ISCA Archive Eurospeech 1989
ISCA Archive Eurospeech 1989

Use of CD-ROM for speech database storage and exchange

John S. Garofolo, David S. Pallett

Speech databases have become important resources for the speech research community. Experience with the use of magnetic media for exchange of these databases has proven to be unsatisfactory. The use of CD-ROM media for speech database storage and exchange is attractive for a number of reasons. At the National Institute of Standards and Technology (NIST), we have produced a prototype CD-ROM version of the DARPA TIMIT Acoustic- Phonetic Speech Database. This paper discusses issues involved in producing this disc and future CD-ROMs, including CD-ROM standards and portability, directory and filename organization, speech file headers, and speech data formats.