Yesterday, I was fortunate to attend a fascinating lecture on audio preservation given by Thomas Reiger of the Northeast Document Conservation Center. The specific focus of the lecture was on a new preservation system known as IRENE, a digital imaging system that converts captured visual data (2D and 3D, depending on the analog format) and converts that data into listenable audio waveforms.
IRENE (short for “Image, Reconstruct, Erase Noise, Etc.”) first utilizes high-resolution 2D and 3D camera technology to safely document with extreme precision the physical state of analog audio discs and wax cylinders. Following this capturing process, the software creates an image map of the whole physical object and determines, based on the physical state of the grooves, what the raw audio waveform would be. This proves to be an extremely effective system when working with fragile or broken audio materials that otherwise would be destroyed through other transcription processes.
The output of the waveform, according to Reiger, is then presented in a “pure” archival state – meaning that little to no audio engineering treatments are applied by the NEDCC. This is an intentional decision on behalf of the organization, as they want to follow a digitization model that prioritizes cultural and material integrity. The end result, as can be heard on the IRENE Seeing Sound Blog, is remarkable – especially when considering that many of the original objects would be otherwise unplayable. (Of particular note is the Helen Hartness Flanders Archive, hosted by Middlebury College, and the work the NEDCC has done with the Woody Guthrie Archives.)
I asked Reiger if the NEDCC was investigating models of crowd-sourcing the engineering of these files, so that a range of desired interpolations could be offered to various audiences. This, to an extent, is being considered and the NEDCC is already working with engineers seeking to offer curated versions of the material. It will be interesting to see how the IRENE technology develops over time and how models of support might shift with a greater adoption of inter-media transcription technologies.