DiapixUK test materials

Picture materials used for the elicitation of spontaneous speech dialogues between two speakers in a collaborative task (see Baker and Hazan 2011 BRM). Examples of recordings made using Diapix can be found here

The complete set of DiapixUK materials are archived in Zenodo (open access) and can be downloaded from there: either High-resolution versions of the pdf files (DOI:10.5281/zenodo.3703202)  or the original photoshop files (DOI: 10.5281/zenodo.3739053)

Some versions of the DiapixUK materials adapted for other languages (French, German, Spanish, Danish) can be found in the ‘Diapix task’ Zenodo community 

Short demo of how Diapix is used in our research

Speech corpora

The following corpora were developed as part of funded projects and are being made freely available to other researchers (for non-commercial purposes), but without further support. Please acknowledge the use of these materials for further studies with appropriate references.

UCL Speaker Database

A corpus containing audio recordings of a wide range of speech materials (words, sentences, read texts, picture descriptions) for 45 speakers of British English with a fairly neutral accent or mild South-Eastern English accent. These include 18 women (mean age: 33;11 yrs), 15 men (mean age: 30;7 yrs), 6 girls (mean age: 13;2 yrs) and 6 boys (mean age: 13;2 yrs). See further details here.

Reference

Markham, D. and Hazan, V. (2002) The UCL Speaker Database. Speech, Hearing and Language: UCL Work in Progress, vol. 14, p.1-17

This corpus was produced by Duncan Markham as part of a study funded by the Wellcome Trust (055651/ Z/98/ JRS/ JP/ JAT), PI: Valerie Hazan.

Access

This corpus is freely available for research purposes only. To get access to zipped folders containing the corpus, please contact Valerie Hazan

LUCID corpus

A corpus containing DiapixUK conversations between pairs of participants produced in good and challenging conditions as well as read sentences and picture naming. Speakers are 40 young adults (F=20) who are native southern British English. See further details here

Reference

Baker, R., & Hazan, V. (2011). DiapixUK: task materials for the elicitation of multiple spontaneous speech dialogs. Behavior Research Methods, 43 (3), 761-770. doi:10.3758/s13428-011-0075-y

This corpus was produced by Rachel Baker and Valerie Hazan as part of a study funded by the Economic and Social Science Research Council (RES-062-23-0681) in 2008-2011.

Access

This corpus is freely available for research purposes. Audio files (wav format), Praat TextGrids and picture materials are all downloadable from the Speechbox Resource at Northwestern University (click on UCL Diapix corpora in left-hand banner)

kidLUCID corpus

A corpus containing DiapixUK conversations in good and challenging conditions for 96 children and adolescents aged between 9 and 14 (46M, 50F, mean age: 11;8 years, range 9;0 to 15;0 years) who were native southern British English speakers. See further details here.

Reference

This corpus was produced by Michèle Pettinato and Outi Tuomainen as part of a study  funded by the UK Economic and Social Research Council (RES-062-23-3106) in 2011-2014 (PI: Valerie Hazan).

Hazan, V., Tuomainen, O. and Pettinato, M. (2016)  Suprasegmental  characteristics of spontaneous speech produced in good and challenging communicative conditions by talkers aged 9 to 14 years old. Journal of Speech, Hearing and Language Research, 59, S1596-S1607.  Hazan et al authors manuscript JSHLR 2016

Access

This corpus is freely available for research purposes.  Audio files (wav format), Praat TextGrids and picture materials are all available from the Speechbox Resource at Northwestern University (click on UCL Diapix corpora in left-hand banner). A spreadsheet containing key acoustic and other measures is available from the UK Data Service. Time-stamped orthographic transcripts are available here. 

elderLUCID corpus

A corpus containing DiapixUK conversations and read BKB sentences produced in good and challenging conditions for 83 Southern British English adult talkers. These included ‘older adults’ (OA) between 64-84 years of age (N=57; 30 F) and ‘younger adults’ (YA) between 19-26 years of age (N=26, 15 F).  OA participants were further subdivided into two groups according to their hearing status. OANH participants (N=27; 14F) had normal hearing while OAHL participants (N=30; 16F) had a mild acquired hearing loss.

Reference

This corpus was produced by Outi Tuomainen as part of a study  funded by the UK Economic and Social Research Council (ES/L007002/1) in 2014-2017 (PI: Valerie Hazan)

Hazan, V. L., Tuomainen, O., Kim, J., Davis, C., Sheffield, B., & Brungart, D. (2018). Clear speech adaptations in spontaneous speech produced by young and older adults. Journal of the Acoustical Society of America, 144, 1331-1346. doi/10.1121/1.5053218

Access

This corpus is freely available for research purposes.  Separate folders containing audio files (wav format) and  Praat TextGrids for the Diapix conversations and read BKB sentences can be requested from Valerie Hazan. A spreadsheet containing key acoustic and other measures is available from the UK Data Service.

Reference for acoustic dataset: Hazan, Valerie and Tuomainen, Outi and Kim, Jeesun and Davis, Christopher (2018). elderLUCID: London UCL Older adults’ clear speech in interaction database. [Data Collection]. Colchester, Essex: UK Data Archive. 10.5255/UKDA-SN-852906

LifeLUCID corpus

A corpus containing DiapixUK conversations in conditions of energetic and informational masking for 114 Southern British talkers aged between 8 and 80. These age bands were: 8-12 years (Young Children, CH-Y, M=10.34 years), 13-17 years (Older Children, CH-O, M=15.94 years), 18-29 years (Younger Adults, YA, M=21.82 years), 30-49 years (Young Middle Aged, MA-Y, M=42.98 years), 50-64 (Older Middle Aged, MA-O, M=59.30 years) and 65-85 years (Older Adults, OA, M=71.19 years). Each of the six age bands included 20 participants (10F) apart from the 13-17 band due to recruitment difficulties (N=14, with only 4 males).

Reference

This corpus was produced by Outi Tuomainen and Linda Taschenberger as part of a study  funded by the UK Economic and Social Research Council (ES/P002803/1) , PI: Valerie Hazan.

Tuomainen, Outi and Taschenberger, Linda and Hazan, Valerie (2021). LifeLUCID Corpus: Recordings of Speakers Aged 8 to 85 Years Engaged in Interactive Task in the Presence of Energetic and Informational Masking, 2017-2020. [Data Collection]. Colchester, Essex: UK Data Service. 10.5255/UKDA-SN-854350

Access

This corpus is available for research purposes. Separate folders containing audio files (wav format) and  Praat TextGrids are available from the UK Data Service.

Configurations files and masker files used in the study are available here.