next up previous contents
Next: Check List Distribution Up: Distribution Previous: Larger Edition vs. Burn-on-Demand   Contents

On-line Distribution

Smaller speech corpora may also be distributed on-line, for instance by a password protected FTP server. Using an appropriate database system it might even be possible to distribute parts or excerpts of a speech corpus. For instance a prospective user might only be interested in the female speech of a large corpus, or even more specificly, only in certain spoken words that might be indexed via a word segmentation of the corpus.

Distribution servers of this kind do already exist for special scientific speech data and are usually free to use. They require a considerable effort to set up and maintain.

For speech resources that are not absolutely freely available there are still many practical and legal problems to solve.

We recommend allowing the free download of the meta data and perhaps also of the annotation data of a speech corpus. Meta data are essential for prospective users to help them decide whether a speech resource meets their special needs. Annotation data are in most cases of not much use for commercial users without the corresponding signal data, but they might be of academic interest.


next up previous contents
Next: Check List Distribution Up: Distribution Previous: Larger Edition vs. Burn-on-Demand   Contents
BITS Projekt-Account 2004-06-01