Doug Boyd and the Power of Digital Oral History in the 21st Century

Digital preservation and Internet access are not only transforming the way we record and convey history, they are also restoring the importance of humankind’s oldest means of storytelling: the oral tradition.

One of the most influential leaders in this modern oral-history movement is Doug Boyd, director of the Louie B. Nunn Center for Oral History at the University of Kentucky Libraries. Boyd (who blogs at http://digitalomnium.com/) is pioneering the use of digital technology to preserve and distribute oral histories. When I asked him what’s so special about oral histories, he said that a historian uses detective work and guess work to piece together resources and draw conclusions, while a recorded interview is a first-hand account told by a witness to history. It is living source material.

Some of the interviewee’s details might be fuzzy –- memory is far from perfect –- but the recording captures something that eludes the historian: how a person felt or what he or she thought about a certain event, place or time that they experienced. Boyd said, “A recorded oral history is more than just a quote on a page in a book. It is a meaningful story expressed by the person who owns that story.” And digital technology makes it possible to hear these stories anytime, just about anywhere.

Until recently, the work of oral historians resulted in a taped audio or video recording stored in box on a shelf in a repository where a researcher may or may not dig it up eventually to listen to it. But modern expectations of immediate access are changing that practice; anyone on the Internet can listen to them on demand. And low-cost, high-quality consumer equipment enables average people to contribute many more oral histories to the historical record. You can use the “C” word and dub it “citizen oral history.”

Boyd tells of a visitor to the Nunn Center who borrowed recording equipment to interview a World War II veteran. Within four days,  the visitor had conducted the interview, dropped it into his computer, added images that he found on the web, created a documentary and put it online. Boyd said, “This guy had about 700 Facebook friends. That quickly, he not only created this thing, he also distributed this interview to more people than we used to brag about serving at the Nunn Center in an entire year.”

Not only is digital technology enabling widespread distribution of recordings, it is also shifting interest to the recording itself and away from where oral historians’ interest had been for years: the transcription.

The value of an interview transcription was that you could skim it and quickly locate the topic you were looking for. After all, interviews can average a few hours in length and you might be interested in just one particular topic that only lasts a few minutes.

Homère by Jean-Baptiste Auguste Leloir

But one of the problems with a transcription is that it may easily contain a number of sloppy errors; given the tedium of the task, humans make mistakes when transcribing. (Standard transcription practice is to listen to a snippet of the recording, stop the recording, type what you heard, then listen to the next snippet, type it up and so on.) And once something is in print, the printed word becomes an accepted fact, especially if researchers don’t bother checking with the source recording. Boyd cites an instance of checking recordings against transcriptions where the transcription of a World War II event described “…an amazing Italian who was fast and quick”. On the recording it was actually an amazing battalion that was fast and quick.

For generations of oral historians, the transcription became the main object of interest, not the recording. Now the focus is shifting away from the written representation of  the recording and back to the recording, “the thing” itself.

That leaves the challenge of searching a digital recording to find specific points of interest. “Recorded speech is not 100% effectively searchable yet,” Boyd said. “Searchable speech is going to transform the way that we capture and understand and access history. But until that happens, we will still rely on manually generated indexed text.”

So you still need transcribed text to search, if only as a finding aid to the recording. To generate transcriptions,  one possible alternative to listening-and-typing is speech-to-text software, which is currently less than perfect. The process of speech-to-text conversion is complicated by diverse regional accents and dialects, lax pronunciation, subtle contexts and poor recording quality. And auto-correct adds even more gibberish to the end result. Boyd compares the present state of speech-to-text software to where OCR used to be a few decades ago in terms of efficiency or lack of it.

The quick and cheap interim solution for audio transcription is the audio equivalent of “dirty OCR,” which was good-enough, hit-and-miss OCR that no one intended to do quality control on. Boyd said that he is considering using automatic speech recognition to generate “automatically created descriptive metadata,” which will be searchable but not visible to the user.  A human reviewer would conduct a second pass for error correction and refinement. The end result will be good enough to get you to the spot you’re looking for in the recording.

Boyd helped create a tool called the Oral History Metadata Synchronizer [See the OHMS overview videos on YouTube.] to enable users to search text that has been correlated to time codes in the transcription His staff begins the process with a transcription and manually drops in time code markers at one-minute intervals, then synchronizes the keywords, terms and concepts with the time codes. The one-minute markers are deliberate. One minute is about the average attention span of someone poking around for the keyword reference…the “sweet spot” as Boyd calls it. The searcher, looking for a keyword, is never more than 60 seconds from the spot in the recording where the keyword is located.

OHMS is highly effective in enabling Boyd’s staff to do a quick turnaround on the recordings, which are then made available and searchable on the Kentucky Digital Library. [For an example of keywords plopping into one-minute markers, see the interview with jockey Eddie Arcaro and search for the term "Preakness."] And quick turnaround is part of the appeal. Again, Boyd is satisfying the modern “expectation of access.” Boyd said, “We are not waiting for two years to raise funds for a project so we can transcribe it verbatim and put it online.  We are syncing and indexing and putting recordings online daily as part of our workflow.”

The new version of OHMS introduces an indexing feature that can be effective without a transcript, creating searchable metadata that corresponds to moments in the interview. This method of indexing can be implemented for a fraction of the cost of a transcription and it is much more effective and relevant than a searchable verbatim transcript because human judgment is a big part of the process; only a human can determine things that are implied in the text but not explicitly stated. For example, there may be a two hour interview about life under segregation in which the interviewee never says the words “segregation” or “desegregation.” An effective indexer  understands the context, surmises the meaning and knows when to link the word “segregation” to an appropriate passage.

Boyd is also one of the pillars of the IMLS-funded Oral History in the Digital Age, a rich resource of best practices for producing quality oral history projects. OHDA offers guidance, through essays and video tutorials, for creating high quality productions. The tone is not preachy but one of modelling examples and suggestions based on best practices. Boyd said, “We realized we could help you choose a camera but the video is still going to look terrible if you don’t know how to use it. So how do we fill that knowledge gap in?”

For example, one video tutorial shows the basic concepts of three-point lighting. Another demonstrates how to process audio but also demonstrates how, when you strip out the high end of a noisy video you lose the background sound of the crackling fire. Once informed, you can decide for yourself how you’d like to approach a situation.

You can compare features of video and audio recorders and make an informed decision based on your budget and your needs. Boyd worked with 45 other authors to cover the wide range of questions he is constantly asked about equipment and production methods. And he likes not only how the resources empower people and boosts the professional quality of their work, but also how easy it is to update the information on the site quickly and cheaply. “It doesn’t require a grant to maintain it,” he said. He plans to update it soon with practical examples of metadata; his article “The Digital Mortgage” already includes a good introductory-level explanation of checksums. And he said he’s gotten great feedback from the Society of American Archivists about the site and has begun to weave their feedback into it.

Boyd’s work is part of a growing body of similar work, such as StoryCorps and the Library of Congress’s Veteran’s History Project – work that honors our ancestors by allowing them speak for the historic record. Their voices are both pushed out into the world and digitally preserved for future generations to appreciate.

For example, much as has been written about the allied invasion of Normandy Beach during World War II. But the words of a veteran such as Jesse Beazley resonate with simple power when you listen to him recall being part of the first wave of the invasion, joshing and cutting up with his friends on the ride over the English Channel and then, gradually, everyone in the boat falls silent. Beazley says, in this recording at the 00:55:00 mark, “I was thinking, would I get killed? Would I make it to the beach? How would I get killed?…I looked around on the ship, it was terribly crowded. We were sitting there on the floor. And all at once there was a blank expresssion on everybody’s face. You look at a fella maybe 18, 19 years old and there was nearly a death look. No talking. No kidding. No nothing.”

 

 

 

2 Comments

  1. Jacky Moore
    January 19, 2013 at 5:53 pm

    I read this article with great interest as my own research involves oral testimonies. Allthough I reference my transcriptions to explain specific instances in their lives it is the actual recordings that give me goosebumps; when I hear comments or detail missed the first time round.I agree with the comment ‘It is a meaningful story expressed by the person who owns the story’, one of a number of insightful observations made by Doug Boyd. Listening to my interviewees, and recording their life histories, the parts pertinent to my research, has been one of the most rewrding experiences of my life.

  2. Mike Ashenfelder
    January 22, 2013 at 12:31 pm

    Thanks for your comment, Jacky. I agree. Hearing the subject’s voice (and seeing his or her face, if it’s a motion picture) is the most valuable result of the recorded interview.

Add a Comment

This blog is governed by the general rules of respectful civil discourse. You are fully responsible for everything that you post. The content of all comments is released into the public domain unless clearly stated otherwise. The Library of Congress does not control the content posted. Nevertheless, the Library of Congress may monitor any user-generated content as it chooses and reserves the right to remove content for any reason whatever, without consent. Gratuitous links to sites are viewed as spam and may result in removed comments. We further reserve the right, in our sole discretion, to remove a user's privilege to post content on the Library site. Read our Comment and Posting Policy.

Required fields are indicated with an * asterisk.