Speech & Math

Most of our research data is covered by restrictions imposed by the Research Ethics Board. However, parts of the research data that are not patient material have been compiled here for public use.

Currently available data set consists of the DICOM-datafiles and corresponding sound samples for all the finnish vowels. Some derivatives obtained from the image and sound data are also provided, this includes the surface models for the vowels.

The vocal tract anatomies of Finnish vowels, imaged from a 30 years old healthy male.

License

The released dataset is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.

Creative Commons License

Data

Subject 1

 
@article{PAAHMSV2012,
  author = {J.~Palo and D.~Aalto and O.~Aaltonen and R.-P.~Happonen 
            and J.~Malinen and J.~Saunavaara and M.~Vainio},
  title = {Articulating {F}innish Vowels: Results from {MRI} and sound data},
  journal = {Linguistica Uralica},
  volume = 48,
  number = 3,
  pages = {194--199},
  year = 2012,
  url = {http://www.math.hut.fi/~jmalinen/MyPSFilesInWeb/L-U12.pdf}
}
Download dataset for subject 1 (~25Mb)

Subject 2

 
@article{AaltoEtalLSDASMRIS2014,
  author = {Aalto, Daniel and Aaltonen, Olli and Happonen, Risto-Pekka 
            and Jääsaari, Päivi and Kivelä, Atle and Kuortti, Juha 
            and Luukinen, Jean-Marc and Malinen, Jarmo and Murtola, Tiina 
             and Parkkola, Riitta and Saunavaara, Jani and Soukka, Tero 
            and Vainio, Martti},
  title = {Large scale data acquisition of simultaneous MRI and speech}, 
  journal = {Applied Acoustics},
  volume = 83,
  number = 1,
  pages = {64--75},
  urldate = {2014-03-10},
  year = {2014},
  url = {http://www.math.hut.fi/~jmalinen/MyPSFilesInWeb/MTR_AA_rev1.pdf}
}
Download dataset for subject 2 (~883Mb)

© 2013 Comspeech@Aalto