Skip to main content

EU Funded · Next Generation EU · UPF · RAC 2026 · LREC 2026 · +RAIN 2026

Signària

Sonic identity for deaf signers: personalized synthetic voice for radio broadcasting

What is Signària?

Signària — «Identitat sonora per a persones sordes signants: creació d'un prototip de veu sintètica personalitzada per a l'àmbit radiofònic» — is a research project funded by Next Generation EU and led by Universitat Pompeu Fabra (UPF). The goal: to create a personalized radio sonic identity for deaf sign language users.

The radio broadcasting and podcasting ecosystem has historically been structured around a phonocentric paradigm that systemically excludes deaf signers. Signària subverts this model by creating personalized — not generic — synthetic voices that represent each person's real identity, providing them with their own unique acoustic signature.

The project combines AI voice synthesis technologies with Catalan Sign Language (LSC) processing, creating a technological bridge between the deaf signing community and radio broadcasting media. The voice is not merely a utilitarian vehicle: it is a first-order human identity marker.

Xavi Vinaixa argues that Signària's technology must be developed «amb mirada ètica, humanista i sobirana»

RAC1 · «Via lliure» · 24 May 2026

Radio sonic identity

The core concept of Signària is the creation of a specific, unique, and non-transferable «sonic identity» for each deaf signer. This sonic identity acts as an interface that enables users to interact fluidly in exclusively auditory communicative spaces — radio, podcasts, media — without depending on human interpreters or robotic, generic synthetic voices that dilute personality and individual agency.

The system is capable of building a personalized vocal fingerprint through the concept of sonic inheritance and voice donors: hearing individuals who lend the acoustic characteristics of their voice so that the deaf person can have their own sonic identity. Decisions about timbre, prosody, rhythm, and fundamental frequency are made sovereignly by the very person who will use it as their interface with the hearing world.

The consortium

Universitat Pompeu Fabra (UPF) serves as the academic institution responsible for applied research, scientific validation, and knowledge transfer. The project is housed in the Department of Communication at Campus Poblenou, with the research groups POLCOM-GRP (Political Communication, Journalism and Democracy) and GLiF (Formal Linguistics).

Sorensen.ai handles the overall coordination and technological development through its AI division, translating complex linguistic and sociological requirements into scalable software architectures and machine learning models. The Laifari association collaborates on the collection, labeling, and structuring of the LSC corpus.

Technical architecture

The technical team has designed and implemented a hybrid architecture that fuses two paradigms: rule-based symbolic systems (deterministic NLP) for the initial processing of LSC — which possesses a three-dimensional grammar with spatial and facial markers — and Deep Learning models (Neural TTS) that transform linguistic representations into Mel spectrograms, decoded by neural vocoders into audible sound waves.

Engineering focuses on optimizing computational latency to ensure real-time voice synthesis, necessary for live radio broadcasts. This includes data ingestion pipeline design, digital signal processing (DSP), and the parameterization of foundational Text-to-Speech models.

Technologies

Hybrid Architecture (Rules + Deep Learning) Neural TTS (Text-to-Speech) AI Voice Synthesis Catalan Sign Language (LSC) Natural Language Processing (NLP) Personalized Radio Sonic Identity Digital Signal Processing (DSP) Neural Vocoders Real-Time Latency Optimization

Funding and institutional context

Signària is funded by Next Generation EU from the European Union, with institutional support from the Generalitat de Catalunya (grant SDC007/25/000093). The project is led by Universitat Pompeu Fabra (UPF) in Barcelona, from the Department of Communication at Campus Poblenou.

This level of funding underscores the strategic importance that public administrations place on digital inclusion, linguistic heritage preservation, and the development of AI systems that operate under strict ethical frameworks.

Milestone: March 2026

In March 2026, the project consortium officially presented a fully operational tool that makes viable the direct presence of deaf signers in conventional radio broadcasts. This presentation marks the transition from the research prototype to the technology transfer phase, enabling real adoption in the media market.

Recognition · Ràdio Associació de Catalunya · May 2026

Honor Mention at the 26th Ràdio Associació Awards

The 26th Ràdio Associació de Catalunya Awards have distinguished Signària with an Honor Mention. The jury recognises the project as «a proposal that facilitates the participation of deaf people in the radio environment through the generation of a synthetic voice, without having to renounce their own language, Catalan Sign Language». The ceremony took place on 21 May 2026 at the Sala La Paloma in Barcelona, where Xavier Vinaixa argued that «Signària is an example of what artificial intelligence should be: a humanist, diverse and inclusive vision with European values», as reported by La Local (Tiana).

It is the first institutional recognition from the Catalan radio sector of the project's technology and consolidates Signària as a benchmark in accessibility, linguistic inclusion and ethical innovation in media. Xavier Vinaixa reflects on it in his article «Another AI is possible», where he explains how Signària proves that AI can be built from humanism and ethics instead of surveillance.

Acceptance speech · Sala La Paloma · 21 May 2026

The acceptance speech

I'm thrilled to share the short speech we gave when collecting the Honor Mention for the Signària project.

Recognition of the engineering and research work is always appreciated, but if I'm sharing this clip it's for a far more human reason: the powerful platform this stage has given us.

Going up to collect this award was a golden opportunity to give visibility to a reality the system too often pushes to the margins. The Deaf community exists, is part of our diversity and has a voice of its own. And that voice is expressed through Catalan Sign Language (LSC), an immensely rich heritage of our country that we have a duty to protect, nurture and defend with the same strength and determination with which we defend Catalan.

Signària aims to be the example of the artificial intelligence I believe in: one that rejects extractivism, that holds a deeply humanist outlook, and that is built not for people, but with people.

Sala La Paloma, Barcelona · 21 May 2026 · 1 min 32 s · Catalan subtitles

«Because now, on the radio, you can sign too.»

Speech transcript (Catalan)

Menció d'honor per a Signària, projecte de la Universitat Pompeu Fabra i Sorensen. El que escolteu és la meva identitat radiofònica. S'ha construït a partir d'una donant de veu de la meva família i el meu timbre real.

Xavi Vinaixa, Marcel Mauri, Roger Cassany i Gemma Barberà. Bona nit. Moltes gràcies, Ràdio Associació. M'acompanyen el Roger Cassany i el Marcel Mauri del grup de recerca PolCom de la UPF i la Gemma Barberà del laboratori de llengua de signes catalana de la UPF.

La llengua de signes catalana s'ha de protegir i cuidar com fem amb el català. És patrimoni d'aquest país. Signària és l'exemple del que hauria de ser la intel·ligència artificial: mirada humanista, diversa, inclusiva i amb valors europeus. Investigadors en comunicació i periodisme, lingüistes i informàtics treballant amb la comunitat sorda signant. No per ells, sinó amb ells.

Gràcies a tot l'equip de Sorensen, de la Universitat Pompeu Fabra i sobretot a la comunitat sorda que ens ha confiat la seva veu. I gràcies a la Berta, que l'heu vist en imatges, que és la primera estudiant sorda signant del grau de Periodisme de la Universitat Pompeu Fabra, universitat pública. Ella ha estat, de fet, el motor i la inspiració de tot plegat. Sense ella no hauria estat possible. Gràcies.

Ara, la ràdio, també signa. Moltes gràcies.

Academic presence

The Signària team has participated in numerous academic conferences presenting the project's advances in international forums on language technology, accessibility, and artificial intelligence. The paper "Identidad sonora personalizada mediante IA para personas sordas signantes" (Roger Cassany, Xavier Vinaixa, Marcel Mauri) was published in the Proceedings of the XVII International Latin Congress on Social Communication (CILCS 2025, Madrid — ISBN 979-13-87819-03-3, DOI 10.15178/CILCS_2025).

In May 2026, the team presents at LREC2026 — one of the world's leading conferences on language resources and technology — the paper "Capturing Methodology for Generating Synthetic and 3D Training Data in Catalan Sign Language (LSC): The Case of Verbal Agreement" (Gemma Barberà, Inés Broto Clemente, Xavier Vinaixa Roselló, Roger Cassany Viladomat), published in the 12th Workshop on the Representation and Processing of Sign Languages: Language in Motion (Palma de Mallorca, 16 May 2026 — ISBN 978-2-493814-82-1, CC BY-NC 4.0). The work proposes a hybrid methodology based on 3D biomechanics and anatomical constraints to generate high-quality synthetic data for LSC, focused on the case of verbal agreement, as a rigorous alternative to purely generative AI approaches.

In June 2026, Xavier Vinaixa presents Signària at the +RAIN Film Festival — Europe's first festival of AI-made cinema, driven by UPF and Sónar — in the MEDIA+RAIN section on the role of AI in journalism and communication. He gives the talk "Signària: sonic identity for deaf signers in a radio context" and joins a roundtable moderated by Roger Cassany with Veronika Solopova (DFKI · TU Berlin), Rob Lang (Reuters) and David Quiñonero (ARA), at the Poblenou Campus on 16 June 2026.

Signària (UPF)

Official project page

Project

Project description and ethical approach

Team

Full consortium team

Next Generation EU

EU press release

European Commission — Radio for deaf signers

European Commission Representation in Barcelona on Signària (April 2026)

Generalitat — Foreign Action & EU Funds

Department of EU and Foreign Action highlights Signària as a tool funded with European funds

LREC2026 Paper (LSC · Verbal Agreement)

"Capturing Methodology for Generating Synthetic and 3D Training Data in Catalan Sign Language (LSC): The Case of Verbal Agreement" — 12th Workshop on the Representation and Processing of Sign Languages, Palma de Mallorca, 16 May 2026 (ISBN 978-2-493814-82-1)

LREC2026 PDF

Direct download of the paper (CC BY-NC 4.0)

CILCS 2025 Paper

"Identidad sonora personalizada mediante IA para personas sordas signantes" — XVII CILCS Proceedings (ISBN 979-13-87819-03-3)

Academic conferences

International academic presence

+RAIN Film Fest 2026 (UPF + Sónar)

Signària presentation and roundtable in the MEDIA+RAIN section of Europe's first festival of AI-made cinema — Poblenou Campus (UPF), 16 June 2026

Sorensen.ai

Company responsible for technological development

Eventum UPF

Project presentation event

Ràdio Associació — Honor Mention, 26th Awards (2026)

The jury distinguishes Signària for enabling deaf signers to participate in radio

La Local — RAC 2026 Honor Mention

«Tiana presence among the winners of the 26th Ràdio Associació de Catalunya Awards»: Signària receives the honor mention (La Local, Tiana)

Substack — Another AI is possible (RAC 2026)

Reflection on the Ràdio Associació Honor Mention and Signària's humanist vision

Substack — Signària: digital sovereignty & inclusive radio

Article on digital sovereignty, personalized synthetic voice and inclusive radio

betevé — Inclusive radio for deaf people

A tool converts sign language into voice for radio broadcasting

ARA — Pioneering UPF tool

A pioneering UPF tool enables deaf people to do radio

ACN — Personalized synthetic voice

A tool enabling deaf people to do radio with personalized synthetic voice

UPF Focus — Signària-UPF-Mortensen

UPF article about the Signària project

UPF Comunicació — RAC 2026 Award

UPF's institutional press office announces the Honorable Mention at the 26th Premis Ràdio Associació de Catalunya

El País — Deaf people enabled to do radio

UPF develops a tool enabling deaf people to do radio

Paréntesis — Signària and synthetic voice

Signària: technology, radio and deaf signers with synthetic voice

BDN Com — AI and radio for deaf people

AI enables a deaf signing student to do radio with synthetic voice

Badalona Digital — Deaf people and radio

A tool enabling deaf people to do radio with synthetic voice

Comunicació21 — Pioneering tool for radio

A pioneering tool enables deaf people to do radio with synthetic voice

Diari de la Discapacitat — Own voice on radio

Signària is born: technology enabling deaf signers to have their own voice on radio

Al Dia — UPF and synthetic voice

UPF creates a tool for deaf people to participate in radio

Minoria Absoluta — Sorensen presents Signària

Sorensen presents Signària, voice technology for deaf people

Xarxanet — Synthetic voice & inclusive radio

Personalized synthetic voice opens the door to inclusive radio for deaf signers (Colectic)

VIA Empresa — The silent revolution

Signària, the silent revolution

YouTube — Signària presentation

Signària project presentation video

RTVE / L'Entrellat — AI and deaf people on radio

Interview on L'Entrellat (RTVE / Ràdio 4) about Signària and AI enabling deaf people to participate in radio

La Local — Xavi Vinaixa and Signària

La Local article on Xavi Vinaixa's role in the Signària project

Guía de la Radio — Pioneering UPF tool

A pioneering UPF tool enables deaf people to do radio

Jornal.cat — Signària is born

Signària is born: technology for deaf signers

Sant Feliu al Dia — Deaf people and radio

A tool enabling deaf people to do radio with synthetic voice

Línia Xarxa — Deaf girl does live radio

The fascinating tool that enables a deaf girl to do live radio

VilaWeb — Signària, sign language and radio

VilaWeb article on Signària, sign language and radio

3Cat/TV3 — Pioneering tool for deaf signers

TV3 report on Signària and personalized synthetic voice

RAC1 — A deaf student does radio for the first time

A deaf UPF student does radio for the first time thanks to AI, on RAC1's Via Lliure show

La Vanguardia — "Until now it was a privilege of those who can hear"

Feature in La Vanguardia's Neo / AI section (27 May 2026): Signària turns Catalan Sign Language into a personalized synthetic voice. Includes a new quote from Xavi Vinaixa: "It wasn't really about finding a voice donor, but a prosody donor"

El Periódico — Berta Viñas and Signària

Feature by Oriol García Dot in El Periódico (27 May 2026): the first deaf signing student of Journalism at UPF and Signària's technical challenge with a doubly minoritized language

Guía de la Radio — Berta Viñas and Signària

Republication on Guía de la Radio (27 May 2026) of Oriol García Dot's feature, in its Equipos y Tecnología section dedicated to the radio industry