Science-Based Voice Technology

BlackVOX is a technology-based company, incubated at the Sensory Research Laboratory (INIGEM, CONICET-UBA), which provides technological tools for hearing and speech applications.

BlackVOX offer various products and services for sound analysis, forensic identification of speakers, and speech recognition and synthesis. During its incubation, it received the support of the Ministry of Science, Technology and Productive Innovation through a FONCyT grant and the granting of a research and development project PID 35891 for the “Development of objective methods for speaker recognition” whose adopting agent was the Argentine National Gendarmerie.

The FORENSIA system of forensic identification of speakers was presented at the First National Symposium on “Science and Justice” organized by CONICET in December 2016 and has been used in the “Candela” case to collate voices. Since 2019, he has been employed by the Voice Collation Section of the Expert Advisory of the Judicial Power of the Province of Buenos Aires.

FORENSIA was also employed in the Fake News checkup. Reverse of Argentina requested an exhaustive investigation of an audio to verify if it was Guillermo Montenegro who was speaking in the piece that went viral. by Whatsapp. Other LATAM countries like Verified from Uruguay and Colombia Check from Colombia also requested similar checks.

Eng. Miguel Martínez Soler from BlackVOX, presented FORENSIA the forensic  speaker identification system  at the VII Argentine Congress of Justice and Technology that was held in the city of Corrientes on November 13, 2019.

Sound analysis

§ Creation of professional laboratories for voice analysis of normal and pathological speech.

§ Training associated with the interpretation of acoustic analysis based on models of speech production and perception.

§ Definition and visualization of the attributes of the voice and distinctive features of speech.

§ Practices in the creation of phono-audiological reports, aimed at ENT and Phono-audiology professionals.

Voice identification systems

§ Creation of forensic teams to identify speakers.

Automatic speech recognition systems

§ Database design.

§ Acoustic and language models of Spanish variants.

Text to speech conversion systems

§ Design of speech bases for the synthesis of particular voices.