BLB home BLB home

Digitization of the Borror Lab’s Animal Sound Archive

Background

Donald Borror, the founder of the Borror Laboratory of Bioacoustics, made over 14,000 recordings of animal sounds from 1948 through the 1980s. The medium that was available then was, of course, analog tape.

collection.jpg (12728 bytes) In a project funded by the NSF, the BLB copied Borror’s entire collection of recorded sounds from 1/4" analog tape to recordable compact discs (CDR). Multiple considerations prompted this decision. First, we needed to preserve the collection. Many of the recordings were approaching 50 years of age, which is about the limit for analog magnetic tape. Other recordings, made in the 1980's, were on tape stock subject to "stick-shed," an undesirable state in which the tape is unplayable unless it has been baked in an oven to drive off absorbed water. Another reason to digitize these recordings was that as the audio world has embraced digital technology, it has become more difficult and expensive to find and maintain analog equipment. Finally, storage of the collection in a digital format offers several advantages. The CD medium offers fast, random access to the data on a disc, unlike a tape which has to be wound to reach the desired location. CDs promise potential great longevity, as the surface is read with a laser beam and is not subject to the wear and tear of being dragged repeatedly over record/reproduce heads as tape media are. The large storage capacity of CDs also facilitates working with and storing the collection.

Overview of the process

Borror’s collection consisted of many reels of 1/4" analog tape, organized by species, with each separate recording (or cut) separated by segments of leader tape. This species organization of tape reels was an innovation of Donald Borror’s. The recordings for some species were housed on single 7" diameter reels, while others, such as the song sparrow, required well over 100 reels.

rtsdscreen.jpg (7473 bytes) During digitizing, a technician played the desired reel, one cut at a time, on a Studer studio tape deck connected into a computer interface. We used an instrumentation-quality analog-to-digital converter (16 bit amplitude resolution, 50,000 samples per second) and RTSD software to convert each cut to a separate, numbered computer file. A scrolling spectrogram (see figure) was monitored as each cut was digitized. This helped avoid signal “clipping.”

Other custom-designed software was used to verify the quality of the digitization process and to add certain information (recording date, location, etc.) from the collection's database into a "header" associated with each digital file. After a sufficient number of cuts were digitized to fill one CD (about 530 MBwe avoided filling CDs to maximum capacity because of the greater error rate due to “wobble” at the outer edges of CDs), the files were moved to a second computer containing the CD writers. We burned two identical CDs simultaneously. One serves as a working copy in the collection, and the other is archived off-site.

Quality control

scorch.jpg (13782 bytes) The short outline above does not convey the full complexity of the process. Errors, both human and machine-caused, can intervene at many steps in the process. We took many steps to minimize, and hopefully, eliminate, errors. The tape recorder and computer digitizer were calibrated and checked at frequent intervals. The technician monitored each cut both aurally and visually (on a real-time audiospectrogram), and compared the catalog data on the tape box and leader to that in the computer database. Each CD was analyzed immediatedly after burning for recording errors using a Clover Systems CD analyzer. Selected CDs are now monitored at three-month intervals to try to identify any degradation in the media over time. Should this occur, we will transfer the data to a new CD (or other digital medium if CDs are replaced in the future). To guard against the possibility of catastrophic loss of CDs, we also made two copies of each CD image on digital tape. The original analog collection has been archived under controlled conditions in The Ohio State University archives. Recordings made prior to 1960, and those suffering from "stick-shed," were copied onto new analog tape stock before being archived. Thus, at the conclusion of the project, we had five complete copies of the collection: two on CD, two on DAT tape, and one on 1/4" analog tape. We have also placed copies of all recordings onto a hard drive in the lab, and these are the copies we now access most frequently.

This material is based upon work supported by the National Science Foundation under Grant No. DEB-9613674. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Borror Laboratory of Bioacoustics
The Ohio State University
1315 Kinnear Road
Columbus, OH 43212-1192   USA

Phone: (614) 292-2176
Fax: (614) 292-7774
e-mail: borrorlab@osu.edu