textav-event-2018
  • textAV - 2018
  • Introductions
  • Projects
    • Hyperaud.io & Studs Terkel Archive - Mark Boas
    • Frame Trail and German Parliament - Joscha
    • Opened Captions & Annotated Articles at the FT - Joanna
    • Fast Forward Audio prototype, BBC R&D New News Team - Tristan
    • Full Fact - Automated Fact checking - Mevan
    • AV work for IIIF, and the context of that work for the British Library - Tom Crane
    • Building a a STT services at BBC - Eyal
    • Subtitles and accessibility at the BBC - Eyal
    • BBC SUMMA Jetski - Eimi
    • BBC Radio Dicer/ Hot Fuzz - James
    • autoEdit Panel for Adobe CEP - Pietro
  • Unconference Projects
    • Un-conference Pitches
    • STT Benchmarking
    • Full Fact - tweet that clip
    • Farfetchd
    • "Selective Hearing" - Concept Clustering in Podcasts
    • IIIF video segmentation
    • IIIF Interactive Transcript - Parliamentary Debates
    • IFF Collaborative podcast annotation workflow
Powered by GitBook
On this page
  1. Unconference Projects

IIIF Interactive Transcript - Parliamentary Debates

PreviousIIIF video segmentationNextIFF Collaborative podcast annotation workflow

Last updated 6 years ago

One of the demos on Day 1 of BBCTextAV was the BBC's FastForward, which synchronises scrubbing of video with the text transcript (navigating through either video or text changes your position in the other).

This would be useful functionality to have available for AV content, where the text transcript comprises textual annotations on the time dimension of the canvas. So we decided to build a proof of concept viewer.

We had access to metadata for ~3500 German parliamentary debates. We transformed this metadata to IIIF Manifests, ending up with one per debate, each with one representing the duration of the debate, annotated with a video file to provide the AV content. Other sources of textual data were available, but for this MVP we used the WEBVTT files as source for the transcripts.

Sample Manifest:

This links to an annotation list for the text content:

The manifests are then shown using this viewer:

Team

  • Joscha Jaeger

  • Tom Crane

The parliamentary source data is hacked into IIIF here:

We produce a IIIF Collection of the manifests, so the viewer software knows what's available:

https://tomcrane.github.io/bbctextav/iiif/ID191002001.json
https://tomcrane.github.io/bbctextav/iiif/ID191002001-transcript.json
https://github.com/tomcrane/bbctextav/blob/master/converter/make_iiif.py
https://tomcrane.github.io/bbctextav/iiif/collection.json
https://openhypervideo.github.io/iiif-interactive-transcript/
IIIF
Manifest
Canvas
https://openhypervideo.github.io/iiif-interactive-transcript/