Audio Archival Methods

For a more technical review of our methods, see the article by Robert Grotke in RLG DigiNews Volume 8, Number 1.

Preservation - The collection storage room where audio and video media are stored is designed for that specific purpose. Limited access and specialized shelving coupled with rigid temperature and humidity control ensure an ideal environment for the long-term storage of magnetic tape, thereby maximizing the life of the recording medium.

Data - Aside from simply documenting who made a recording, identifying the species, and noting where and when the recording was made, we also store valuable information on the recording equipment used, detailed behavioral and habitat descriptions, and even individual animal identification when available. We also do some data enrichment, where information not originally identified by a recordist (such as call type or behavioral display) may be added to the record by an archivist using an authority in the field, such as the Birds of North America, or an expert researcher in the field.

Protocols for consistency - Over the years, the Macaulay Library has developed protocols to ensure uniformity in the way specimens are created and associated data are stored. The aim is that if an archivist from 1960, 2004, or 2040 were to be given the same field recording, all would generate nearly identical specimens. While advances in technology have and will alter and expand the way in which the work is done, the essence of the archived specimen should remain consistent.

Restoration - Over time, problems develop with recording media, both analog and digital formats. For example, tapes may become sticky and/or shed, resulting in the oxide (material the sound is stored on) separating from the tape backing, resulting in lost information that can never be reclaimed. Tape stocks may also dry out and become brittle making playback without breaking nearly impossible. We have the expertise and dedicated facilities on-site to temporarily correct many of these problems thereby allowing the digitizing process to be undertaken on otherwise unplayable recordings.

How does the digital archival process work?
Audio field recordings are ingested into a digital editing system at sampling rates of either 96kHz or 192kHz and a bit depth of 24 bits using either Prism Sound or dCS analog to digital converters. While these may seem like extraordinarily high sampling rates for many species, the resulting process provides unparalleled fidelity and signal resolution. Additionally they provide the capability to handle many of the ultrasonic signals produced by toothed whales, bats, and other wildlife.

Specimen creation and editing - After ingesting the raw field recording, the archivist uses standardized protocols for determining where one specimen ends, and another begins. Some example criteria used for these decisions are whether the recording equipment was turned off and for how long, or whether the species or the individual of interest during a recording changes. Field tapes are edited into audio specimens with a powerful and flexible digital sound-editing program, SonicStudioHD. After identifying the specimen on the original field recording, the archivist assigns a catalog number (adds a voice announcement of the catalog number for audio specimens), digitizes the desired audio segment, and, in the process also optimizes the amplitude of the sound of interest. Other than level optimization, archivists strive to maintain the specimen as it came from the field. No additional manipulation of the sound occurs and filtering is only used when amplitude level setting on a particular sound proves challenging due to low frequency interference such as wind or handling noise.

Data entry - The associated data record for each audio specimen is created by the archivist, often in parallel with digitization. In cases where the recordist has previously generated the data record, the archivist adds any extra information to the record. Using the recordist's notes, archivists enter detailed information on time and location, habitat description, behavioral context, equipment setup, and any other information available. Data are stored in a relational database, allowing powerful searches on multiple fields and text entries.

Output - The output of the digitization process is high resolution (96kHz/24-bit) AIFF audio files, that are written onto two, first-generation DVD-ROM discs. One DVD is loaded into a DVD jukebox system for local access, and the second serves as a safety copy that is sent to a secure, off-site, climate-controlled facility. The high resolution files are also down-sampled into more readily accessible and more easily distributed lower resolution formats such as .wav, Real, mp3, and QuickTime.

Storage/Delivery - DVDs are stored in a 12-jukebox array, with a 27 terabyte storage capability. This equates to approximately 25,000 hours of high-resolution monaural audio storage. Lower-resolution files are stored on a 25 TB Apple Xserve/Xserve RAID array, where they are currently accessible to staff on an internal network. Additionally, all lower resolution digital files are backed up via an Exabyte LTO tape library located off-site. The digitized portion of the archive will soon be available on-line to the public through a friendly graphic user interface that is currently under development.

Quality control - Quality control is a critical part of our archival process. Archival studio equipment is frequently calibrated to known International Standards. Internal operating protocols are in place to ensure consistent quality standards throughout the archival process. Additionally, archivists double-check each others work after archiving a body of recordings. The storage media (DVDs) are also quality control checked pre and post writing. All DVD-Rs are analyzed by using an AudioDev CATS DVD-R/RW Pro audio analyzer system to ensure that they meet or exceed industry standards. Discs that either fail or are of marginal quality are rejected. To monitor the stability and longevity of the digital collection DVD discs are randomly selected and re-tested. We then compare test data from initial disc creation to current tests looking for any signs of change or manufacturing batch related problems.