National Recording Registry: PowerBI

The National Recording Registry is a program of the Library of Congress wherein the librarian acknowledges audio recordings as aesthetically, culturally, or historically significant, to encourage preservation of American's oral and musical history. Over almost two decades, the librarian has inducted over 500 works, ranging from individual songs and albums to famous speeches to old radio shows to the entire collections of field recorders, producing a list that is both extraordinarily fascinating and nearly impossible to quantify. I for one, have had years of exciting discoveries using the list as "recommended listening," finding new favorite albums and entertainers - but was disappointed with the lack of organization in the Registry. As such, I started a side project, not to dive into the recordings, but to skim the whole pool, collecting, quantifying, and classifying the recordings, with the aid of Microsoft Excel and PowerBI, and drawing some conclusions about the Registry (the nature of the content, trends in induction over time, and accessibility to the recordings), just in time to anticipate the 2022 announcement of the 2021 inductees. Graphics to be updated soon!

As of 7 March 2022, the entire contents of the National Recording Registry (which I encourage everyone with free time to sift through!), when measured in duration and sorted by genre and year of recording, form the beautiful curves below:

The .pdf above was developed using PowerBI (if my organization licensed the software; you could well have seen a live-updating dashboard with working sliders). Figure 1a displays the total duration of the recordings in the registry, separated by when they were first recorded and colored by genre; Figure 1c does much the same for strictly musical recordings from after the popularization of the phonograph. Figure 1b compares the lengths of the longest individual entries in the registry, many of which can be identified in the main figure. The data, it should be noted before conclusions are drawn, came from many sources, and some of it is only estimation - for details, see this page, especially prior to drawing conclusions.

To draw some conclusions from this - the most obvious facets of each graph are the long-duration entries. The Lyndon B. Johnson presidential recordings, an estimated 850 hours of telephone calls and cabinet meetings, appear as a tower over the sixties in Figure 1a. The vernacular wax cylinder collection of UC Santa Barbara makes a dark cloud over the 1900s and 1910s, as it contains assorted cylinders from various years. The Highlander Center's 300 hours of civil rights seminars and discussions covers the 1950s and 1960s about as well, while Ronald Reagan's radio commentary stands up for the late 70s, before he left radio to run for president a second and successful time. Ignoring these whole categories of spoken word and field collection-type recordings, as in Figure 1c, we see many of the longest entries are classical works, such as the landmark 1965 recording of the complete Ring Cycle. Jazz and pop pepper the 20th century, with a great deal of rock emerging around the 1960s, though this reflects as well as the decisions I made in classifying works by genre (it is not quite the same as the system used by the Library in its listing; see here for details).

History of the registry itself

While these recordings may reveal a great deal about the history of the American people, it should be remembered that these data do not; they reflect only the interests held by the Librarian of Congress when selecting recordings to pique interest in audio preservation. As such, it may be interesting to note changes in the recordings based on year of induction, as is done in the figures below.

It is worth noting that in the first four years of running the Registry, the librarian was tasked with finding fifty entries worthy of induction; this was reduced to twenty-five in the following years. Unsurprisingly, those earlier entries still comprise much of the registry's interests in older genres, such as classical, jazz, and blues; rap, on the contrary, has seen its presence in the registry grow much more in recent years, with the addition of a new album each year. Country finds most of its inducted works have been so for a long time; this reflects only the bold 2002 decision to induct the entirely of the 1927 Bristol, TN recording sessions as significant, followed by a tendency to induct mainly singles to represent country music. Similar observations can be made of other lengthy early inductions, such as Louis Armstrong's Complete Hot Five and Hot Seven recordings (2002, Jazz), the complete recordings of Robert Johnson (2003, Blues), and the Blanton-Webster era recordings of Duke Ellington's orchestra (2002, Jazz), all of which would be released later as large compilations that dwarf normal albums.

Similarly, the Spoken Word and Field collection categories are dominated by their respective long entries, the aforementioned Lyndon B. Johnson recordings and the vernacular cylinder collection, inducted in 2013 and 2014, respectively. For some time prior, the longest entry in the registry had been Alexander Scourby's 1946 recitation of the complete King James Bible, 72 hours long. Figure 2c shows the spoken word and field collection recordings tend to be longer than musical offerings, echoing the same conclusion from Figure 1a.

Accessibility

Another topic of perhaps greater interest is the accessibility of these recordings to the public, which can prompt debates about the purpose of audio preservation - if some very precious wax cylinder is locked and kept untouched in a vault, what difference would its disappearance make? Should preservation consist of keeping things available to select historians should these things become relevant, or does it consist of keeping things present in the minds of the whole population?

The answer may be a mix of the two, as reflected by the nature of the preserved recordings. Many of the entries (in fact, about thirteen and a half hours' worth) are singles, which are often re-released on compilations and thus easy to find. Many of the albums benefit more recently from release on streaming platforms, such as Spotify. Many collections, such as Elvis's Sun Records sessions, have been compiled into albums. Other collections remain partially digitized across various websites - the Harvard Vocarium series, for example, is available on Harvard's website to those with a login - while others are held only in libraries which cannot offer streaming generally for copyright reasons - the New Music Quarterly series, for example, is archived in the New York Public Library's Rodgers and Hammerstein archive of recorded sound.

The charts below show what portion of the entries in each genre are available for streaming on Spotify, elsewhere online (see this page for where) as of 5 March 2022. Note that this chart was developed - coincidentally - during a small controversy in which artists (including Joni Mitchell, whose album For the Roses was inducted in 2007) removed their music from Spotify; these data are subject to change soon.

Note first that, for convenience, the partially digitized vernacular wax cylinder collection is shown with the later recordings as unavailable; actually, it is a portion of the recordings in each year that are such, including the lone 1888 cylinder in the collection.

Nonetheless, the most evident conclusion of this graph is that collection and the other salient entry are most available for streaming elsewhere than Spotify. Spotify covers music fairly well, especially more recent music, including the entirety of the rock, R&B, and Latin entries. It would cover all of the rap if not for outstanding copyright claims against De La Soul for their sample-heavy album, Three Feet High and Rising, and would include all the musicals if it picked up the early incomplete versions of Show Boat and Cradle Will Rock. Regarding music, many of the older pop, country, and folk songs and classical records not on Spotify can be found on YouTube. Many genres have at least one entry that cannot be streamed online - in jazz, these are the rare recordings of Lovey's Trinidad String Band from 1912 and Maria Schneider's 2004 album Concert in the Garden, of which she would prefer people pay for a copy; in country, it is a handful of lost records from the Bristol sessions; in classical, it is the Mapleson cylinders and the New Music Quarterly series, both rather old collections; in pop, it is a set of obscure early Yiddish songs recently released as a CD - not streaming; in gospel, it is an episode of long-forgotten radio show, Wings Over Jordan, as well as some Alan Lomax recordings of the 1959 Sacred Harp singing convention; in folk, it is the selections from the Ali Akbar Khan college archives.

Regarding the non-musical entries, many of the spoken word recordings - news broadcasts, radio shows, and of course presidential recordings - can be found between YouTube and the Internet archive, while the field collections, besides the UC Santa Barbara vernacular wax cylinders, are not online. Most of these are large archives of Native American music recorded by field researchers for universities, which preserve them principally for research purposes. They are not made available for streaming or released commercially, as they were not recorded for such purposes, and the consent of the performers would be difficult to obtain. Nonetheless, it is this idea of the sealed, protected archive that sometimes characterizes audio preservation, and can be trusted to fulfill the broadest aims of the National Recording Registry.