels do. Music creators demand that sort of functionality, sooner or later distribution services will start to comply.

* Safer choice as it's an older codec.
Yes, at least for some time, sure. Eventually it'll get obsoleted from hardware decoders, etc. Note how I was talking about a more gradual, less apparent, and in quicker succession occurring transition phase. We will still see transition phases, but for things played on smartphones, tablets, and desktop computers (or rather, decoded by those), it really doesn't matter as much. When we talk about hardware that's locked-in, like hardware decoders in stereo systems, car stereo systems, digital broadcasting, etc. that's a different issue. The consideration is also one of quality. Sure, an older codec might be a "safer" choice when it comes to compatibility, but the quality might not be ideal, either, etc.

Quote
If we look at (lossy) formats/codecs that once had some momentum (sorry, Musepack fans) and that we are maybe closing in on getting rid of, then what?
Nothing, really. If a format is crap (like HD-DVD), it'll simply get lost to time. Too bad if you adopted something early on, and gotten rid of the original medium.

Quote
Real. A stellar example of how not to do it.
MP1 on DCC, and ATRAC.
Those are locked-in technologies, similarly, how many people still use audio cassettes in their cars. Some even still use 8-Track for the same reason. The same issue arises with things like DVB-T(2), which brings pretty much the same issue to the table. Eventually you're forced to get a new device, if the firmware can't be updated in a reasonable amount of work.

Quote
Vorbis served the purpose of being a threat. Kept MPEG-formats cheap to costless. But I have only a few hours' worth of .ogg music. And video? (Can the Theora users please raise a hand? Both of you, please.)
WMA ... I have more .wma than .ogg, and even more when we count video files. Most downloaded from careless musicians who posted on their website in what their Windows computers would default to. But despite all the hardware players that support it, I don't see much .wma anymore. Abandonware it is as well. (I wish I could blame WMA on Steve Ballmer.)
That's already quite a lot and quite complex compared to the average user, using their smartphone as their primary playback device. Just the fact that you keep files is actually rather unusual to the casual user. The entire concept of keeping a private music collection is actually on the downswing. Just using an on-demand streaming service is much simpler. Sure, it eats bandwidth, but then again, that's gotten reasonably cheap, so why bother with files? Sure, some people prefer just that.
Same thing with musicians not being available through streaming services like Spotify. They might offer their music for download on bandcamp or just on their website, but it's actually smart to use something like Soundcloud similar, where people can listen to their music in its entirety without downloading (and dealing) with files.

Re: AAC-LC patent expiry?

Reply #7
Has anyone tried to assemble a list of AAC patents and when they expire, as several people did for MP3? Given that AAC-LC was published in 1997, isn't it likely that all AAC-LC related patents have expired and binaries for LC encoders may now be published without legal qualms?

And I suppose HE-AAC (v1) will follow in ~3 years.

Fedora-Legal cleared FDK-AAC for inclusion with only the LC profile enabled. So the patents in the US must have expired some time back. Not that there's much reason to use AAC with Opus being around and basically able to do everything all profiles of AAC do and more.

If anything, patent expiration is useful because it undermines companies like Apple coming out and telling everyone it's illegal to make any codec that does anything without paying them.

The problem with MPEG codecs in general is that they're made to be stuffed full of patents, and they deliberately release a base profile with missing stuff so they can add it throughout the years and string everyone along with patents. That's exactly what the relationship between Advanced Video Coding (2003) and High Efficiency Video Coding (2013) is. There's like 6,000 and some odd patents on HEVC, and more than half are just AVC patents that you have to license twice if you support both codecs in your device. Companies are getting sick of it and that's how the Alliance for Open Media happened.

Steve Jobs liked to come out and threaten Free media codecs like Theora and declared that it was illegal to compete with the iPhone (and they lost 80% of the smartphone market to Android since then), but they never actually took anyone to court, despite the "nuclear war" threats.

Amusingly enough, Microsoft sold smartphones that were way more than a crummy web browser on a small screen with no app store well before the original iPhone (which had all those problems) and nobody even noticed.

Re: AAC-LC patent expiry?

Reply #8
Fedora-Legal cleared FDK-AAC for inclusion with only the LC profile enabled. So the patents in the US must have expired some time back. Not that there's much reason to use AAC with Opus being around and basically able to do everything all profiles of AAC do and more.
Good too see LC-AAC patent-free.
While I'm agree that Opus does great however many companies use AAC. It's well positioned standard. Spotify still uses LC-AAC in its web player, Tidal's lossy streams, Apple Music and so on. https://caniuse.com/#search=AAC

IMO LC-AAC profile is the most interesting by far because there isn't much  (if any) interest  in HE-AAC with today speed connections.
 

Re: AAC-LC patent expiry?

Reply #9
IMO LC-AAC profile is the most interesting by far because there isn't much  (if any) interest  in HE-AAC with today speed connections.

Yeah, HE-AAC only becomes useful at bitrates that people are not likely to be using in practice. Parametric Stereo (in HE-AACv2) isn't useful until you are talking <48kbps and Spectral Band Replication and PNS are not recommended until you go lower than 96kbps. At bitrates higher than 96kbps, LC-AAC is the way to go because the HE stuff is more likely to degrade audio quality at that point instead of helping.

64 kbps has proven itself to be a bridge too far for any codec to sound like the CD. Microsoft was claiming WMA could do it 20 years ago and there's still no codec that can. Not even Opus. I can't imagine who would be using <128kbps with anything for "storage" though. Those are already some pretty darned small files.

 

Re: AAC-LC patent expiry?

Reply #10
Fedora-Legal cleared FDK-AAC for inclusion with only the LC profile enabled. So the patents in the US must have expired some time back.
The LC-Patents expired in 2017. Red Hat (staffing the Fedora legal team) declared FDK-AAC is able to be included into the main repos in October of 2017 (it took them a couple months to ensure no infringement is possible I assume). Three months after Fedora 26 got released.

Not that there's much reason to use AAC with Opus being around and basically able to do everything all profiles of AAC do and more.
That's kindof a naïve take on it. AAC is the standard in things like DAB and DVB. When implementing algorithms on hardware, systems become locked-in while they're developed. There is simply no other way, than use whatever has been standardized, or wait for the next standard (which isn't a very efficient way to go about these things like that).

If anything, patent expiration is useful because it undermines companies like Apple coming out and telling everyone it's illegal to make any codec that does anything without paying them.
That's a very course and simplistic way of putting it. Patents as such are a form of publication, however there are legal protection claims attached to it. However, details of patenting work different in different countries. Certain design templates have essentially an indefinite expiration date, since there are further legal protection attached to them (famously for certain logograms and architectural shapes, etc.).

The problem with MPEG codecs in general is that they're made to be stuffed full of patents, and they deliberately release a base profile with missing stuff so they can add it throughout the years and string everyone along with patents. That's exactly what the relationship between Advanced Video Coding (2003) and High Efficiency Video Coding (2013) is. There's like 6,000 and some odd patents on HEVC, and more than half are just AVC patents that you have to license twice if you support both codecs in your device. Companies are getting sick of it and that's how the Alliance for Open Media happened.
That's not really how AOM and things like AV1 happened. The main reason wasn't exactly companies getting sick of patenting issues, it's that the business model as driven along by MPEG imploded (famously, the founder of MPEG has written about this: http://blog.chiariglione.org/a-crisis-the-causes-and-a-solution/). As people like David Rowe have written about this in the past, the actual investment to improve a codec is actually relatively small. This might sound like it'd be very profitable, but it actually isn't anymore. The issue is, that it's increasingly cost un-effective to invest in codec development and then collect royalties. In many ways, AOM is very similar to a patent pool. Except this time, it's relatively open and with no actual patenting attached - quite similar to how Microsoft is now a member of the Free Software Foundation. In other words, we entered a time, in which not making standards like this restricted is more cost effective, than working and releasing an open standard (in this case, it's a different story in other realms entirely. Two-way radio communication like TETRA is still thriving on those kinda things, however there's Codec2).

When it comes to publishing "unfinished" codecs, well, that's a bit of a stretch. The debacle with MP3-U-Boat-Patents have been widely solved. One aspect that has to be considered, is the time it takes to actually make a patent issue "water proof" and wait for the patent grant to be accepted, etc.

MPEG-LC was granted in 1997, MPEG-4 AAC-LC (AAC-LC + PNS) in 1999, HE-AAC (AAC-LC + PNS + SBR) in 2003, HE-AACv2 (AAC-LC + PNS + SBR + PS) in 2004. One reason a "simpler" version of a format (simpler profile) is to be first to the game. I wouldn't call them "unfinished" but they go from one use-case to the other. Does this improve royalty collection" Sure. But it's also not [i[all[/i] greed...

Speaking of which, the latest in the "bunch" is xHE-AAC, which has a very limited use-case in narrow-band broadcasting (it's all the aforementioned profiles, plus USAC). Since 2016, xHE-AAC can be licensed, but I'm not sure when the USAC patents were filed (I vaguely remember 2009-2012, but I might be incorrect). Other profiles have been released in the mean time (SLS, LD, etc.) however, those get almost no attention by anyone these days.

Steve Jobs liked to come out and threaten Free media codecs like Theora and declared that it was illegal to compete with the iPhone (and they lost 80% of the smartphone market to Android since then), but they never actually took anyone to court, despite the "nuclear war" threats.

Amusingly enough, Microsoft sold smartphones that were way more than a crummy web browser on a small screen with no app store well before the original iPhone (which had all those problems) and nobody even noticed.
Well, this is kinda orthogonal to the prior discussion, but Windows CE based PDAs (which could also make calls) were quite popular as company phones here, where calendars, appointments, etc. were synced to the employee phones/PDAs. This was kinda expensive for the regular consumer, but companies were quite happy to foot the bill for their employees, as Internet through GPRS was quite expensive.

In case you're talking about phone speech compression, things are even more let's say complicated. Most G.7xx codecs are still royalty bound.

Codecs like the ones you could license through MPEG-LA, was a cost effective idea back in the day, but the world and the internet have simply caught up with it, and it now simply doesn't make sense. However, people (and companies) will keep improving on their work. MPEG-H is one thing, and  I'm pretty sure, there'll be a successor to Opus eventually.

In general, having opinionated political discussions like this, isn't really getting us anywhere.

Oh, perhaps one thing: Remember when MPEG didn't charge for MP3, but then they just kinda started out of the blue? Now that was a true dick-move.

 
SimplePortal 1.0.0 RC1 © 2008-2019