From time to time I get asked what set up I recommend for linguists’ to use when they are making field recordings. The requests often go like this:
Greetings, I’m writing because I am just about to head out to do field work for my degree and I’m wondering if you have any advice on the latest audio recorders for linguistic fieldwork.
My response is invariably to pivot and respond to the question by asking what sort of products they would like to make. That is, linguists need to think about what can be done with the recordings, and how they need to influence the audio product’s creation and future use.
Making good audio recordings is about making recordings someone will want to listen to again, and again, and again. Invariably researchers (linguists) will listen to their recordings over and over again as they transcribe them. However their recordings gain value when others also want to listen to them. I recently was listening to a presentation where the presenter from Brazil, demonstrated some audio recordings from his field work. He showed a picture of the recording set up. In the picture it showed that he was using a Zoom H4n. In the recordings it was evident that the gain on the mics was pumped really high. The picture showed that he was using the built in mics. In the recording, the background noise was really full front while the speaker was perceptually distant — the chicken and the noise floor (the sound of “air” in the recording) were louder than the language speaker. Controlling for these factors are fundamental issues in basic crafting of audio recordings. Craftsmen (and women) who appreciate their art will get the tools they need to make their art (and artifacts) usable and enjoyable. This presenter from Brazil, played a recording from a language in which there are only about 8 speakers left… linguistically the language is already like a classic car with low mileage—RARE. However, a good recording of a rare language is like a classic car dipped in gold. Microphone choice and placement allows the audio craftsman to tune down the gain, get closer to the subject and thereby record at levels where the signal to noise ratio (the sound of “air”) and the chickens are not heard in the recording or are perceptively in the far background rather than in the foreground1. The ability to attenuate a listener to specific parts of the audio stream is where the artistry and the skill comes into full view. It is this applied artistry that makes audio recordings valued and worth listing to over and over.
In contrast to audio craftsmen, linguists are generally not interested in good sounding audio, they are only interested in completing their dissertation or the next published article which proves their theoretical hypothesis or generates points with their tenure and promotion board. Audio artifacts (recordings) are not primary products of interest to a linguist, rather it is what is done with them that is of their concern. This whole approach to language recording militants against good audio recordings.
One linguist with whom I collaborated, when asked how they wanted to manage the recordings (both the technical aspect of production, and the file system organization of the recordings) was not concerned with either. That person was (at the time) ambivalent because their advisor was advising the PhD dissertation not some fieldwork project. That linguist’s goal wasn’t to produce a body of work (recordings) which could be used beyond the dissertation, the goal was to do enough fieldwork to complete the dissertation.
From the advisor’s perspective, as long as the fieldwork produced some written data to present in the dissertation the core requirements were satisfied. I followed the career of that linguist and now they are in a post-doc at a reputable organization. The post-doc was advertised pre-COVID and included funding for fieldwork. However, due to COVID this linguist has relied on their prior field recordings in (presumably) unexpected ways. The organization and quality of the recordings impacts their productivity as they now need to reference those recordings. Within the context of the current post-doc and the global climate, job performance productivity is not evaluated as being above average or below average. However, the career building blocks for the next post-doc or position is impacted, by how the fieldwork situation and production process was managed, and the quality of the product produced. I notice that career scholars produce about four publications a year in addition to the grants that they write and win. This is a very different interactive context than negotiating interpersonal relationships on a funded project (in a lab or in a village).
As a audio craftsman, I use an an MMaudio DPSM-SIL mic or one of two Røde M3 mics with a Zoom H4n digital audio recorder. The MMaudio mic is a special set up of the mmdpsm mic with longer cords (7 feet). To order it one must call the company because they don’t put the part on-line. I think there are a few other tweaks with the DPSM-SIL and the regularly advertised DPSM.
Do not use Phantom power with the DPSM-SIL there is a setting through the menu of the Zoom H4n that says “plug-in power”. Turn this on. Also be careful of plugging in the mic into things like computers. I do not know what specific computer’s mic jack setting are but they might over power the mic and damage it.
There are technical reasons related to mic sensitivity and faithfulness to the input which lead me to choose these two mics in 2008. However, for me the ability to plan how audio should sound and the ability to create “three dimensional texture” within the audio recording is more important than a specific microphone2.
As a craftsman I also suggest the use of headphones because live recordings need active monitoring. Hearing the recording is import to the production production process — what is on the recording is not what is heard in the recording environment. See my write up on headphones I have used here: https://hugh.thejourneyler.org/2011/headphones-for-language-documentation
I found a different audio recording in a language archive which demonstrates the “air” and high-pitch noise induced into the recording. Presumably the high frequency pitch comes from the computer fan which is running nearby while recording. However, given the date of recording it might be the noise of the tape deck. The archive record does not make clear the original medium of recordings. From a sound engineer’s point of view, this is not great recording because the noise to signal ratio is really out of proportion. The recording is by Laurent Fontaine. Created in 2005. Titled: Histoire de Kanumá. https://doi.org/10.24397/pangloss-0000495 ↩︎
I also inevitiably get asked about USB mics and mics which plug into mobile phones or tablets. I find that these tools are not worth the money or the added risks they bring to workflow managment. There are some cool features and I have used plugin devices with iOS for recording as backups. The major issue with USB plugin mics is fan noise from the computer. People often look at podcasters and say “but they are using USB mics”. To which I say, “are you making a podcast, from the far bedroom in your home in the USA? What kind of post processing do you plan on doing?” I found a really well written article which discusses these issues: https://www.lowdownpublishing.com/post/musicians-guide-for-video-conferencing . ↩︎