Slashdot Mirror

...or at least I do hope it should be obvious (but unfortunately it is not true for everyone, I don't see any other comments talking about it yet), whichever software and operating system you choose, please consider using Ogg Vorbis format to store your music.

Many people still keep using proprietary MP3 file format, which is unfortunate not because of its lower quality to size ratio, which is hard to hear for an average person, who doesn't even usually hear any difference between 192kbps and 256kbps constant bitrate MP3 files, but because of its legal issues. While Vorbis is technically similar to JPEG graphics format, i.e. it's a "lossy" compression, it is legally similar to PNG, while MP3 is in that analogy similar to GIF, using LZW compression patented by Unisys. See the MP3/MP3PRO Patent and Software Licensing Information website and search Google for "MP3 patent" to find more informations about this issue.

Also, I hope, and I'm sure most of the people here will agree with me, there will be a little "(O)" mark, next to your copyright statement! Good luck! We'll be looking for your links in the Open Music Registry!

(And please, don't post this old stupid joke that we should use double OAL logo, "because (o)(o) looks better," because it doesn't. It looks like an immature joke made by a 15-year-old child, while EFF's OAL should be taken seriously if we ever want it to successfully compete with the recording industry at large.)

[BMG owns] RCA Records

Strange. BMG, a major record label, owns RCA Records. Thomson Multimedia owns the rest of the RCA brand, and Thomson is also the exclusive USA sublicensor of the MP3 patents. Does that point toward a new method of fighting "Music Piracy 3" (the first two were player-pianos and tape decks)?

Well, [decoding Vorbis on DSP chips] is already taken care of with the release of the BSD-licensed "Tremor" integer decoder.

Three reasons why it may not help:

1. Some players decode MP3 audio with an ASIC that isn't LBA-complete[1]; they take MP3 on one pin and produce WAV on the other, and they cannot be reconfigured for any other audio format.

2. Though the iPod player, uses a pair of ARM processors for decoding the audio and running the menus, and those ARM processors can be upgraded in firmware, the flash chip may not have enough storage to hold both the MP3 decoder and the Ogg decoder.

3. What if the player maker got a sweeter unit royalty deal with RCA, the U.S. sublicensor of the MP3 patent, for pledging to keep the device MP3-only?

[1] "LBA-complete" denotes a machine that can run any algorithm that fits into RAM, that is, a general purpose computing device. It's a weaker form of Turing-completeness which cannot be achieved because it requires infinite storage; a Linear Bounded Automaton restricts the available memory to a multiple of the size of the input.

The licensing fee DOES NOT apply to software decoders, only hardware decoders.

You're getting your information from a PR person. I'm getting mine from the licensing page. I see no such exception for free decoders.

If you check out the list of licensees at mp3licensing.com, you'll find Microsoft right there. And if you look at the royalty rates, you'll find out how much they paid for a one-time paid-up software decoder license.

If you check out the list of licensees at mp3licensing.com, you'll find Microsoft right there. And if you look at the royalty rates, you'll find out how much they paid for a one-time paid-up software decoder license.

This story has obviously been overhyped, as it will not mean the end to free (as in free beer) mp3 players (and some encoders, most likely). It will, though, mean the end to mp3 encoders/decoders in GPL'd software, in other words the vast majority of open source software. It will also mean no default mp3 support in your favorite GNU/Linux distro, most likely.

This is unfortunate for the open source community, but it will help to promote newer open standards with no patent or licensing problems to worry about. It will most likely not affect the general public, who will still be able to download winamp for free.

It suprises me that nobody in the open source community noticed this discrepancy before, as it has been listed on the licensing page for quite some time. As the Thomson lawyer said, it has been incompatible with the GPL for some time now. I suppose from now on people will be more careful before building patented proprietary technology into open source software, despite the fact that this type of patent makes no sense and hurts innovation by creating a closed standard.

Gee 0soft already paid these dumb license fees, have a look:
http://www.mp3licensing.com/licensees/index.asp

So here's the solution: write your code, set up a distribution site in India and lean back. When that's too far, try Poland or Mexico.

Don't forget to register a company or other legal entity in India, since you may be sued in your own country where there's still a patent. Fraunhofer and ST may not win, but you'll sure be broke looking at the costs of litigation.

It's basically the same trick as KaZaa:
When they were sued in The Netherlands, they moved the whole stuff to Australia (however, in appeal, KaZaa won).

IANAL (yet)

Can anyone clarify?

Let them come to me to cough-out 75 for my license.

If they win, they get not 75 cents but $15,000 in actual damages, as that's the annual minimum royalty. From the software royalty rate page:

Annual minimum royalties are payable upon signature and each following year in January and are fully creditable against annual royalties: US$ 15 000.00 per calendar year

Browsing around the licensing website for a bit it appears as though a number of companies (Nullsoft -- makers of winamp) are already licensed, although it makes no mention of whether or not they will retain their license under the new scheme. Although Thomson does not sell the mp3 libraries for non windows/mac operating systems they do offer their SDK for linux and other operating systems (for a price). It is also interesting that winamp has plans to come out with a version for linux soon. Although many windows mp3 players are licensed, I would guess that no open source players/encoders are. Therefore those who will get shafted will be linux users who rely upon these. I sincerely doubt win users will have any obstacles.

Browsing around the licensing website for a bit it appears as though a number of companies (Nullsoft -- makers of winamp) are already licensed, although it makes no mention of whether or not they will retain their license under the new scheme. Although Thomson does not sell the mp3 libraries for non windows/mac operating systems they do offer their SDK for linux and other operating systems (for a price). It is also interesting that winamp has plans to come out with a version for linux soon. Although many windows mp3 players are licensed, I would guess that no open source players/encoders are. Therefore those who will get shafted will be linux users who rely upon these. I sincerely doubt win users will have any obstacles.

you want to cough up the $0.75?

I can do that for the first copy.

You can pony up the Annual Minimum Royalties of "US $15,000.00 per calendar year".
http://www.mp3licensing.com/royalty/software.html

Q. Do I need a license to stream mp3/mp3PRO encoded content over the Internet?

Yes. A license is needed for commercial (i.e., revenue-generating) use of mp3 / mp3PRO in real time broadcasting (terrestrial, satellite, cable and/or any other media), broadcasting / streaming via Internet, intranets and/or other networks or in other electronic content distribution systems, such as pay-audio or audio-on-demand applications.

However, no license is needed for private, non-commercial activities (e.g., home-entertainment, receiving broadcasts and creating a personal music library), not generating revenue or other consideration of any kind or for entities with an annual gross revenue less than US$ 100 000.00.(emphasis mine)

MP3 only came up because it was available at low-to-no-cost. Regarding some of the patents, of course. Nobody would've had used it if they had charged this decoder fee from the very beginning, and they know!

Do what I am going to do: Write a letter (paper!) to Fraunhofer and Thomson and explain your concerns.

Yes, I know about Ogg Vorbis and stuff, but there's no reason not to protest against changed mp3 licenses.

I don't want to re-compress all my mp3s to Ogg because this will reduce quality. So I will still have mp3s around in several years (don't mention all those CDs I burned). So this is an issue, since I will need a player/decoder to access them.

Contact Fraunhofer:

(Interesting: On the English homepage, their postal address doesn't show up - only eMail addresses. On the German homepage, it does.)

Contact Thomson:

Do you license mp3/mp3PRO software to end users?
No. We license mp3/mp3PRO software and patents to developers and manufacturers of software applications and hardware devices.

Note: This license does not cover the right to distribute, broadcast and/or stream mp3 / mp3PRO encoded data. These rights are covered by the licenses described under Electronic Music Distribution / Broadcasting / Streaming.

Note: No license is needed for private, non-commercial activities (e.g., home-entertainment, receiving broadcasts and creating a personal music library), not generating revenue or other consideration of any kind or for entities with an annual gross revenue less than US$ 100 000.00.

Do you license mp3/mp3PRO software to end users?
No. We license mp3/mp3PRO software and patents to developers and manufacturers of software applications and hardware devices.

Note: This license does not cover the right to distribute, broadcast and/or stream mp3 / mp3PRO encoded data. These rights are covered by the licenses described under Electronic Music Distribution / Broadcasting / Streaming.

Note: No license is needed for private, non-commercial activities (e.g., home-entertainment, receiving broadcasts and creating a personal music library), not generating revenue or other consideration of any kind or for entities with an annual gross revenue less than US$ 100 000.00.

Do you license mp3/mp3PRO software to end users?
No. We license mp3/mp3PRO software and patents to developers and manufacturers of software applications and hardware devices.

Note: This license does not cover the right to distribute, broadcast and/or stream mp3 / mp3PRO encoded data. These rights are covered by the licenses described under Electronic Music Distribution / Broadcasting / Streaming.

Note: No license is needed for private, non-commercial activities (e.g., home-entertainment, receiving broadcasts and creating a personal music library), not generating revenue or other consideration of any kind or for entities with an annual gross revenue less than US$ 100 000.00.

I'm still not convinced that these guys are coming down on open source players, but here's some concrete items worth mentioning:

If you use a COMMERCIAL MP3 decoder, you're OK. Nullsoft, Apple, Microsoft, Real, Musicmatch, and probably any other manufacturer ou can think of has a license. It's all listed here.

Whether or not the freebies will be forced into licensing is another question. Yes, the clause has been removed from their page that freebie players don't adhere.

Is Ogg completely open? Is there ANYONE who can claim a patent on it 10 years from now (see JPEG)? If not, what are we waiting for?? Ogg rules!

It made me laugh anyways...

Note: No license is needed for private, non-commercial activities (e.g., home-entertainment, receiving broadcasts and creating a personal music library), not generating revenue or other consideration of any kind or for entities with an annual gross revenue less than US$ 100 000.00.

Yes, NullSoft is a licensed company.

And to clarify, end users are not responsible for the payment of any licensing fees for software that uses the mp3 patents.

Bammkkkk

Yes, NullSoft is a licensed company.

And to clarify, end users are not responsible for the payment of any licensing fees for software that uses the mp3 patents.

Bammkkkk

Go here for a list of companies w/ licenses. AOL is on that list, as is Napster and NullSoft

Loads of companies have already paid for a licence. Look here.

Annual minimum royalties are payable upon signature and each following year in January and are fully creditable against annual royalties.
US$ 15 000.00 per calendar year.

Now that's a pain. I emailed them to see if I could get a "hobbyist license" for more per app, but without the $15k minimum (wanted to make "iTunes 3 for Classic Mac OS"). They allow you to release up to 5000 units of a game that uses mp3s royalty free, so I was hopeful. The reply? No dice. (I was impressed they sent a reply!)

Fwiw, here's a list of the licensees.

I imagine they use the same technology, if my memory serves me, the copyright notice says its licenced in the PS2 version. I don't think it would be economical to use a differant compression technology for the same game on a differant platform.

Yes it would. If Sony is willing to license its ATRAC3 codec (used in MiniDisc LP decks) to PS2 game developers cheaper than Thomson Multimedia is willing to license MP3 technology for a video game, but Sony is not willing to license ATRAC3 for use in PC software, then you bet developers will re-encode their wav files into ATRAC3 for the PS2 version and MP3 for the PC version.

In addition, platform limitations may come into play. The NES supported only one codec for samples, a simple predictive codec, at a few specific bitrates. The GBA is more flexible (with ability to play PCM from RAM; thus, ability to decompress in real time), but it still has only a 16 MHz processor. I've read that ATRAC and ATRAC3 are less complex to decode than MP3, which could conceivably free the PS2's vector units to process more triangles instead of audio samples.

hardware companies ... don't have to pay to implement MP3

THOMSON multimedia, the sublicensor of the MP3 patents in the United States, charges a royalty for decoders, at $15,000 for the first 20,000 annual units and 75c/unit thereafter.

The latter option is infeasible until they get their act together and put out a specification.

Wait a few days for the Ogg Vorbis 1.0 release to be finished, and you'll apparently be able to download the specification as part of the libvorbis manual.

hardware companies ... don't have to pay to implement MP3

THOMSON multimedia, the sublicensor of the MP3 patents in the United States, charges a royalty for decoders, at $15,000 for the first 20,000 annual units and 75c/unit thereafter.

The latter option is infeasible until they get their act together and put out a specification.

Wait a few days for the Ogg Vorbis 1.0 release to be finished, and you'll apparently be able to download the specification as part of the libvorbis manual.

This ambush of the digitial imaging industry will probably stand as the worst public relations nightmare a company can inflict upon itself.

Rather like the Unisys LZW (GIF) patent fiasco from, when was it, 1994? Or the more recent Fraunhofer MP3-compression patent enforcement?

Either the company will cave in, or the software developers will, and we'll all move on to a different JPEG compression technology. Same as before.

Because the author of LAME has'nt licensed the right to use the Fraunhofer MP3 Patent. If the commercial software companies license the patent, they're allowed to use any encoder they want in their products, including LAME. OGG Vorbis is'nt covered by this patent. Details about MP3 patent licensing are availalble here.

I had owned and immediately given away the original, incredibly flawed, RCA Lyra because it simply stank so much I couldn't stand to own the thing. If RCA wants to send one of the new ones, I'd be pleased to check it out, but the shoddy original made me vow never to give RCA more MP3 player money.

RCA already has your MP3 player money, and there's little you can do about it. RCA's parent Thomson Multimedia administers the patent rights for MP3 technology and charges royalties to all manufacturers of hardware MP3 players.

As cool as the iPod is, why can't it record?

• Encoding takes much more processing power than decoding.
• Writing to flash memory is quite a bit slower than reading because it takes a while to erase a block.
• Just the patent royalties for hardware MP3 encoding run upwards of $2.50 per unit.

AFAIK, no one can make the format ... expensive to use

Not even Thomson?

but we will still have open source alternatives (such as LAME) for encoding.

Which are illegal to use in the United States and other countries where the algorithms necessary to encode MP3 are patented. Fraunhofer's patents cover more than just its codec and ISO's codec.

From their legal page, it seems that their encoding software apparently uses LAME. Isn't that kind of sad?

Thomson Multimedia (parent company of RCA and GE electronics division) controls patents on MP3 technology and charges a royalty of 2 percent of revenue on electronic music distribution. Thomson must be making a wad of dough on this deal.

This is why I copy all my CDs (legally) to my hard drive in MP3

Which program do you use to encode them? Has the publisher of the program paid the Thomson royalties for use of MP3 technology? This is why I copy all my CDs (legally) to my hard drive in Ogg format.

and use WinAmp across the board

It's spelled Winamp (small a), and the "Win" has nothing to do with the name of any Microsoft operating system product, just as the name "Windows" has nothing to do with DOS.

Are there more than 1000 people worldwide who care about the Ogg Vorbis format?

I'd bet there are at least that many on the developers' mailing list, let alone users who have downloaded oggdrop. [/me goes to vorbis.com to pick up oggdrop RC3]

as long as the authors insist on retaining that stupid name for it

How is .wma any less stupid than .ogg? At least the latter is less clumsy to pronounce.

MP3Pro

The free encoder runs at only 64 kbps (nearly equivalent to 112 kbps MP3, fine for flash-based portables but not for anything that uses a disc), the better encoder is $7.50 per unit straight to RCA in addition to the OEM and retail markup, the decoder is $1.25 per unit ("I can't play this in Winamp?"), and all encoders cut out at 15 kHz and have trouble with wideband signals such as cymbals and distorted guitars.

There are no restrictions on MP3 decoders the way there are on encoders

Not anymore. The royalty-free license for free(beer) software mp3 decoders appears to have disappeared from Thomson's royalty schedule. Winamp and XMMS as free products seem no longer legal.

Even then, hardware decoders have never had a royalty-free license.

• Decoder: 75 cents per unit or $50 thousand one off
• Encoder: 2 dollars 50 cents per unit

This does not include the right to stream that content, for which you have to pay more. In contrast, to decode, encode, stream, store Vorbis you have to pay $0.

Getting away from licensing issues, a recent listening test concluded that at 128kbps Vorbis RC2 was right up there with LAME encoded MP3, and better than Xing encoded MP3 -- and RC2 still has a couple of minor issues that will be fixed before the release.

(okay, what the FUCK kind of lameness filter reason is 'Junk character post'? I had to get rid of some dollar signs to get it to post)

Not true. The main hold up is that a v.1x reference encoder/decoder hasn't been released yet. While the file format is stable, and has been forward/backward compatable for some time, some final features are still being added. For commercial use, adding the support in too early would cause customers to want to upgrade...too much of a hassle for most companies when v.1.0 is so much more marketable.

See the Ogg Vorbis FAQ for more details on pending hardware support for Ogg Vorbis. On the Rio device front, there are hints at Ogg support in the near future (nothing formal yet ^ ).

Another reason why Ogg Vorbis support is likely is that MP3 and MP3Pro licencing costs are fairly high. For devices where MP3 support isn't even used -- say small digital players used by joggers -- Ogg is quite interesting.

^ - Rio Volt firmware and Ogg Vorbis

Ummm, what exactly is ILLEGAL about MP3 encoders and CD-recording software that we can use now? The only main difference is that the XP-included utilities are likely of lower quality than ones you get for free elsewhere.

the music industry for producing the music being encoded into MP3s without their license. They should definitely go for per track damages.

so, does this make any mp3 generating software liable for "viral" infringment?

No, but if you don't turn on the "ogg" option in abcde, you commit patent infringement. In the United States of America and several other countries, inventors can patent algorithms by patenting any device running the algorithm, and Fraunhofer owns patents on MP3 in several jurisdictions.

My game freepuzzlearena contains some "viral infringement" of its own, as it includes a clone of the patented game "Dr. Mario" whose object is to remove viruses from a bottle.

no one is really into ... encrypting their band's music (since being noticed is what they *do* want)

If an independent band inserts .ogg files into Freenet and then publishes the KSK (or whatever Freenet retrieval keys are called) on its web site, bam! Instant legitimate use, and no more need for the Fraunhofer patent license required to get your work onto MP3.com. (LAME doesn't work for bands in countries that recognize Fraunhofer's patent on "coding an audio stream by doing a spectral transform on each block then allocating bits per frequency based on fixed sub-bands".)

MP3 patents: http://www.mp3licensing.com/patents/index.html Royalty rates: http://www.mp3licensing.com/royalty/index.html

MP3 patents: http://www.mp3licensing.com/patents/index.html Royalty rates: http://www.mp3licensing.com/royalty/index.html

Does anyone know what patents are used by MP3

When you license MP3 encoding rights for $2.50/unit, you get these patents such as 5,742,735, which covers the process of transforming audio into a spectrum, quantizing the amplitude by frequency band so as to minimize audible noise, and allocating any leftover bits to the most powerful frequency bands. (Skillfully engineered VBR may avoid infringement by skipping step 3 altogether.)

and when they expire

Patents expire 20 years after they're filed or 17 years after they're granted, depending on various factors. Because it typically takes three years to get a patent granted, it really doesn't matter. This patent, filed in August 25, 1994, will expire in 2014 or so.

Have you ever heard about all the patents that exist around MP3 recording? Playback has licensing problems itself (up to $100,000 for a one time payment), and for encoding MSFT would have to pay $5 per copy of WMP8 for full MP3 encoding capabilities.

For the amount of copies Microsoft will sell of XP and will give away of WMP8, that means they're paying $50m (at least) in patent fees to Thomson Multimedia, and then there's all the fees to include Fraunhofer's own encoding engine. They're not going to do that.

The limited MP3 encoder has been removed from the latest build of XP and a plugin will become available. It will not be limited - any idea that WMP8 will automatically limit MP3 bitrate is complete FUD.

This patent encumberment is why Ogg Vorbis exists, of course...