In this paper, we describe the best practices for digitizing speech recordings for archival purposes. Many current guidelines do not give adequate rational and are not specific enough about audio digitizing practices for spoken word digital repositories. By analyzing frequency response, dynamic range, formant bandwidths, noise, and psychoacoustic, and perceptual quality, spectral evaluation of recordings was used to develop digitization best practices. Given the available technologies and the demands of archival preservation, weighed against expense and storage, the following digitization practices were determined acceptable for digitizing spoken word resources; sample rate: 96,000 Hz; bit-depth: 24-bit; and oversampling delta-sigma A/D converter with dither added prior to sampling. Best practices for digitizing equipment and digital format are also described.
- Discusses digital audio basics and hardware
This tutorial is intended to assist researchers in implementing several important aspects of audio technology in the field, studio, and research lab. It presents a set of best practices in the recording, processing, and analysis of acoustic speech signals. The author has no commercial interest in recommending particular pieces of hardware or software. Some brand names and manufacturers have been mentioned to help researchers find appropriate tools more easily.