The speech and language processing division at British Telecom Research Laboratories (BTRL) has extensive experience in collecting databases which range in size and complexity and have been recorded in different environments. This paper will describe many of the options which are available when embarking on a database recording. Suggestions for helping to make the correct choice amongst those options will be made. The contribution will conclude by describing the collection and processing of the CONNEX Alphabet Database, a database collected for research into neural networks, which is available (at BT's discretion) on 5.25" optical disk format to other researchers in the field.
Cite as: Walker, G.P., Millar, W. (1989) Database collection: experience at british telecom research laboratories. Proc. Speech Input/Output Assessment and Speech Databases, Vol.2, 71-74
@inproceedings{walker89_sioa,
author={G. P. Walker and W. Millar},
title={{Database collection: experience at british telecom research laboratories}},
year=1989,
booktitle={Proc. Speech Input/Output Assessment and Speech Databases},
pages={Vol.2, 71-74}
}