accessible av in the duke digital repository · shortcode support for embeds in any sites@duke...
TRANSCRIPT
Accessible AV in the Duke Digital RepositorySean Aery & Jim Coble Duke University Libraries Samvera Connect, Nov 2017½ ½
ı Playback
JWPlayer so�ware (free)
HTML5 video w/Flash fallback, HTTP pseudo‐streaming
Mul�‐file AV renders as playlist, with external controls
Uses Apache mod_xsendfile module, supports byte range requests
Rails controller uses send_file with path to file on mounted storage
Interac�ons captured (via JS API) as Google Analy�cs events
ɲ Cap�ons & Transcripts
WebVTT cap�on files uploaded as file datastream
ǫ bu�on in video; cap�ons render in player UI
Interac�ve transcript: cap�on parsed to HTML via webv�‐ruby gem
Đ Embedding
Embed code (iframe) copyable from Share menu
Shortcode support for embeds in any Sites@Duke Wordpress site
Embedded view is responsive; includes some metadata, permalink
Ⱥ Rights
Store URI for rights & text note
Crea�ve Commons &RightsStatements.org
Displays icons, short text, re‐use label
- Access Control
Masters & deriva�ves can be protected
Role‐based access control
Tie‐in with Shibboleth, campus groups
ĝ Integra�on
Export digital object info to ArchivesSpace
Finding aids UI displays DDR AV inline
DDR item page links to spot in finding aid
Over the course of 2017, Duke University Libraries developedsupport for audio and video objects in the Duke Digital Repository(DDR). The DDR is presently a suite of applica�ons built on a(Fedora 3 based) Samvera Hydra head Rails engine.
We needed to implement AV solu�ons in 2017 to accommodatetwo grant‐funded digi�za�on projects and to enable migra�ngexis�ng AV collec�ons from our legacy digital collec�ons pla�orm.Projects include:
• Duke Chapel Recordings (Lilly Endowment Inc.)
• Voices of Change: Preserving & Presen�ng Radio Hai� (NEH)
• Adver�sements & oral histories from legacy pla�orm
OVERVIEW Ë Goals
• Support hybrid collec�ons of AV, images, documents
• Model, ingest, preserve, & play mul�‐file AV objects
• Accommodate intermediate / mezzanine files
• Meet WCAG accessibility guidelines for �me‐based media
• Enable manual provision of low‐bitrate deriva�ves
Next Steps
• Export WebVTT cap�ons as .pdf or .txt
• Advance player via linked �mecodes in descrip�on field
• Upload & display .pdf or .doc transcripts
GOALS / NEXT STEPS Ë While we have succeeded in mee�ng our goals on a rela�vely
short �meline, there are some notable limita�ons to what wedeveloped:
• Not a Hyrax solu�on
• One deriva�ve per file, no auto‐genera�on
• No adap�ve bitrate streaming (e.g., HLS)
• No mul�lingual cap�ons (yet)
• Cap�ons not indexed in Solr (yet)
LIM
ITATIONS Ë Contact
[email protected] [email protected]
Visit repository.duke.edu
Code
github.com/duke‐libraries
Project Blog blogs.library.duke.edu/bitstreams/
LEARN M
ORE Ë