Number of historical recordings issued
The 20-plus major discographies used as the basis for this study, which are widely used by scholars and collectors, list more than 400,000 recordings of interest issued prior to 1965. That figure may grow as additional fields are documented that have not been adequately researched (e.g., White gospel recordings, recordings made in the 1890s, and post–World War II ethnic recordings).
Number of historical recordings extant
Experts believe that the vast majority of recordings commercially issued in the United States-probably more than 90 percent-still exist in some form.4 Often this is in the form of commercial pressings, since in the past, record companies routinely destroyed their older masters, and sometimes even the files that documented what these companies had recorded. However, archives and private collectors have assembled large collections of these recordings, and they serve as the source for many reissues. Relying on a worldwide network of collectors, Europe’s Document Records has issued more than 600 CDs documenting the contents of the book Blues & Gospel Records, 1890–1943, alone.
Many of the discographers whose works were consulted for this study physically examined most of the records they listed or received reports from others who had done so. Rock-era compiler Joel Whitburn claimed to own a copy of every record listed in his huge Top Pop Singles; individuals working in earlier periods commonly derived take numbers5 and label credit from copies examined. This was the case with the author’s Columbia Master Book volume covering 1901–1910. Most likely to have been lost are cylinder recordings made in the 1890s, but even these continue to turn up. Preservation specialist Glenn Sage reported that he has seen approximately 1,200 such cylinder recordings and believes that 8,000 to 10,000 are in existence nationwide. Another expert, Bill Klinger, puts the figure at about 7,000 to 8,000.6 Interestingly, 50 percent of the more-than-600 1890s cylinders that Sage has transferred were made by the Columbia Phonograph Company. If this sample is representative, approximately half of the cylinders from this era are still protected, even though the rights holder, Sony BMG, has retained or reissued almost none of them.
Though many early cylinders are in poor condition, they can still provide future preservation experts with the opportunity to retrieve their sound using technologies yet to be developed. Several advanced technologies for this purpose are in development.7
Percentage of the recordings in this study controlled by someone under current U.S. law and percentage in the public domain
An important provision of the 1976 copyright law is Section 301(c), which provides, for recordings made prior to 1972, that “any rights of remedies under the common law or statutes of any State shall not be annulled or limited by this title until February 15, 2067.” This means that recordings issued before 1972 will not be covered by federal copyright until 2067. Until that time, they fall under state laws, generally those regarding copyright, property rights, and unfair competition. State laws vary, and there is no known summary of their current coverage of sound recordings. However, these laws are universally interpreted to grant permanent ownership of recordings to the creating entities until such time as federal copyright law takes over in 2067.8
The net effect is that pre-1972 recordings are treated in the United States as protected from the time of their creation, no matter how long ago that may have been, until 2067. At that time, barring a change in the law, all such recordings will become part of the public domain. Inasmuch as the first commercial recordings were made around 1890, this is a de facto term of between 95 and 177 years, depending on the date the recording was made. All recordings included in this study were made before 1972 and therefore fall into this category.
Despite bankruptcies, abandonment, and long-dead record labels, under current U.S. law an overwhelming majority of historic recordings-in this study 84 percent-are still owned by someone. This is true even for the earliest periods, with more than 60 percent protected for every period after 1895. The figure exceeds 90 percent after the 1930s. It is also true for every genre of music studied. Most of America’s recorded musical heritage of the last 110 years, even recordings made in the nineteenth century, is protected by state and common law until 2067.
Reissues made available by rights owners
Ten percent or less of listed recordings have been made available by rights holders for most periods prior to World War II. For periods before 1920, the percentage approaches zero.
The demand for historic recordings
Despite laws discouraging unauthorized reissue activity in the United States or the importation of reissues of U.S. recordings from other countries (parallel import laws), foreign labels and small entities in the United States have made available a considerable number of such recordings. Our study found that other entities have exclusively reissued 22 percent of historic recordings, versus rights holders’ 14 percent. To the extent that rights holders reissue older recordings, they concentrate on recent periods with larger potential markets, while third-party distributors serve all periods more or less equally. As a result, third-party entities reissued more than rights holders did in every five-year period prior to 1945. But current copyright law has made such activity difficult and risky for small organizations and so has driven much of this activity overseas. It is worth noting that a label such as Europe’s Document Records, which has made available thousands of rare, pre–World War II American blues and gospel records not reissued by rights holders themselves, could not exist in the United States because of copyright restrictions.9
Accessibility of historic recordings to scholars and other interested parties wishing to reissue historic recordings legally
The first task for a third party wishing to reissue a recording is to find the rights holder. This can be difficult or effectively impossible. In this study, despite a major effort to trace ownership and consultation with several experts, 25 percent of the labels in the sample were untraceable and many of the other labels had to be assigned probable protected or nonprotected status. The uncertainty introduced by current copyright law has prevented many reputable companies, institutions, and associations from engaging in any reissue activity at all.
Conclusion
Evidence uncovered in this analysis suggests that a significant portion of historic recordings is not easily accessible to scholars, students, and the general public for noncommercial purposes. There are many reasons for this, but the primary one appears to be a convergence of two factors. The first is that the physical barriers created by recording technologies change often and have rendered most such recordings accessible only through obsolescent technologies usually found only in special institutions. Second, copyright law allows only rights holders to make these recordings accessible in current technologies, yet the rights holders appear to have few real-world commercial incentives to reissue many of their most significant recordings. The law has severely reduced the possibility of such recordings entering into the public domain, at least until 2067.
While there is no reason to assume that the law intended to create or sustain such an imbalance between the private and public domains, the evidence suggests that it has, in fact, created such an imbalance. This study indicates that there is an active and hardy network of foreign and small domestic companies, associations, and individuals willing to make historic recordings available; indeed, some do this in spite of laws that force them underground or overseas.
These circumstances create a complex policy environment. The time for sorting out these matters grows short as the recording formats become more difficult to maintain. We hope that the evidence we have gathered and the analyses of reissue practices we have made will help inform an urgent and much-needed policy debate.
FOOTNOTES
4 Correspondence with Sam Brylawski, Library of Congress, October 11, 2004.
5 Several “takes” are usually required to obtain a recording suitable for issue and these are generally numbered sequentially. Sometimes multiple takes survive for a given recording.
6 Telephone conversation with Glenn Sage, October 10, 2004, and correspondence with Bill Klinger, October 12, 2004. Sage provided label counts for the more-than-600 1890s cylinders he has transferred. Figures reflect the number of different titles, not the number of copies of those titles.
7 Examples are the experiments at Syracuse University in New York and at the Lawrence Berkeley National Laboratory in California. For more, see the New York Times, May 6, 2004, and the ARSC Journal, Vol. 31 No. 1 (2000), p. 161; and Vol. 35 No. 2 (2004), p. 308.
8 See, for example, Capitol Records, Inc., v. Naxos of America, Inc., United States District Court, Southern District of New York, 02 Civ 7890 (RWS)(2003), Section I(B)(1), Applicable Law.
9 Document Records avoids reissuing material readily available from rights holders.