BlackVOX is a company that based on Research and Development (R&D) develops those new technologies that the market requires. It is nourished by its own investigations and those carried out in the Sensory Investigations Laboratory (LIS) belonging to INIGEM-CONICET, which is based at the Hospital de Clinicas «Gral. San Martín ”of the University of Buenos Aires.

New publications

Can we identify a suspect by his voice?

By Pedro Univaso, published in La Plaza Redonda on Jan 11, 2015

PC y micrófono

In Argentina, cases of identification of suspects by voice that are requested by the justice system are processed by the Argentine Federal Police and the Argentine National Gendarmerie. The Argentine Federal Police, through the Voice Identification Cabinet, belonging to the Scopometry Division of the Scientific Police Superintendency, employs a semi-automatic methodology through collaboration that provides the expert with an interactive analysis and processing system of acoustic signals developed by Dr. Sergey Koval. plus…

FORENSIA, a software prepared to check fake audios of Whatsapp, Youtube and Twitter

Pedro Univaso. Posted in Research Gate on Sep 29, 2019

The task of fact checkers in journalism is to confirm and verify facts and data that are used in speeches (especially politicians) and in the media and other publications. The purpose is to detect errors, inaccuracies, simulations and lies (fake news), although nowadays we are beginning to speak of misinformation (misinformation) instead of false information, since the interest is focused on lies as voluntary acts that they try to generate a wrong vision of reality. The British media report «Cairncross Review» describes fake news as «disinformation [understood as] the deliberate creation or dissemination of false or manipulated information that seeks to deceive or mislead audiences, whether for the purpose of cause harm, or to achieve a political, personal or financial return ”. The other variable of importance at this time is the viralization of disinformation through social networks, especially Facebook and WhatsApp in Argentina. The main success of a fake news lies in the number of times it is played and the number of «impressions» it causes. The main format of fake news is text, although they begin to appear in videos and audios (Whatsapp, Facebook and Twitter). The category of fake videos have escalated in such a way that they already have a particular name: deep fakes, due to the deep learning algorithm that generates them. FORENSIA is a system of forensic identification of speakers that is beginning to be used for news checking. See more…


Jorge Gurlekian, Pedro Univaso and Miguel Martínez Soler. Posted in Research Gate on May 23, 2018

The forensic expert Daniel Corach from Argentina has developed a guide (Corach, 2018) for the reception of DNA samples, where he states that “Legal assistants and acting experts are usually unaware that there are procedural guidelines that must be followed to ensure that the results of forensic expertise have evidence value at the time of trial and are not declared invalid due to errors in procedures. Upon receiving an indication of expert interest, a series of essential requirements must be fulfilled that will allow granting evidence value to the results of the studies derived from its analysis. ” Based on the aforementioned guide, on the foundations of forensic identification of speakers by voice (Rose, 2002) and on the recommendations (Drygajlo et al, 2015) of the European Network of Institutes of Forensic Sciences (ENFSI), it is presented here a guideline protocol for the stages involved in voice expertise ranging from the receipt of evidence to the creation of the expert report. See more…


  1. Determining the Likelihood Ratio From Perceptual Attributes of Voice. Jorge A. Gurlekian, Stefania Suligoy, Pedro Univaso, Humberto Torres, Evangelina Masessa, and Nancy Molina. Journal of Voice (2022).
  2. A multiparametric system for forensic speaker comparison. Eugenia San Segundo, Pedro Univaso, Jorge Gurlekian. Revista de Estudios de Fonética Experimental XXVIII, ISSN 1575-5533, pp. 13-45 (2019).
  3. Hacia la evaluación perceptual forense. Transformación de puntajes de similitud a relaciones de verosimilitud. Parte 1: hablantes masculinos. Nancy Molina, Stefanía Suligoy, Evangelina Masessa, Humberto Torres, Pedro Univaso, Jorge Gurlekian. Gaceta Internacional de Ciencias Forenses, Nro. 42, Universitat de València ISSN 2174-9019 (2022).
  4. Hacia la evaluación perceptual forense. Transformación de puntajes de similitud a relaciones de verosimilitud. Parte 2: hablantes femeninos. Evangelina Masessa, Stefanía Suligoy, Nancy Molina, Humberto Torres, Pedro Univaso, Jorge Gurlekian. Gaceta Internacional de Ciencias Forenses, Nro. 42, Universitat de València ISSN 2174-9019 (2022).
  5. Grabaciones indubitadas y dubitadas en las pericias forenses de voz: un trabajo preliminar. Pedro Univaso, Jorge Gurlekian. Research Gate (2018)
  6. Cómo presentar la evidencia científica a la comunidad judicial: factor de Bayes. Univaso, P. Research Gate (2018)
  7. Towards a Unified Methodology for Forensic Speaker Identification. Univaso, P. Research Gate, DOI 10.13140/RG.2.2.31571.58402  (2016)
  8. Identificación forense de hablantes en Argentina: un tutorial. Univaso, P. ResearchGate, DOI 10.13140/RG.2.1.4252.3768 (2016)
  9. Data Mining applied to Forensic Speaker Identification. Univaso, P., Ale, J. M., y Gurlekian, J. A. Latin America Transactions, IEEE (Revista IEEE America Latina), 13, 4, pp. 1098-1111 (2015)
  10. El alfabeto fonético SAMPA y el diseño de córpora fonéticamente balanceado. Gurlekian, J.A., Colantoni, L y Torres, H. Fonoaudiológica. Asalfa. 47,3, 58-69 (2001)
  11. Comparison of Two Perceptual Methods for the Evaluation of Vowel Perturbation Produced by Jitter. Gurlekian JA, Torres HM and Vaccari ME. Journal of Voice. Elsevier, doi:10.1016/j.jvoice.2015.05.009 (2015)
  12. A preliminary approach to forensic speaker recognition using phonemes. Univaso P., Martínez Soler M., y Gurlekian J. A. En IberSPEECH 2012, VII Jornadas en Tecnología del Habla and III Iberian SLTech Workshop, Editorial: Escuela Politécnica Superior, Universidad Autónoma de Madrid, España (2012)
  13. Human Assisted Speaker Recognition Using Forced Alignments on HMM. Univaso, P., Martínez Soler, M., y Gurlekian, J. A.. International Journal of Engineering Research and Technology, ESRSA Publications, 2, 9 (2013)
Anuncio publicitario