----------------------------------------------------------------------------------------------- AVOID Age-related Voice Disguise corpus http://cs.uef.fi/~rgonza/avoid.html (Version 1.0) RELEASE January 2019 School of Computing University of Eastern Finland Copyright (c) 2019 Tomi Kinnunen tkinnu@cs.uef.fi, tkinnu@uef.fi Rosa Gonzalez Hautamäki rgonza@cs.uef.fi ------------------------------------------------------------------------------------------------- OVERVIEW This dataset includes speech data uttered by 60 Native Finnish speakers (31 females and 29 males). Each speaker reads 78 sentences (66 Finnish, 12 English). The sentences correspond to Finnish translations of The rainbow passage and The north wind and the sun, and two selected English sentences from TIMIT[1] corpus (SA1, SA2). "The rainbow passage" and "The north wind and the sun" text in Finnish language are: Sateenkaaritarina (The rainbow passage) Kun auringonvalo osuu sadepisaroihin ilmassa, ne käyttäytyvät kuin prismat, ja muodostavat sateenkaaren. Sateenkaari muodostuu valkoisen valon jakaantuessa useiksi kauniiksi väreiksi. Nämä muodostavat kauniin pitkän kaaren horisontin yläpuolelle päättyen jonnekin sen taakse. Legendan mukaan sateenkaaren päässä on padallinen sulaa kultaa. Ihmiset etsivät sitä kuitenkaan mitään löytämättä. Kun joku etsii jotain mahdotonta, sanotaan hänen etsivän kultaa sateenkaaren päästä. Pohjantuuli ja aurinko (The north wind and the sun) Pohjantuuli ja aurinko väittelivät kummalla olisi enemmän voimää, kun he samalla näkivät kulkijan, jolla oli yllään lämmin takki. Silloin he sopivat, että se on voimakkaampi, joka nopeammin saa kulkijan riisumaan takkinsa. Pohjantuuli alkoi puhaltaa niin että viuhui, mutta mitä kovempaa se puhalsi, sitä tarkemmin kääri mies takin ympärilleen, ja viimein tuuli luopui koko hommasta. Silloin alkoi aurinko loistaa lämpimästi, eikä aikaakaan, niin kulkija riisui manttelinsa. Niin oli tuulen pakko myöntää, että aurinko oli kuin olikin heistä vahvempi. The selected TIMIT sentences in English: SA1:"She had your dark suit in greasy wash water all year" SA2:"Don't ask me to carry an oily rag like that". The speakers read the 78 sentences in two recording sessions. In each session, the speakers performed the following tasks: Read 13 sentences in modal voice. Read 13 sentences attempting an elderly voice. Read 13 sentences attempting a child voice. The age range is from 18 to 73 years. All the participant signed a consent form to allow the use of their speech for reasearch purposes. RECORDING INFORMATION The speech was recorded with Zoom H6 Handy recorder, at a sampling rate of 44.1 kHz and 32 bits precision using a omnidirectional headset microphone (Glottal Enterprises M80). Parallel recordings were collected using apps in two smartphones: Nokia Lumia 635 (The Sound Recorder app) and Samsung Galaxy Trend 2 (Smart voice Recorder app). All recordings were collected in a silent semi-anechoic lab of the School of Computing, University of Eastern Finland. AIMS This dataset was collected to study the effect of voice modifications caused by the speaker in speaker verification carried out by machines [2] and human listeners [3]. COPYING You are free to use this dataset under CLARIN RES (Restricted)+BY+NC+PRIV+NORED+DEP1.0 The Copyright holder grants the End-User a free, non-exclusive and perpetual (for the duration of the copyright) right to use and make copies of the Resource for personal use as such, as modified, or as part of a compilation or derived work. The permission applies to all known or future modes and means of communication and includes a right to make modifications enabling the use of the Resource on other devices and in other formats. Additional license terms as defined in the Terms of Service Agreement: Identification and Access conditions ID: The user needs to be authenticated or identified. PLAN: The right holder requires a research plan for granting access. General Use conditions BY: Attribution, i.e. acknowledgement of authorship, is required. NC: The content is available for non-commercial purposes only. PRIV: There are personal data in the resource. Distribution conditions NORED: The user is not permitted to redistribute the resource. DEP: The user may distribute derivative works via CLARIN. This license has been made in compliance with copyright agreements by WIPO – the World Intellectual Property Organization. The rights granted in this license shall be so interpreted that in case applicable intellectual property laws grant rights not mentioned in this license, they are also regarded as part of the rights to be licensed; the purpose of this license is not to restrict any rights intended to be licensed within different legal systems. Additional rights to the Resource may be agreed separately in writing. THIS DATASET IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS DATABASE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ACKNOWLEDGEMENTS The dataset was constructed by: Tomi Kinnunen (1) Rosa Gonzalez Hautamäki (1) Md Sahidullah (1) Ville Hautamäki (1) Stefan Werner (2) Maria Bentz (2) Affiliations: (1) School of Computing, University of Eastern Finland (2) School of Humanities, University of Eastern Finland This research work was supported by the Academy of Finland (projects no. 253120 and 283256) and the Finnish Scientific Advisory Board for Defense (MATINE) project no. 2500M-003. REFERENCES [1] J. Garofolo, L. Lamel, W. Fisher, J. Fiscus, D. Pallett, N. Dahlgren, V. Zue. 1993. TIMIT acoustic-phonetic continuous speech corpus LDC93S1. Web Download. Linguistic Data Consortium, Philadelphia. [2] R. Gonzalez Hautamäki, Md Sahidullah, T. Kinnunen, V. Hautamäki, "Age-Related Voice Disguise and its Impact in Speaker Verification Accuracy", Speaker Odyssey 2016, Bilbao, Spain, June, 2016. [3] R. Gonzalez Hautamäki, Md Sahidullah, V. Hautamäki and T. Kinnunen, "Acoustical and perceptual study of voice disguise by age modification in speaker verification", Speech Communication, Volume 95, December 2017, Pages 1-15, doi: doi.org/10.1016/j.specom.2017.10.002