Should new digital versions of classic old analog recordings sound essentially the same as the originals, or provide a fresh take on them?
For some record labels with valuable old sound recordings, licenses that should have established the labels’ decisions on this important business and aesthetic question appear to have gone missing. Or were not submitted in evidence because they were unhelpful. But this crucial gap in the record allowed a federal court in California to rule in favor of a bold theory of defendants CBS Corporation and CBS Radio, Inc.: post-1972 remasterings of pre-’72 sound recordings with the modicum of originality required after Feist get their own sound recording copyright and thus can be broadcast under the federal compulsory license rules. This is in sharp contrast to the direct negotiated license required for pre-’72 recordings governed solely by state copyright laws.
With a federal court in New York now considering a companion case, and the possibility of appeal in the California case–not to mention the stakes for the huge back catalog of pre-’72 recordings that have been remastered since then–it’s important to get this right. But it is a complex matter that sits at the confluence of multiple areas of technical detail.
First, the players. ABS Entertainment and other record labels separately own various original sound recordings of legends including Al Green, Andy Williams, The Everly Brothers, Jackie Wilson, King Floyd, and Mahalia Jackson. They licensed third party sound engineering firms to “remaster” these valuable pre-’72 recordings, who then rightly or not did more than simply remaster the tracks for digital media. CBS Corporation and CBS Radio, Inc. broadcast these remastered recordings through their digital transmission stations, relying on the statutory compulsory license for non-interactive digital broadcasts. Thus, they did not obtain any licenses directly from the labels (nor it seems from the remastering distributors).
Second, the background copyright laws. Contrast copyrights in musical compositions from those in sound recordings. The former are rights to the song as the composer wrote it; the latter are for a sound recording of that piece. Before a 1971 amendment to the Copyright Act of 1909 then in force, sound recordings were not protectable under federal law. Thus, labels relied on state law alone for the recording right (but of course federal law for the underlying composition right). This meant that in practice, traditional terrestrial broadcast radio stations paid royalties only on the composition right. But even once effective in 1972, the sound recording right was limited to reproductions and did not extend to performances. So stations still did not need to pay royalties on sound recordings.
In 1995, as streaming digital transmissions became a reality, Congress passed the Digital Performance Rights in Sound Recordings Act. This added an exclusive right to public digital audio transmissions for post-’72 sound recording owners (but did not address pre-’72 recordings). It also did not affect existing terrestrial stations still broadcasting analog signals. On a side note, that distinction set the stage for digital broadcasters’ advocacy of the Internet Radio Fairness Act, which would have amended the sound recording public performance right. The bill did not pass, and the better remedy is for Congress to amend the Copyright Act to add a general sound recording performance right so that terrestrial analog broadcasters have to pay royalties as well.
The new law set up a distinctions between “interactive” digital transmissions (the listener can select the exact song), in which the provider has to negotiate rights directly with the rights holder, and “non interactive” transmissions (the listener can only suggest or “like” tracks to influence a playlist determined by the provider’s algorithms or DJs), where there is a compulsory license with rates set by a government process.
The system applied only to federal sound recording rights (i.e. post-’72) and thus left open the question of whether digital broadcasters needed to secure licenses from pre-’72 recording owners under state law. But neither terrestrial analog nor satellite or online digital broadcasters had paid royalties on sound recordings under state laws.
This issue came to a head when Flo and Eddie (of famed 60’s band the Turtles) sought to enforce their pre-’72 recordings against digital broadcasters including Sirius XM in multiple states. While defendants raised issues such as laches, in part based on the uncontroverted evidence that Flo & Eddie–and indeed apparently no other major sound recording rights holders–had ever enforced royalties against either digital or analog stations ever since terrestrial stations began broadcasting music recordings, both the California and New York courts were unpersuaded. These twin decisions rocketed through the broadcast world and presumably licenses are now being worked out–either through the collecting agency of SoundExchange or otherwise.
The ABS v. CBS case suggests that pre-’72 sound recording owners are indeed demanding royalties–from both digital and analog broadcasters–and that some broadcasters are resisting. Given that California and New York are the lead state copyright law markets for this enforcement and that the Flo & Eddie rulings pretty well established the rights and enforceability of pre-’72 recordings, broadcasters will have to come up with some creative alternative defenses.
Enter the CBS post-’72 remastering defense.
Because CBS could show that it was only broadcasting copies of such remastered recordings, and that they had substantive changes from the original mastered copies, the judge found them to be derivative works deserving of their own copyright. Having been made after ’72, they then fell within the federal public performance rules. Thus, the terrestrial broadcasts would be exempt altogether, and the non-interactive digital ones would be subject only to the statutory royalty. CBS does not appear to be providing interactive services at this time. The kicker is that even the statutory royalty would then presumably be paid to the third party remastering firm as the proprietor of the sound recordings actually performed.
Of course, the remastering had to be authorized by the labels (who would also have to provide the the original analog tapes on which the sound recordings were made for any high quality remastering to be possible). But an open question is whether the the labels should have provided only an original stereo 2 track master, and not the multitrack source tape, if they truly sought only remastering for digital distribution. Doing so would have prevented much of the substantive changes that convinced the court to find a derivative work in the “remasters.”
To understand this requires a brief review of the sound recording production process.
The mastering stage is the final step before mass copying onto distribution copies such as vinyl or cassettes. The process optimizes the sound recording for that particular format. This is crucial because most distribution formats traditionally had inferior sonic characteristics compared with the source material.
For example, the “headroom” of professional 2″ 24 track audio tape is much greater than that of a cassette tape. This means that the 2″ tape can handle a wider dynamic amplitude range (soft to loud) than the cassette tape. Even the open reel or reel-to-reel tape deck had wider tape than the cassette and it used both sides of the tape, so only two tracks were recording per side (by contrast the two sets of stereo tracks– four tracks in all–were recorded on one side of cassette tape). So, part of the mastering process is designed to “compress” the sound recording by raising the amplitude of soft passages and decreasing that of the loud ones. Without compression, soft passages might be inaudible, while loud passages might be distorted from “clipping,” as the limitation of the format cannot handle the signal peaks and simply “clips” them off. Given the different sonic characteristics of various formats (vinyl, 8 track, cassette, CD, MP3), multiple duplication “masters” of a sound recording might be made–one for each format.
Thus, traditionally there were at least three recording stages for professional releases. First, the original capture of the acoustic or electrical waves on tape. Second, the “mix” of the raw recorded tracks down to a single stereo or mono 1/4″ open reel tape recording at high speed that represented the final product that the artist, producer, and label ideally sought to distribute to the public. But, because few consumers were willing to acquire high fidelity open reel decks necessary to reproduce studio master quality, and given the widespread adoption of acetate and vinyl record platters from before multitrack recording became commonplace, the labels of course had to release singles and albums in other formats as well. Thus, third, the mix tape had to then be copied onto yet another 1/4″ open reel tape that captured further processing of that mix to compressed, equalize, and otherwise optimize it for high speed duplication onto copies in a particular commercial release format, or, in the case of vinyl, to create the acetate master for duplication.
On a side note, this helps explain why artists were often frustrated by the sound of the analog commercial duplication copies of their recordings distributed to the public. After (at least) two generations of analog copying, combined with the usually inferior nature of the distribution format and consumer playback equipment, the copy sounded a lot less live, dynamic, and punchy than the original raw source material played off the 24 track machine, and run through the pro mixing console and monitor speakers in the studio.
When we entered the digital age–first with CDs and then with digitally transmissible files–artists, labels, and audiophiles wanted to maximize the greatly enhanced headroom and dynamics available to distribute copies much closer to the original source material. Thus, many previously released classic recordings were “remastered” for these new formats. For some recordings, the new digital distribution formats were so good that they could actually reveal the limitations of the source material. This led to a practice of printed disclaimers on CD cases warning the listener–who was now expecting “flawless” audio–that limitations such as hiss or clipping were in the source material and thus were not the fault of–and could not be removed by–the digital distribution format.
But in some cases, artists, producers, or labels were tempted to engage in some remixing as well. This was partly because the potency of the new digital formats allowed instrument takes that didn’t work in older formats to be reevaluated. While “remix” as a term is familiar to anyone who follows digital media debates–usually paired with “mash-up”–I am using it in its narrower sense as a new processing of the individual raw instrument or vocal tracks from the original multitrack source tape, relative to each other, and not simply processing the final stereo mix tape as whole. This should be familiar to anyone who has played with GarageBand or similar music apps. You record a piano track, and then separate drums, bass, guitar, vocal etc. tracks. Once the tracks are “cut,” then you can go back and change the EQ (equalization, as in treble and bass response), spatial effects (e.g., reverb, delay, position in the stereo field), volume relative to other tracks, and other things. You might even decide you don’t like the piano track and turn it off for the final stereo mix exported into MP3 or other format. But so long as you don’t delete that piano track entirely, you can always turn it back on for a later remix of the recording.
So remixing is starkly different from remastering. A remaster is supposed to be a faithful reproduction of the original mix, but optimized for a new format (such as digital). No changes in the relative sounds of the various instruments should occur. Any changes in EQ etc should be applied universally across the final stereo mix (because the mastering process is applied to the mix not to the source tracks).
Turning back to ABS v. CBS, then, it is clear that the third party engineering and distribution firms did not limit themselves to true remastering. Further, the court seemed highly confused by the remixing/remastering distinction when it cited plaintiff’s expert:
“Finally, in their depositions, Plaintiffs themselves also agreed that a sound engineer’s choices made during the remastering process are creative and subjective. (See Gressel Decl., Ex. C at 80 (“[T]he mixing engineer at least has to decide the levels of each track. Which vocal performances do you want. Which instruments do you want. When do you want them out. There’s a lot of latitude at that point in the production process.) Ex. B at 60, 75).”
ABS v. CBS at 9 (emphases added). The quote is clearly discussing mixing and not mastering. A mastering (or remastering) process would not sub in different vocal or instrument tracks from the original mix, or change when they are on and off throughout the recording. (For some tracks, you may want to drop out only a portion of the performance in the final mix.)
But the judge may have also been confused because some of the experts were correctly pointing out that even a true remastering for a new format required some engineering. It was not simply a “‘drag and drop’ conversion of an analog recording to a new digital format.” Id. at 7. Post recording signal processing can involve a spectrum of activities ranging from that kind of electro-mechanical “drag and drop” of simply copying an existing analog mix into a digital format and then duplicating/distributing, at one pole, and extensive remixing of the source instrument tracks (including perhaps recording or overdubbing new instrument or vocal tracks) at the other pole.
The key principle is whether the “mastering” engineer is simply trying to optimize the mix–as decided solely by the artist and producer–for a new format or whether she is actively changing that mix. Or, is the engineer applying signal processing universally across a final stereo (or mono) mix, or differentially on to distinct instrument tracks.
This is why I suggest that the true mastering engineer need not be working from the multitrack source material anyway, but rather only from the two channel stereo mix (i.e. left and right tracks). In this case, it would be difficult–although hardly impossible–to alter relative EQ, volume, etc. And it would be easier, as the mastering engineer would not have to recreate all the signal processing and mixing of the separate tracks first, and then engage in the further mastering processing after that. Of course, some mastering engineers might want access to the multitrack source because some adjustments to individual tracks might be necessary to accommodate the limitations of the distribution format (or maximize its benefits), even as the engineer is working from a reference mix done by the artist or producer and simply trying to make as faithful a reproduction of that as possible, given the final distribution medium.
But here is the kicker: the court’s central confusion about the difference between remix and remaster doesn’t matter because–as it rightly points out–plaintiffs had licensed the third parties to produce the remasters. Thus, as the court said, the plaintiffs presumably obtained whatever compensation and rights they sought under the license to produce the remasters. Without such license, of course, any “remastering” or remixing would result in an unauthorized infringing copy of plaintiffs’ sound recordings, even if it was also a derivative work in its own right.
Which finally brings us to the ultimate question: what was licensed from the various plaintiffs to the respective third party sound engineering firms? Unfortunately, none of the parties produced any of the licenses for the trial. It may be that no one has them anymore. Presumably they existed at some point. Hopefully in a formal writing. (If they were not exclusive, then there was no need for a signed writing.) Whether memorialized or not, the key question then is whether they authorized a true remastering only or instead the substantive remixing reported by the experts: e.g., “additional reverberation”; “different key and at a faster tempo”; and “differed in the musical performance” (elsewhere in the opinion the court suggests that different vocal or instrument takes were used). Id. at 12. Given the substantive changes that led to the finding of sufficient originality, plaintiffs should have a case against the third party engineers if the license was restricted to true remastering. These were big changes that the rights holder (and ideally the artist and producer) would had to have authorized, particularly because of the derivative work issue. Because even the Copyright Office’s guidance on derivative works in sound recordings limits such status to remasters where “the preexisting recorded sounds must have been rearranged, remixed, or otherwise altered in sequence or character, or there must be some additional new sounds” and in particular that “involves multiple kinds of creative authorship, such as adjustments of equalization, sound editing, and channel assignment.” United States Copyright Office, Circular No. 56, “Copyright Registration for Sound Recordings”
A lingering question is why the plaintiff labels would have allowed such changes to have been made to the original sound recordings. The short answer from them in the litigation was that they did not: Plaintiffs “never would have permitted a distributor to make any substantial or non-trivial changes to the sound of a Recording” and “to the best of [the declarant’s] knowledge none of the Licenses authorized any of the Distributors to prepare a derivative work.” Id. at 14. But given that these new “remasters” appear to have been around for a while, it seems odd that plaintiffs had not sought to remove them from the market before the current litigation against CBS. Further, the inability to produce the licenses is problematic. It has become a common practice to increase the commercial value of “remasters” by billing them as the equivalent of a movie “directors cut”–in the music genre this would be a more accurate representation of what the original performers and producers sought to record and release (perhaps against the original decisions of the label to keep the material to a certain length or within certain other parameters). Some of these remixed remasters are indeed revelatory. Others seem like marketing gimmicks to prod reluctant fans to buy the same recordings they already have in new formats/versions.
Finally, though, even if the “remasters” were derivative works in their own right, they clearly still contained pre-’72 recorded tracks. Thus, CBS would need a license under the Flo & Eddie ruling to play them. The court accepted defendants’ preemption argument that the federal rights in the remasters trumped any state rights in the original tracks. But the Copyright Act’s own section on preemption expressly provides that with regard to pre-’72 recordings “any rights or remedies under the common law or statutes of any State shall not be annulled or limited by this title until February 15, 2067.” 17 U.S.C. 301(c). Further, the section on compilations and derivative works states that “The copyright in a compilation or derivative work extends only to the material contributed by the author of such work, as distinguished from the preexisting material employed in the work, and does not imply any exclusive right in the preexisting material. The copyright in such work is independent of, and does not affect or enlarge the scope, duration, ownership, or subsistence of, any copyright protection in the preexisting material.” 17 U.S.C. 103(b).
Accordingly, the court’s preemption decision should be reversed on appeal, and hopefully not followed by the New York court.
But, once again, the licenses–should they surface–could be dispositive. If they expressly license the remastering firms to not only make and distribute copies of the underlying pre-’72 tracks embedded in the remasters, but also to publicly perform, or license others to publicly perform, those tracks, then the firms could sublicense CBS and other broadcasters to do so. But because these licenses appear to have been executed before the Flo & Eddie decisions, it is unlikely that express provisions about public performance rights were included. Instead, the firms were likely just authorized to make and sell copies. Further, the firms themselves were a bit vague about there own sense that they had derivative work sound recording rights. Ultimately, it may be that the licenses were themselves vague about exactly what was meant by “remastering” and both the labels and the firms may have been hoping to boost sales in the new formats by being able to promote at least some changes beyond simply the general higher fidelity and reliability of digital. It would be interesting to see whether any of the CD covers in fact promoted substantive changes beyond the shift to a new format.
Regardless of whether any licenses eventually surface, and of what they contain, neither CBS nor the firms seem to have contended that the latter did in fact license CBS to publicly perform not only the derivative work, but the underlying pre-’72 tracks. They could grant such a license now–again if they can successfully argue that they secured such a licenses themselves from the labels–but it is unclear whether that should be allowed to be retroactive to cover the period of infringement that has already occurred.
It will be interesting to see how the New York court rules.