For several years the Library of Congress By the People program has featured sheet music campaigns, offering volunteer transcribers the opportunity to interact with digital sheet music collections by transcribing nearly all text found within. Ultimately, these crowdsourced transcriptions will be published on the Library of Congress website, making select sheet music collections keyword searchable. This is true of the Women’s Suffrage in Sheet Music digital collection, which is now keyword searchable across all lyrics, tempo markings, and advertisement text. Before being transcribed by By the People volunteers, suffrage sheet music was only searchable by its bibliographic information (title, composer name, etc.). For an example of this newfound searchability, if someone wanted to find songs in the digital collection that reference President Woodrow Wilson, searching “Wilson” in the “This collection” search bar in the upper right corner of the page will yield six results. About five minutes of investigating these results will let you see that three pieces of sheet music reference President Wilson in the lyrics; imagine the amount of time saved when the alternative would be downloading over 200 pieces and reading the lyrics in search of these references!
We are currently crowdsourcing transcriptions of the Music Division’s Sheet Music of the Musical Theater, a digital collection consisting of more than 16,000 musical theater piano-vocal selections from shows produced between 1880 and 1925. Sheet music transcriptions pose some unique questions for our volunteers, so let’s review our guidelines:
We want to capture nearly all text found in the sheet music: title information, composer and publisher information, lyrics, dedications, advertisements, copyright notices, etc. We also invite volunteers to transcribe tempo and expression markings; these are words that indicate the speed and mood of a piece (e.g. moderato, vivace, allegro, tempo di marcia, etc.). We do not see a need to transcribe dynamic markings, which dictate the loudness or softness of the music (e.g. f, mf, mp, p, cresc., dim., <, >, etc.). Transcribers can also disregard time signatures and key signatures indicated by symbols at the beginning of each staff line (i.e., treble and base clefs, sharps and flats, and the numbers that indicate time). And, of course, we are unable to tackle musical notation with these transcriptions.
After transcriptions are published on loc.gov, thus making the material keyword searchable, the final stage of a transcription campaign’s life cycle involves the Library of Congress publishing a dataset for each completed campaign that researchers can download as a .CSV file. These By the People datasets include all of the transcriptions and tags that volunteers created and reviewed during the transcription process, making computational research possible. For example, the Library’s mostly recently-published dataset concerns writings by, from, and to Leonard Bernstein. Rather than just sheet music, this campaign called for transcriptions of correspondence, speeches, scripts, essays and other writings from the Music Division’s Leonard Bernstein Collection. Read more about By the People datasets in this 2024 blog post from the Library’s The Signal blog, for more details about what research possibilities are unlocked with these products.
40% of the 16,000 Sheet Music of the Musical Theater digitized titles are completely transcribed, but the remaining 60% are in need of review by a second set of “By the People” eyes. Once we collectively review and approve these transcriptions, researchers will be able to keyword search and more efficiently analyze almost half a century’s worth of musical theater repertoire. Join us in closing out this tremendous transcription campaign!