## Doubletalk ## The Doubletalk corpus consists of recordings of paired speakers carrying out a variety of tasks in order to elicit a range of styles/registers in a discourse/interactive situation (map task, repetition, diapix, reading and shadowing). Ten speakers make up five pairs where one member is a speaker of Southern Standard British English and the other member is a speaker of Scottish English. There are also recordings of the one pair pilot study where one member is a speaker of American English and the other member is a speaker of Northern British English – these are task aware linguists. Alignment was carried out at Glasgow as part of the SPADE project using MFA (note therefore that there will be some American/British English mismatches in transcription). **Number of Speakers:** 12, 7F \ **Hours of Speech:** about 7 \ **Year Recorded:** 2010 \ **Data Guardian:** public, available at http://espf.ppls.ed.ac.uk/ \ **Speaker Dimensions:** age, gender, accent, accent descriptor #### Corpus references: #### Scobbie, J.M., Turk, A., Geng, C., King, S., Lickley, R., & Richmond, K. (2013). The Edinburgh Speech Production Facility DoubleTalk Corpus. Proceedings of 14th Interspeech, Lyon. \ Geng, C., Turk, A., Scobbie, J.M., Macmartin, C., Hoole, P., Richmond, K., Wrench, A., Pouplier, M., Bard, E.G., Campbell, Z., Dickie,C., Dubourg, E., Hardcastle, W., Kainada, E., King, S., Lickley, R., Nakai, S., Renals, S., White, K., Wiegand, R. (2013). Recording Speech Articulation in Dialogue: Evaluating a synchronized double Electromagnetic Articulography Setup. Journal of Phonetics, 41, 421-431.