Project structure

One platform, several language corpora

Pronunciation Matters is designed as a multilingual platform. The web app brings together the project languages Spanish, French, English, and German in one shared structure. Each language has its own corpus area with its own responsibilities, task materials, and disciplinary priorities. At the same time, central structural principles remain stable across the project.

This combination of a shared platform and language-specific design is important. Pronunciation research cannot be organised identically for all languages. Spanish, French, English, and German differ in their sound systems, patterns of variation, didactic traditions, and typical challenges for learners. A general project design must not cover up these differences. For this reason, the concrete research designs of the language corpora are presented within the respective corpus areas.

The general project pages, by contrast, describe the overarching logic: why the project exists, how the platform is structured, what kinds of data are produced, how research and teaching are connected, and who is involved in the project.

Areas of the web app

The public web app is divided into three main content areas: Project, Research, and Teaching.

The Project area explains the basic idea, the development of the project, the platform structure, the methodology, and the collaboration behind it. It is intended for all users who want to understand what Pronunciation Matters is and how the project works.

The Research area leads to the individual language corpora. The respective corpus designs are explained there. The protected areas contain the actual speech data and research tools: the Speakers page provides access to the participating informants and their session-scoped player entries, comparison views support contrastive analyses, and phenomenon-based item sets allow users to work with targeted excerpts of the material. These areas are not openly accessible because they work with pseudonymised speech data and recording-related metadata.

The Teaching area presents selected content for pronunciation teaching. It is deliberately leaner than the Research area because the focus here is not the full research environment, but thematically focused access to especially relevant aspects of pronunciation. Such topic pages can use contrastive comparisons, make typical difficulties visible, and provide material that teachers can integrate pragmatically into existing lessons. The project does not assume that pronunciation will regularly be given entire teaching units in school-based foreign language instruction. Rather, the aim is to provide clear and academically sound extensions for moments where curricula, teaching situations, or concrete learning problems allow room for them.

Public orientation and protected research access

Pronunciation Matters distinguishes between public information and material areas on the one hand, and protected research data on the other. This separation is not merely a technical decision; it is necessary for disciplinary and data-protection reasons.

Speech recordings are personal or pseudonymised research data. Even when clear names do not appear in the web app, voice, metadata, and recording contexts can contain sensitive information. For this reason, detailed research access is not made freely available to the public. Access to the player, comparison views, speaker profiles, and phenomenon-based work surfaces remains controlled and is intended for authorised users from research and educational institutions.

Publicly accessible content includes project information, general methodology pages, language-specific design descriptions, and released teaching materials. In this way, the project remains transparent without publishing research data in an uncontrolled manner.

Research and teaching as separate but connected spaces

The platform deliberately separates research and teaching. This prevents teaching materials from appearing as a reduced research database and prevents research surfaces from having to take on didactic functions for which they were not designed.

In the Research area, users can work with detailed data. They can open session-scoped player views, select excerpts of material, build contrastive comparisons, and create or modify phenomenon-based sets. The Research area is therefore a workspace for analysis and university teaching.

In the Teaching area, selected content is presented didactically. The focus is on clear explanations, tasks, listening examples, and materials suitable for school or university teaching contexts. These materials can emerge from work with the research data, but they are not simply delivered directly from the protected data space.

Technical infrastructure as part of the project idea

Pronunciation Matters is also a digital humanities project. The technical infrastructure is not a layer added after the fact; it is part of the scholarly and didactic method. Data modelling, structured metadata, annotations, audio processing, web app design, and protected access spaces are what make the collected speech data practically usable in the first place.

The platform is intended to support research and university teaching comfortably. Users working with the data should not first have to understand folder structures, filenames, or raw formats. Instead, the web app offers structured access through languages, persons, session-scoped player views, comparison views, and pronunciation phenomena. Technical complexity is not made invisible, but organised in a way that facilitates scholarly work.

A supplementary technical project page on Pronunciation Matters is maintained in the context of Hispanistica @ Marburg: https://hispanistica.com/projects/promat/. The public web app itself, however, foregrounds the project idea, the research and teaching logic, and the usable content.