As most people are well aware, a variety of copyright and trademark disputes have prevented the online distribution of the most valuable catalogue of musical recordings known to humanity — viz., the songs of the Beatles, whose value in the online marketplace surely exceeds a billion dollars. So imagine everyone’s surprise when the online seller Bluebeat.com began selling remastered Beatles tracks for $0.25 each a couple of weeks ago. Four for a buck!! Christmas in July (er — October, anyway)!
Alas, it was not to be — or at least not to be for too long. Bluebeat.com had not, in fact, beaten the industry giants like iTunes and Amazon.com to the rights to the Beatles’ songs. Instead, it claimed the right to distribute the songs without any permission from the copyright holders because the songs had been re-recorded using the technique of “psycho-acoustic simulation,” described by Hank Risan, head of Media Research Technologies (owner of the Bluebeat site) as ““my synthetic creation of that series of sounds which best expresses the way I believe a particular melody should be heard as a live performance.” Risan described the technique to the LA Times this way:
“Make a single copy. Analyze it. Destroy it. Create a new simulation based on parametrics of sound. You’ll be shocked at how the brain, in terms of its perceptual coding in the central nervous system, turns these sounds into electrical impulses, which then affect your mood, your cognizance, etc. Pyschoacoustic simulation exploits aspects of perception that are not present in the original work. It’s an art. The first simulations we did were awful. It’s an art. It’s not a copying at all.”
Uh-uh. Nice try. The copyright holders were not amused — EMI and Capitol Records, among others, rushed into federal court and, on November 9, obtained a temporary restraining order against Bluebeat forbidding any further distribution oof the Beatles’ songs. The problem for Bluebeat is not simply that “psycho-acoustic simulation” sounds like a crock of nonsense — people who heard the recordings being distributed claim that they sound exactly like the Beatles’ originals. The problem is that even if it’s not a crock of nonsense, there’s still an obvious copyright infringement in what Bluebeat is doing. Each of the Beatles’ recordings, like virtually all recorded music, contains two separate copyrighted works: the “musical work” (i.e., covering the song itself, music and lyrics) and the “sound recording” (covering the actual sounds placed on tape by the performers). Now, even assuming that psycho-acoustic simulation is a real technique that creates new sounds expressing the underlying musical work, that would just mean that distribution of the recordings wouldn’t infringe the sound recording copyrights; you’re allowed, under US law, to re-create the sounds of a recording and avoid infringement liability to the owner of the sound recording copyright. But there’s still the musical work copyright to consider, and there’s no argument on the planet that is going to help Bluebeat avoid infringement liability on that score.
Too bad, I guess — along with maybe 200 million or so others, I’m still waiting for those classic tunes to make it to iTunes or the equivalent. Guess I’ll just have to find my old LPs and start ripping . . .
[Thanks to Ed Butkovitz and Jerry Lewis for the pointer]
mg says:
If I could use my laptop in your class this would definitely be the way that I wasted time.
Quote
November 13, 2009, 10:25 ammph says:
“Yuri Gagarin” cosmonaut
returns 1,680,000 results.
Quote
November 13, 2009, 10:41 amkrs says:
meh. nauseatingly overrated
Quote
November 13, 2009, 2:05 pmDOuglas2 says:
I have music-recording students attempt “sound alike” recordings. To do so, they must either find a score (with its own issues of performance and recording rights) or create one themselves by analysis of the original recording. Once they work out the electric guitar part, they must also analyze and surmise enough about the performance to duplicate the effects and timbre of the original. A vocalist with a voice timbre appropriately similar to the original needs to be found. The exercise is a fantastic skill-builder, not a commercial product. Nevertheless, many hold that the resulting recording could be released with the agreement of the music (composition) publisher, and no agreement of the original record label or artists would be needed except where they hold publishing rights in addition to their rights to the (original) performance and recording. We certainly could not represent it as being a preformance by the original artist, however.
Say an artist released a Wendy-Carlos style or Mannheim-Steamroller style instrumental work composed entirely of “sequenced” instruments. By listening to the recording, one could analyze the notes and the timbres and reverse engineer it so that with similar synthesizers the “patches” that control the timbre and the sequencing-computer instructions that control “what note is played when” are duplicated exactly. A resulting recording could in theory be indistinguishable from the original work, but would it be a separate new performance?
Old low-fidelity recordings of piano works can be electronically analyzed to create a set of instructions (or an “electronic piano-roll”) for a modern player-piano. With appropriate artistry in the tweaks, the resulting “re-recording” can sound nearly exactly like the originally recorded performance, except without the distortions and noise of the old cylinder recordings. Is it a new recording? It is certainly made by a computer playing a piano rather than by the long-dead artist.
I figure that Hank Risan/Bluebeat are either using (or claim to be using) some implementation of “structured audio” a-la MPEG-4
(http://users.ece.utexas.edu/~bevans/courses/ee382c/projects/spring04/MihirAnandpara/LitSurveyReport.pdf is a decent overview)
The “analysis” identifies the pitches associated with at particular element of the auditory scene, and identifies a model that could be used to resynthesize a similar sound. From that point parameters are derived that are sufficient to recreate those pitches with that timbre. Additional layers can reverberate particular elements of the scene etc. The vocal parts of a song are just synthesis of appropriate fundamental frequency, overtones, and formants, using a software model of the sufficient parameters of the human vocal tract to recreate a close-enough timbre upon “decoding”.
Undoubtedly some people think that once you have thrown away every part of the original and recreated it “from scratch” then it is not really the same work anymore. Others will argue that it is the same work — if your screen print is a low-resolution copy of a press photo, for example.
I’m with the record company on this, by the way. But belittling the technical process as something akin to a “proprietery MP3 recorder” could possibly be off the mark. The singers on those bluebeat MP3’s could be the brothers of the voice on your GPS, rather than John and Paul.
Quote
November 13, 2009, 4:14 pmM. Gross says:
So, let’s assume that BlueBeat.com was going into bankruptcy before this whole fiasco. What exactly are the drawbacks of their actions? I’m sure the copyright holders will have limited success recovering damages.
Quote
November 13, 2009, 5:39 pmLeo Marvin says:
David Post,
What happened to the compulsory mechanical license?
Quote
November 13, 2009, 6:05 pmfooburger says:
Just FYI, ‘psychoacoustics’ means “MP3 encoded”.
In MP3 encoding, a Fourier spectrum of the signal is obtained. That spectrum is analyzed based on a ‘psychoacoustic model’, which basically drops frequencies components of low amplitude which are very nearby high amplitude frequency components.
It is basically thinning out the frequency spectrum, allowing the signal to be represented by a smaller number of numbers (the remaining frequencies).
So... no wonder these samples sounded exactly like the original songs. They pretty much were.
Quote
November 13, 2009, 6:38 pmDavid Nieporent says:
Personal liability.
Quote
November 13, 2009, 6:52 pmDOuglas2 says:
Psychoacoustic encoding can mean “MP3 encoded”, but in my boring long winded comment above I tried to explain that it can mean more than that. After all, why send digital audio to the “virtual reality” console if it takes less data-rate or data-storage to just send instructions to the synthesizer in the console to re-create a close-enough representation of the sounds that you need.
http://www.chiariglione.org/mpeg/faq/mp4-aud/mp4-aud.htm#41
Under copyright law they have no right to “duplicate the sound recording in the form of copies that directly or indirectly recapture the actual sounds fixed in the recording...” My suspicion is that they are trying to leverage the phrase “actual sounds”, and say that while it may “sound like a duck” it is not really a recording of a duck or a copy of a recording of a duck, but instead a recording of a duck-call.
Quote
November 13, 2009, 7:44 pmBruce Hayden says:
The theory seems to be that they can exempt themselves by copying the music to a score, and then preforming from the score. Or at least preforming from the score. BUT, they are consciously trying to mimic the sound of the original.
Which, sounds like an attempt to hide the fact that what they are doing is copying the original performance. And, worse, by having someone listen to the original, they have made the job of proving copying that much easier by admitting to access.
No, this isn’t going to get around copyright. It is reproduction, and if not that, at least the creation of a derivative work.
Better luck next time.
Quote
November 13, 2009, 10:34 pmDOuglas2 says:
For my educational purposes they are trying to mimic the sound of the original in the same way that the art students in the museum are trying to mimic the appearance of the old master. Commercial release is not envisioned. Sure does build skill and understanding, ‘tho.
Copyright exists in the musical composition, for which anyone can get a compulsory license to release a recording.
Copyright also exists in the printed score of musical parts, if such exists.
Copyright also exists in the lyrics.
Copyright exists for the phonographic recording. It is the copyright in the phonographic recording that is at issue in this case.
17 U.S.C. § 114(b) states that
Now, IANAL, but it looks to me like this is saying my students and I are in the clear for our attempts to imitate the copyrighted recording, even if I admit to access of the original.
Quote
November 13, 2009, 11:23 pmfuu says:
intellectual property is not real property. get real.
Quote
November 13, 2009, 11:34 pmFub says:
For some long dead pianists and composers, an even simpler “re-recording” method is already long in use, with remarkable results.
Thousands of performances for which composition and performance copyrights have expired were recorded on player piano rolls. Some piano roll recording methods (notably those of Welte and Ampico, but others as well) captured not only the notes and durations, but also the note velocities, expression pedal actions, and more. When played back properly, these rolls yield music that is as faithful to the original performance as any modern recording or recreation of a performance. Piano rolls were, quite literally, the first digital recordings of musical performances.
Some years ago, Mr. Terry Smythe, led an effort in collaboration with the International Association of Mechanical Music Preservationists and others, to collect, scan and transform ancient piano rolls into MIDI representation, for reproduction on modern computers. Some tens of thousands of results from this effort can be found here.
From that collection, one can enjoy piano performances in exquisitely subtle clarity and detail, by the some of the greatest pianists and composers of the early 20th century.
Quote
November 14, 2009, 1:48 pmShane says:
I’m pretty sure fooburger is right. The “psychoacoustic” process described sounds a lot like the technical details on how ordinary lossy audio codecs (like MP3) work.
With this line of reasoning, I could probably justify any kind of copying of almost any work. It’s amazing how similar a PNG file can look to a JPG file although the UNDERLYING DIGITAL BITS ARE COMPLETELY DIFFERENT! It’d even work with text — run the original through some kind of upside down converter like this one, destroy the original (apparently this is key), and then print the new file out. And then flip your page upside down.
Example:
˙ʇuǝɯǝƃuıɹɟuı ʇou sı sıɥ⊥
Quote
November 14, 2009, 2:35 pmBruce Hayden says:
But of course, it would be infringement if the original were still under copyright. Running something through a filter is either reproduction and/or creation of a derivative work. Copying in one guise or another.
Quote
November 14, 2009, 4:36 pmpunditius says:
The real question is, where is the market? The Invisible Hand ought to be resolving this issue, given how much money would be generated by internet availability of the Beatles’ catalog. Shouldn’t it?
Quote
November 14, 2009, 6:51 pmMartha says:
The Invisible Hand has been giving us The Finger, I think.
Quote
November 14, 2009, 9:09 pmBrett Thomas says:
But, the underlying song copyright comes with a compulsory license — http://www.law.cornell.edu/uscode/17/115.html
I certainly think it’s a reasonable question whether they could afford to pay it at the $.25/song price point, and obviously their rationale on the sound recording is bunk. You say, though, “There’s no argument on the planet that is going to help Bluebeat avoid infringement liability on [underlying song].” I think the argument “we were just going to pay the compulsory license fee” (which is generally remitted quarterly as I recall) is a pretty compelling one.
Quote
November 14, 2009, 10:23 pmHerb says:
I think you overestimate the value of the digital Beatles catalog by...oh...a lot. A billion dollars? Oh, I know, I know. They’re proven cash cows in the music biz, but is that due to their enduring popularity...or the various media format changes over the years? The Beatles fans of yesterday bought the LPs, the 8-tracks, the cassettes, the CDs. The Beatles fans of tomorrow? They’ll just buy the catalog once. (That is, if it’s not crippled by DRM.) Or, with “music on demand” type services, they won’t buy it at all.
I look at the Beatles catalog, a comforting stand-by unchanged since the 70s, not as a cash cow, but as a diminishing asset. It was great...while it lasted.
Quote
November 15, 2009, 4:06 amM. Gross says:
Shouldn’t they be shielded from that by incorporation?
Quote
November 15, 2009, 12:43 pmyankee says:
Perhaps, but waiting until some indefinite point in the future to put it online isn’t making it any more valuable.
Quote
November 15, 2009, 1:05 pmShane says:
Yeah that’s what I mean. I was sorta making fun of the “we aren’t infringing because we destroy the original” line of reasoning. It’s absurd because it can be extended to almost any kind of infringement.
Quote
November 16, 2009, 1:20 amShane says:
I dunno — the Beatles Rock Band game sold over a million copies its first month. I’m not sure how much of the $100 million or so in sales that translates into royalties for the copyright owners. And just because consumers may own the entire recorded Beatles catalog doesn’t mean that they aren’t open to paying for albums like the remixed Love album, or that they won’t watch films/television/plays that license the music. A billion dollars is a bit of a stretch, but maybe not by that much.
Quote
November 16, 2009, 1:41 am