CrowDescription

flat =Definition= CrowDescription is a neologism coined by Saveria Arma (of [|CulturAbile]) to indicate the crowd-sourcing of audio descriptions. =Basic idea= The idea of a CrowDescription project is based on the following considerations: It would therefore be useful to do a crowdsourced / collaborative audio description of a video and to discuss the concrete issues arising, and its first part - writing and timing - is feasible with the mentioned collaborative captioning tools. The written audio descriptions could be rendered audible by recording either a human reading them, or a text-to-speech software voicing them. =What kind of video?= The video for such a project should Possible videos:
 * Audio descriptions raise interesting issues as to what should be audio-described and how: i.e. what is relevant to the understanding of a video without seeing the pictures, and how can the description of these relevant elements be fitted in moments without speech or other relevant audio info.
 * Audio descriptions must first be written and timed, and this can be done with online collaborative captioning tools.
 * be relatively short (otherwise people might get discouraged);
 * not have too much speech, to facilitate the insertion of descriptions:
 * under a license authorizing derivative works (because adding audio descriptions to a video means creating a derivative work) and possibly commercial use (because otherwise, several hosting platforms would be barred because they have ads, and hence are commercial). E.g. a Creative Commons Attribution - Share Alike (BY-SA) license http://www.creativecommons.org/licenses/by-sa/3.0.
 * Sita Sings the Blues (Full Movie 720p) Part 1 / 10 http://www.youtube.com/watch?v=PfS2p1vFics.
 * 6:57 long, with the possibility to start audio descriptions after the initial credits (from 5:15 - "San Francisco")
 * almost no speech
 * under a BY-SA license
 * ... (add more)

=Writing and timing= Both [|Universal Subtitles] (univsubs) and [|DotSUB] are collaborative online captioning/subtitling tools. For this project, univsubs is probably more suitable because: Other tools ....
 * all language tracks can be time-coded independently from the original language
 * there is a wiki-like history for all tracks, which means the same track could be used by different people, and all versions would be kept, with the names of their authors, even if they forget to download their work.

=Text-to-speech tools= =Audio editing tools= =Video editing tools= (to be written) =Repository and discussions= A repository where all participants can publish their work and discuss it is needed. A free wiki like this one could do: Other ideas .... =Possible work flow 1=
 * http://www.ispeech.org/ (online application which does the TTS and allows its recording)
 * OS voices (computer OS have synthesis voices in their accessibility features)
 * ... (add more)
 * Audacity http://audacity.sourceforge.net is free software and multiplatform. Moreover it has a timed label track in which the timed written descriptions can be inserted to check the length of each audio file
 * .... (add more)
 * the written timed descriptions could be uploaded in the wiki's files,
 * audio files comprising the audio descriptions could also be uploaded in the wiki's files, though it would be better to host audiodescribed videos elsewhere and embed them in the wiki, because of storage limits and for further visibility,
 * pages could be made for each participant and for general discussion (each page has a discussion feature)
 * 1) Publish a complete description of the project and work flow somewhere online
 * 2) Upload the video on an online video hosting platform, specifying in the title that it is meant for this CrowDescription project in the title, and shortly explaining the project in the description of the video, with a link to the complete description
 * 3) Create a page for it by streaming at an online collaborative captioning tool (e.g. univsubs: see above)
 * 4) Start a few time-codable "captioning tracks" for different languages in that page, specifying in the 1st "caption" that the track is meant for writing audio descriptions, not captions, and with the URL of the complete description
 * 5) Invite volunteers to start audio describing, asking them to download the resulting file when they have finished, and to upload it - with a name identifying the authors and the language used - in the chosen repository, and to discuss them there.
 * 6) Invite volunteers to make audio files of the timed written descriptions and to upload them - with a name identifying the authors and the language used - in the chosen repository, and to discuss them there.
 * 7) Invite volunteers to integrate the audio descriptions in the video, to upload the described video in an online platform, and to embed them in a dedicated page of the repository, and to discuss them there.

Possible work flow 2
and further ones: add yours.