Speaker and Producer

The next relationship to consider is that between the producer of a speech corpus and the speakers. A speaker is recorded for the purpose that his/her voice is used for scientific investigations or the development of new technology. Basically, every speaker (if he is no public figure) has the right to decide whether recordings of his voice, knowingly or unknowingly, are stored on a media and distributed to others. Therefore it is of paramount importance for the producer of the corpus to clarify the legal situation with his speakers before he starts the recording.

In most countries it is sufficient to advise the speaker about the purpose of the recording and what will happen with the data in the future (data protection, anonymity) and then let him sign a short declaration in which he waives his rights to the recorded data and explicitly agrees to the intended usage of the data. This might be not sufficient in your country; please contact a legal advisor about this.

It might be also an good idea to include in this declaration information about the anonymity granted for the speaker. If you are planing to produce more than one speech corpus, you may also ask the speaker on the signed declaration whether he/she would like to participate in future recordings and whether the producer is allowed to store the speaker's contact information for this purpose.

Legal problems may arise in the case of Wizard of Oz (WOZ) recordings (see also section , p. ) where the speakers might not be aware that they are being recorded.

For details see [2], pp. 143 - 145.