Verksamhetsberättelse 2020/Story: Raise your voice

Access
2020

In this Story we are focusing on the continuous work done by Wikimedia Sverige to improve the structures for audio files on the Wikimedia projects, which includes support to share newly recorded music, batch uploads of out-of-copyright music, inclusion of sheet music, information about musicians and tools and structures for speech recordings.

One of the pieces of sheet music shared by the Swedish Performing Arts Agency.

The Wikimedia projects are currently not well developed to handle audio files and supporting material; as a result, there is not much audio material available and potential contributors do not have access to sufficient resources and guidance. We have tried to address this gap for a couple of years now through a number of initiatives which we continued to scale up in 2020 and 2021. The issues include how to compile the information about audio files in a structured way to make it useful and searchable, how to ensure full coverage of a thematic area, how to source the information and files properly and how to be able to both upload existing recordings and to encourage the production of new recordings in a fast and efficient manner.

We believe this to be of crucial importance as music has been part of civilization for a long time, and this is true in most cultures. As a phenomenon in society, it has an increasing value and is generally perceived as positive. Music is played by many people and it can be said to belong to the cultural expressions that both distinguish and unite groups of people all over the world.

Playing music is a way of preserving local and national culture and it is an area that is closely associated with other cultural expressions such as film and dance. As digital production of media increases, so does the need for more free material that can be used.

The Swedish cultural heritage institutions have been good at documenting music and especially Swedish folk music. There is no clear definition of folk music as it is alive and developing while it is being played. But one way to preserve knowledge about it is to document live performances. Together with the Swedish Performing Arts Agency, we have been able to share some of the recordings made by the Swedish Radio in the 40’s and 50’s under a free license.

There is another possibility here as well: creating new recordings of old music. In this way we get a modern interpretation of sheet music and works that were created hundreds of years ago. It is very fun that both old music and new interpretations can be found on the Wikimedia platforms, allowing users to hear music while reading an article about it. Since music is a common and popular way of expression, many can participate and contribute. We would like to see that music played locally across the world has an obvious place in the free knowledge sphere.

By describing the audio files in a structured format and connecting them to other resources we believe that the Wikimedia platforms will be able to provide a unique learning environment while also improving searchability overall around some specific areas of music and speech recordings.

By describing the audio files in a structured format and connecting them to other resources, the Wikimedia platforms become a unique learning environment that at the same time improves the searchability overall in certain specific areas for music and speech recordings. We have worked with institutions and created processes and structures for files and metadata. It is in this initial work where we tried and tested different solutions of, for example, modeling on Wikidata and structured data on Commons. It takes a few tries before you find a "best practice" that you are happy with and can use for the next batch of work. It is an iterative work with constant improvements.

The work we are doing with audio files of music will also help us to prepare for structuring speech recordings, which we hope to collect as part of the project Wikispeech – Speech Data Collector 2019 in 2021. In this project we will collect speech data to build better natural language processing solutions. This will also allow us to improve the different Wikimedia platforms by adding pronunciation recordings to Wiktionary and Wikidata lexemes, as well as to Wikipedia articles (e.g. pronunciation of place names). The structure of the recordings might also be possible to reuse for oral citations later on.