Augmented Vocality: Recomposing the Sounds of Early Irish and Old Norse

Lead Research Organisation: Birmingham City University
Department Name: ADM Birmingham Conservatoire


Phrases such as 'the early medieval period' and 'the Viking age' are rich in imagery. For many people, these words conjure a mental landscape of swords, helmets, longboats and thatched huts. They form part of a shared cultural imagination that encompasses primary school projects, the historically-inspired fantasy world of Tolkien and popular series such as Vikings (2013-19) and The Last Kingdom (2015-). The legacy of such imagery in European fine arts is both deep-rooted and perennial, a fact that was highlighted by exhibitions such as the British Museum's Vikings (2014) and Celts: art and identity (2015-16). In contrast, the ephemeral nature of sound means that there is no clear legacy of the sonic past. When a language ceases to be spoken its literature becomes increasingly restricted to the scholars and specialists who have learned to read it. The focus shifts from performance and dissemination to translation and discussion. Texts that once existed as sonic artefacts of a vibrant oral tradition become fossilised in the silence of the printed page. The language loses its voice.

'Augmented Vocality: Recomposing the Sounds of Early Irish and Old Norse' proposes a novel programme of practice-based research and a methodology to analyse and explore the sounds of early medieval languages. Combining linguistic expertise with sophisticated voice processing technologies the project aims to give new life to early languages and help reclaim the oral quality at the heart of medieval literature. In particular, vocal music composition with live electronics is a powerful tool to develop new insights and reanimate texts from early languages for audiences well beyond the field of literary studies. The project focuses on two languages, Early Irish and Old Norse - both chosen for their particular sonic qualities and the richness of the surviving texts - and comprises three integrated strands:
Cataloguing and sampling of words and phonemes from selected texts;
Analysing the vocal samples to inform the development of voice processing and live electronics software;
Creation of musical compositions for voices, ensemble and live electronics in response to the original sounds and texts.

Each strand of the project will result in specific outputs, including:
Digital audio databases and sample libraries of Early Irish and Old Norse words and phonemes;
Vocal processing software tools adapted to the specific sounds of the chosen languages;
Live electronics software to support composition and performance with the vocal source material;
At least two musical compositions for one or more voices, ensemble and live electronics;
A project website with musical scores, recordings of performances, downloadable software and access to the digital audio databases;
Public concerts in partnership with three renowned music ensembles in the UK, Ireland and Norway;
Conference presentations and workshops in the UK, Ireland and Norway;
At least two peer reviewed journal articles.

As an interdisciplinary project, Augmented Vocality addresses the following research questions in the fields of linguistics and music technology, composition and performance:
How to design a digital audio database and a sample library of words and sounds from medieval languages?
How to develop voice processing techniques to respond to the specific sonic features of the source material?
How could words and sounds from a medieval language be used as a resource for musical composition?
How could live electronics help illuminate the meaning and the sonic qualities of medieval texts, words and sounds?
Is it possible to reclaim the oral nature of a medieval text in a contemporary music performance context?

Planned Impact

Augmented Vocality offers unparalleled opportunities for engagement with early medieval languages, music technology and music performance. The project's varied, accessible outcomes (as described in the Pathways to Impact) facilitate engagement with areas of knowledge that have often been difficult to access from outside academia. Key beneficiaries include those with specific interests in the project's three main strands of languages and literature, technology and music. However, engagement with any strand of Augmented Vocality will invariably result in contact with one or more of the project's constituent fields, further broadening the project's interdisciplinary outreach.

Augmented Vocality will provide intelligent, academically-rigorous audio databases for people who want to hear the sounds of Early Irish and Old Norse. The freely available, online databases will allow these language to be heard and studied in far greater detail than is currently possible with the often inaccurate, variable-quality amateur recordings that exist on platforms such as YouTube. The existence of YouTube recordings, and the popularity of television series including Vikings (2013-19) and the Netflix series The Last Kingdom (2015-), point to widespread interest in the cultures and languages that predate modern European states. Exhibitions such as the British Museum's Vikings (2014) and Celts: art and identity (2015-16) further highlight a popular fascination with early cultures, while also drawing attention to the ways in which institutional engagement with these cultures has tended to focus on visual art and literature rather than on recreation of the spoken word. The online audio databases of Augmented Vocality will make the sounds of Early Irish and Old Norse accessible to audiences beyond the relatively small number of university specialists. In addition to amateur linguists, historians and students, further beneficiaries of the audio databases (described in detail in the Pathways to Impact) include actors, and those involved in the heritage and tourism sectors.

Following the creation of audio databases of linguistic corpora, the development of new vocal processing software tools will be of particular interest to both amateurs and professionals engaged in software development and experimentation. New software developments will be presented at ICMC and NIME. The project will also serve as a template for developers who want to work on new software in response to specific collections of sounds. As Augmented Vocality's software tools are incorporated into the freely available Integra Live software, they will become available to the growing community of Integra Live users spread around the world (as evidenced by more than 30,000 downloads of Integra Live since 2010). Current users of Integra Live, who will benefit from software developments, include music and performance students, sound engineers, and amateur and professional performers.

Performers will also benefit from Augmented Vocality's musical compositions, which will be freely accessible to all via the project's website. For students of singing, the compositions' use of languages other than English will develop familiarity with the International Phonetic Alphabet. The compositions will also help performers to become accustomed to live electronics, empowering them to work with new technology. Live performances of the compositions, initially by project partners Birmingham Contemporary Music Group (Birmingham), BIT20 Ensemble (Bergen) and Hard Rain SoloistEnsemble (Belfast), will make musical outcomes accessible to regular concert-goers and new audiences in the UK and abroad.


10 25 50
Description Creation of a searchable audio database of Early Irish and Old Norse words. The dataset includes more than 1,400 words and 9 poetic texts in Early Irish, and more than 3,000 words and phonemes and 9 poetic texts in Old Norse. All word samples are recorded twice by both female and male actors.
Development of several vocal analysis tools and live electronics processing tools to explore the characteristics and the potential for musical use of the speech samples in the database.
Exploitation Route The methodology that we developed to record, analyse and repurpose the speech samples for artistic creation can easily be adopted to explore other linguistic corpora, both in the context of ancient languages and contemporary endangered languages.
Sectors Education,Culture, Heritage, Museums and Collections

Title Augmented Vocality Archive 
Description Open access online dataset of audio recordings of Old Norse and Early Irish phonemes, words and texts. The dataset includes in excess of 4,000 phonemes, words and texts in Old Norse, and 3,000 in Early Irish. All the words in the database have been recorded by both a female and male voice for both languages. A web interface developed by RA Joe Wright allows for easy access and search of all the audio recordings in the dataset. 
Type Of Material Database/Collection of data 
Year Produced 2022 
Provided To Others? Yes  
Impact The archive is a new resource for students and scholars of medieval languages, and as such is being promoted by the project partners. The online archive will be integrated with other resources (online dictionaries) by staff of the Department of Anglo-Saxon, Norse & Celtic at the University of Cambridge. 
Title ?pnF?nR - Dataset on Alveolar Rhotics from Sud-Oranais (Berber language) 
Description Between 2015 and 2017, I carried out several fieldworks in Sud-Oranais (Algeria). Hence, I have created this web-app, made with Shiny, in order to share a sample of my data and make my work available to the scientific community. The development of this interface follows on from my thesis that I achieved at INALCO: Linguistic description of endangered Berber varieties of Sud-Oranais (Algeria) - A dialectological, phonetic and phonological study of the consonantic system. The purpose of this web application, called ?pnF?nR, is to offer an interactive signal processing of a phonetic corpus on the alveolar rhotics. 
Type Of Technology Software 
Year Produced 2018 
Open Source License? Yes  
Impact Adaptation of the original ?pnF?nR tool to the corpora of Old Norse and Early Irish speech samples of the Augmented Vocality project 
Title Songeriser 
Description "Songeriser" is a computer program developed by Research Assistant Mohamed El-Idrissi implemented to generate a new sonic output by fusing a music sample and a spoken word/text sample. This new approach is based on AI, and specifically using Generative Adversarial Network and Transfer Learning to combine the audio features of both input sources and create hybrid sonic results. "Songeriser" has been used in Augmented Vocality by matching the Old Norse and Early Irish speech samples with relevant music samples chosen by the composer, and then comparing the results to select those that presented the most interesting musical characteristics. 
Type Of Technology Software 
Year Produced 2022 
Impact The software tool has been used so far only by the project team, and in particular to enrich the music composition workflow. 
Title TextGridShiny - Create interactive annotation interfaces for speech with R and Shiny 
Description TextGridShiny allows to read TextGrid files, from the software PRAAT, and to display annotations describing representations of sound recordings. In addition, it is possible to interact with the created visual annotation by adding tiers and labels. 
Type Of Technology Software 
Year Produced 2018 
Open Source License? Yes  
Impact Web application interface updated in 2022 to work with the speech samples in Early Irish and Old Norse created as part of the Augmented Vocality project 
Description RBC Public Research Seminar by Dr Edmund Hunt 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Postgraduate students
Results and Impact As part of Royal Birmingham Conservatoire's Public Research Seminar Series, Edmund Hunt, CO-I on the Augmented Vocality project, delivered a seminar on Tuesday 30th November 2021, 3.30pm-5.00pm, online. The title of the seminar was "Twenty-first century music for a twelve-hundred-year-old voice: text, technology and (un)translatability", presenting his research on how to engage with Old Norse and Early Irish texts and vocal sounds in the context of contemporary music composition with electronics. The seminar was recorded.
Year(s) Of Engagement Activity 2021
Description UCL Workshop 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Postgraduate students
Results and Impact Hybrid workshop delivered at University College London's Department of Scandinavian Studies, as part of their Medieval Scandinavia Seminar Series, on 10 February 2022. The PI Lamberto Coccioli, Co-I Edmund Hunt, RA1 Mohamed El-Idrissi and RA2 Joe Wright all took part in the workshop presenting the research outcomes of the Augmented Vocality project so far. Coccioli and El-Idrissi took part remotely, while Hunt and Wright attended in person in London.
Year(s) Of Engagement Activity 2022