Re-usability of recorded images and sound

One of the potential influences that preparation for later re-use may have on data-originators is the awareness that others may encounter difficulty in understanding one’s image and sound records. As they are more intrinsically ambiguous than written records, researchers may be quite wary of depositing them (quite aside from the considerable ethical issues involved).

One of the potential dangers here is that originators may be too aware of aesthetics when producing a video record, so that the quality of the image assumes greater importance than what is actually being represented. The originator may select shots on the basis that they ‘work visually’ rather than purely on the basis of the study’s methodological needs. Once a dataset is already constructed, the originator may feel the need subsequently to ‘tidy up’ or edit these records so that they become more ‘readable’.

Technique and technology

There is a purely practical issue here. In order for image and sound based records to be usable by others, the quality does indeed have to be sufficiently high to allow later re-users to decode the archived information. Poorly produced sound recordings, incompetent photographs and wobbly video footage may be decodable by originators due to their field-recollections and experience-derived insights. But they may be frustratingly opaque to re-users. Written records, by contrast, do not involve such a potential range in quality.

This issue involves questions both of technique and technology. On the one hand, we would advise that anyone intending to use audio-visual media in the field undergoes some preliminary training in how to use a camera effectively. There are short courses available at a range of organisations and institutions. On the other hand, we would also recommend that he or she obtains professional advice about what kind of equipment to purchase. One of the perennial problems is ensuring sufficiently high and consistent sound quality, so equipping oneself with good microphones is essential. This usually means buying a video camera that can have external mics attached, which is likely to increase the price accordingly. We advocate using semi professional models rather than consumer-level ones, as they enable settings to be customised, sound to be monitored and more than one external microphone to be attached.

There is a set of guidelines on choosing good audio recording equipment on the Qualidata website.

Viewability and when to use the camera

The other issue is more of a methodological one: how to decide which recording medium to use: when to use a visual recording medium? When to use only writing? When to use fieldnotes?

• How should we deal with the demands of aesthetic ‘viewability’ in deciding how to record diverse data?

For example, in a complex and fast-moving fieldwork setting (such as a school, a hospital or a science museum), one might well have to choose between observing one event over another. If one event is more amenable to the camera, i.e. more ‘visual’, does this mean we should make our choice based on a ‘viewability’ criterion? This would clearly be problematic. Choosing what to film should be decided on methodological/theoretical grounds, not on grounds of its visuality.

• How should we judge the merit of using visual recording media in the context of re-use?

In our own work we decided to select the recording medium that best suited the diverse kinds of data we were dealing with. For example, for live interactions between users and exhibits, cameras were ideal. But in sit-down interviews with the science museum staff, we most frequently used audio recordings alone, and decided not to use video. In terms of our particular research questions, there was little to be gained in terms of qualitatively important supplementary information by the use of video in interviews. This may not, however, be the case with other studies, and – crucially - studies that re-use existing data, where the nuances of participants’ gesture, facial expression, posture, and so forth might be of greater significance.

To Exemplar 4-->