Wikisound

From Meta, a Wikimedia project coordination wiki
Other languages:

Project status[edit]

Work is underway creating spoken versions of articles on various projects. Activity is being coordinated on a per-project basis.

See:

Original proposal[edit]

Wikisound is a proposed project to provide spoken versions of pages from Wikipedia and its sister projects.

As well as appealing to blind or partially sighted users, a spoken wiki project could benefit people with reading difficulties and those who prefer listening to material to reading it.

Wikisound could be distributed via Wikiradio. This would give access to a far greater number of people than the Internet does.

Project possibilities[edit]

This idea does not include only Wikipedia. It could be particularly viable for Wikisource. For example, the poems in Wikisource would be prime candidates for recording as they are short, often meant to be read aloud, and usually in the public domain so you know there are no copyright issues.

Wikibooks would also be very suited to a spoken version, especially its languages bookshelf as this would allow learners to hear native speakers saying the vocabulary, which will be far more useful for learning the language than reading it on screen would be.

For Wikipedia, we may want to focus only on "featured articles" for the initial stages of this project, and produce a very selective version of Wikipedia.

For Wiktionary, having at the minimum a pronunciation of the word being defined could be useful, even if the whole definition was not read aloud.

Self-guided audio tours could be recorded for museums, sightseeing, and tour groups. The files could be downloaded onto devices such as portable audio players, cell phones and PDAs. Short descriptions could be read from Wikipedia articles. Playlists could be used to arrange separate audio files into a tour.

Costs[edit]

A grant may be possible to support such a project from an organisation who said they may be willing to help if we are providing a service to the disabled.

One potential expense involved in this project will be the amount of disk space needed. There is currently a 5MB upload limit.

If a high-quality recording were required, studio time would be another expense, although it may be possible for studio time to be donated. Local radio stations might be willing to donate recording studio time in return for having a copy of the content they could put on air. This could also provide some free radio publicity for Wikipedia and reach new audiences.

Quality and format[edit]

There are varying levels of quality we could aim for; anything from users reading a text into a microphone in their home to trained professionals reading the texts in a recording studio.

Whether to provide the recordings as streaming audio or download only would need to be decided. Wikisound might also prove popular in CD format.

For downloadable versions, it may be worth recording the introduction separately from the entire article. This allows people to decide if they want to download it before waiting for the whole file, and is useful for people who are just browsing and do not need the entire article.

Lower quality versions could be provided for non-broadband users.

Adapting articles[edit]

One major change would be the way in which links are dealt with. It would not be easy to incorporate links into a spoken version, and if only a select few articles were being recorded, many of the links would not be useful anyway. If links were used, they could be marked by using a different voice for them, or by having a sound whenever a link occurred. An alternative to links might be to list related articles at the end rather than throughout the text, with a note about whether it is available as a recording. How to deal with external links also needs to be considered.

Some articles might benefit from being adapted for the spoken version. If there are to be no links included, perhaps parts of other articles should be merged with the main article so the information is in one place.

Reducing the sentence length would make it easier for the person recording the text to read it aloud.

As well as reading the text, the images or diagrams used in page should be explained. This might be a good time to look at improving the quality of alt text used across the projects. "Audio photographs" could also be used instead of ordinary images. Increasing the collection of sounds available, perhaps in conjunction with the Wikimedia Commons, could benefit spoken articles.

Lists, tables and mathematical formulae might need to be presented in an alternative way if they are to be read out.

The list of authors and the text of the GFDL would need to be recorded as well.