Editor’s Note: Unlike other posts that start off text or images, Wendy Hsu @WendyFHsu opens up her third post in her guest series on digital ethnography with sound. She wants us to click PLAY before reading on. It doesn’t matter if you don’t use sound in your fieldwork, you’ll still find this to be a useful exercise in opening your ethnographic ears.
After you click PLAY, you’ll appreciate Wendy’s message: our fieldsites are rich with sound data that carries a lot of meaning. She closes her post with a great discussion theorizing digital ethnography as horizontal versus vertical immersion.
Above is a field recording of mini pinball machines that I collected in the Lungtan township in Taiwan. In it, you can hear the sounds of the machines, scooters, and a conversation that I had with my father while we were trying to figure out how to play the pinball machines. This field recording is rich in texture and meaning.
I know that not all ethnographers work with sound. But I do think that it could be useful to reconsider the sonic (and by extension, the visual) dimensions of our work. I propose an engagement with the textures of human speech in its original sonic. This approach counters the traditional emphasis on text and its impulse to textualize sound in ethnography. This perhaps is most on conspicuous in the practice of transcribing interviews.
You all can probably recall moments of dealing with the complexity of meaning embedded in the tone or the delivery of oral content in interviews. There are sounds of the environment that the informant has chosen to carry out an interview or interact with. Do these sounds reveal anything about the speaker and her relationship to her physical and social environment? Are there other voices in the room? Incidental sounds? Does the tone of the speaker react to and interact with the sounds of the environment in anyway? Where are the points of dissonance and resonance?
Sounds are rich in meanings. Daniel Makagon and Mark Neumann wrote in their book Recording Culture , “getting any kind of recording is also a mode of exploration and investigation in its own right.” Sound recording can be productively considered as an ethnographic practice because it can enlighten us on social life. Textualizing interview recordings can run the risk of flattening the sonic meanings that orbit around a person’s voice and its resonance in an environment. I recommend that while we transcribe (if it has to be done), we pause and listen closely to the textures of sounds and voices that inform our ethnographic understanding.
In this post, I give you some tips on how I use digital tools to augment my engagement with the sounds in my research. I encourage you all think about how you may do this with your own fieldwork materials and data.
In addition, as the third post of my series on digital ethnography, I will focus on how we might use digital technologies to deepen our engagement with the textures and materiality of culture. Digital technologies afford us the capability to engage with content in multiple modes and at multiple levels. With the example of my webscraping project, I captured data in the form of text, and then translated it into lat-long coordinates that subsequently became visualized as pins and clusters on a map. These clusters then become resized, forming perceptible patterns, as I zoomed in and out on the map. The combination of scalability and multimodality that digital technology offers makes visible both (micro) details and (macro) patterns of cultural content that were previously imperceptible to us.
Humans are better than computers at finding patterns that are meaningful because our minds are able to handle contextual thinking and calculating really well. But we also pale in comparison when it comes to switching between modes and scales. So why
shy away from utilizing computational power to help facilitate our analytical processes?
Seeing the textures of sound
As an ethnographer who works primarily with sound, in the form of field recordings, interviews, and music recordings, I have found utility in leveraging digital audio workstation [DAW] for the analytical purposes. On my last trip to Taiwan, I was lucky enough to get my hands on a set of cassettes of nakashi (postcolonial itinerant) musicians entitled The Wandering Blind Singers [Figure 1].
During the process of digitization, I came to notice sounds that were extraneous to the content of the music. Using Audacity, a free and open-source audio workstation program, I quickly saw that the recordings in this tape set are done in mono. This is unusual since most studio recordings in Taiwan have been produced stereo since the late 70s. Looking at the waveform, which displays the change in amplitude (volume of sound) over time, I found that many of the tracks get cut off before the natural decay of the instruments complete. You can see an abrupt amplitude reduction [figure 2].
Upon closer listening, I discovered an occasional environmental sound seeping through the walls in the studio. In one song, the sounds from a vehicle moving on the street got picked up in the recording. All of these observations add up to an amateur low-fidelity and low-budget production quality. These characteristics, in the context of the recording history in Taiwan, are rather unusual. My conjecture is that this low-budget production is telling of the lower-class status of these nakashi musicians and their fringe position in both the music industry and the society at large.
Inspired by Tanya Clement‘s on distant listening — as a technique to obscure sonic patterns of Gertrude Stein’s poetry — I began looking at the spectrograms of these recordings. In contrast with waveform, spectrograms reveals patterns of frequency (as well as amplitude) changes over time. Thankfully, Audacity makes this mode of sound visualization available through only one click. Spectrograms are particularly useful when I am analyzing the stylistic qualities and vocal timbres of nakashi singers. Notably, these singers’ tight, enka-inspired vibrato stands out as horizontal wiggles on the spectrogram [Figure 3]. You can listen to this particular portion of the song by playing it back on the SoundCloud player below the image. I imagine that this mode of visualization would be useful in exploring the stylistic contours of vibrato across the related musical genres. [If you’re interested in playing with spectrograms in Audacity, I recommend this video tutorial by Matt Thibeault.]
Seeing leads to further inquiries
I imagine that the same attention can be applied to visual content (video and photographs), especially after seeing Tricia’s successes with using Instragram as a field image annotation method. I stumbled upon surprises when I was scanning the album art and liner notes of some of the nakashi tapes. After scanning and then zooming in on the image, I gained a closer view of the granularity of the album art to discover traces of its print production.
I will take a cassette entitled “Sounds from Taiwan’s Underclass Vol. 2”, a mix of field recordings and low-budget studio recordings of nakashi artists, Catholic worship songs of Taiwanese aboriginal groups, among others, compiled by renown music critic He Yingyi. The cassette was released by the Taiwan’s historically important indie label Crystal Records in 1995. After zooming in on the artwork, I discovered an unusual dotted pattern in black ink that underlays a few color layers [Figure 4].
According to my archivist colleague, this pattern indicates that the album art was printed via the halftone technique. Without knowing the history of print in Taiwan, I gathered that this print work was done cheaply. By the mid 1990s, Taiwan was the major site of manufacturing of computational components. Halftoning, a common print technique a few decades prior to the 1990s, would have been considered as obsolete in the time of cassette’s production. Given the fact that this cassette was printed as a laser disc, and its Vol. 1 counterparts printed as a double-CD set in the 1991, I suspect that there might have been an another intention, besides saving cost, behind the choice of halftoning the album art. My guess is that the halftone technique was deployed to signify the music’s associations to the social underclass of Taiwan. This hypothesis, of course, is only the beginning of further explorations of cassette and print culture in Taiwan.
Before we hurry into tagging and filing media content based on their relevance to our research questions, why not spend some contemplative time deepening our sensory engagement with field materials in their various modes and scales? Small observations can lead to big discoveries.
Empiricism, immersion, and all the things we care about
A deepened engagement with cultural content in sonic, visual, and geographic registers has allowed me to see patterns of social linkage and cultural meanings that I had not anticipated in traditional forms of field data analysis. I think of these data processing methods as ways to achieve “radical empiricism,” a term that I use to describe my goals in finding and documenting socio-musical processes with empirical specificity and precision.
I consider the strive toward empirical precision a form of immersion, a point that Jenna Burrell raises in one of her posts on big data. She cites Howard Becker’s statement to evoke the ethnographic interest in empirical closeness — “the nearer we get to the conditions in which [the people we are studying] actually do attribute meanings to objects and events, the more accurate our description of those meanings are likely to be” (Becker 1996). To this notion, Burrell draws a helpful distinction between ethnographers and big data scientists by saying that former “do a whole lot of complementary work to try to connect apparent behavior to underlying meaning.”
To add to that discussion, I’d claim that computational immersion is not exclusive to the domain of big data. We ethnographers, too, can achieve immersion — through a computational means — in communities that are digitally embedded or not. I like to think of this computational immersion in two different directions: horizontal and vertical [Figure 5]. Horizontal immersion explores general contours of social actions and events. It typically engages with a large number of data points and provides categorical answers. I consider projects that are heavy on the side of datamining (like my Myspace webscraping project or Alan Lomax’s cantometrics) to take on a horizontal contour.
A vertical approach, on the contrary, engages with a single artifact (potentially event, dialog) that represents patterns of cultural life on a larger scale. The techniques that I discussed in this post would fall into the vertical category. Visualizing sound and magnifying images as an interpretative intervention afforded me a closer access to the subtle nuggets of meaning that often resists categorical thinking. Most ethnographers I know engage with their objects of study in both horizontal and vertical manner, but we usually call it something different, generality-specificity or macro-micro.
As ethnographers, we study the social and cultural life. With the tools at our disposal, we could digitize virtually all elements of life. I’m confident that sources of data will continue to proliferate (although our access to them may not). Through the use of recently refined sensing technologies, we could collect data pertaining not only clicks on websites (big data type sources), but also embodied data related to sound, sight, gestures (e.g. digitizing musical performance gestures), positions, etc. So I invite other ethnographically minded colleagues to innovate together the use of emerging computational means to extend our immersive ethnographic practices.
I’m a proponent of playing with and repurposing technologies out-of-the-box and beyond, to deepen our data exploration and engagement. The reason I emphasize “playing” is because it often leads to discoveries outside of the mode and scale of conventional data process. And I find these surprises most fruitful in generating further inquiries and speculations.
So far in this series, I have discussed the role of emerging technology in facilitating field data acquisition and processing. In my final post, I will talk about how we can think outside of the monograph as a medium while harnessing the powers of digital platforms to present ethnographic narrative and argumentation. I can’t wait to introduce some creative multimodal ethnographies!
ALSO IN WENDY’S SERIES:
- Part 1 On Digital Ethnography, What do computers have to do with ethnography?
- Part 2 On Digital Ethnography: mapping as a mode of data discovery