The State Library of NSW had been investigating opportunities to engage online volunteers to help with the transcription of over 11,000 hours of historical sound archive collections. An ongoing initiative of the Library had enabled them to create digital versions of their sound collection, but finding a way to deliver time-coded transcripts for each recording was important in ensuring the collection was truly accessible. As a non profit institution, human transcription services are unaffordable for the Library, so crowdsourcing offers a sustainable way to achieve the task; while also giving people an opportunity to engage with their collections in a way not previously possible.
The State Library of NSW conducted extensive research to find out what sound collection platforms were already available for transcribing and delivering audio transcripts, and although they found quite a few, the most suited was the New York Public Library’s Transcript Editor. Most of the platform is written in Ruby on Rails which is open-source, with 70-80% of their codebase matching the State Library’s requirements. It made sense to modify and customise New York Public Library’s existing codebase rather than try to reinvent the wheel, but due to the limited capacity of internal resources and the fact that the Library was venturing into a new technology space, a Ruby on Rails agency was required for this project.
After researching agencies, the Library decided to engage reinteractive due to our enthusiasm about the project and our quick response to the Library’s brief. There was a strict timeline for the project as funding had been granted but needed to be spent before the end of the financial year, giving us a limited time frame to make magic happen!
”The expertise, the direct communication with the developer and the seamless project management has been fantastic. Although I wasn’t familiar with Ruby on Rails, reinteractive really put me at ease.”Jenna Bain - Digital Projects Leader, State Library of NSW
reinteractive kicked off the project with the project manager and one of our lead developers, mapping out the build requirements. The scope included the “must have”, “would be great” and the “if we can get to it” priorities. One of the desired customisations was a ‘play all’ function so that people would have the ability to listen to a full audio recording before committing to any transcription. That feature plus others are unique to the State Library’s installation, built as added functionality on top of the already robust and feature-rich platform developed by the New York Public Library.
Another key feature of the project was to link up existing computer-generated transcripts with the new platform. VoiceBase was used to produce machine-read transcripts that are 70-80% accurate, which are then uploaded to the platform to be reviewed and corrected by the public as required. Once enough people have corrected each line of the transcript and made the same changes, a consensus is calculated and that line is marked as finalised.
Every single feature outlined in the project scope was achieved with ease, which was a surprise and a major achievement for the Library. They were able to break patterns and work with an unfamiliar agency and have a delightfully positive experience.
- Transcription editing platform completed in just 4 weeks ready for launch, enabling the State Library of NSW to realise a project they had been working on for over a year
- Achieving 100% of the initial scope within budget
- Ability to create a new dataset of accurate transcripts for a historical sound collection
- Creating an opportunity to generate more interest in and engagement with the Library’s rich sound archive.
”We were really surprised that everything in the scope, even the things we thought may not be possible, were completed - it just wasn’t a problem for the developer. It means we are able to promote our collections and the new transcripts like we have never been able to do before. That is a game-changer for us and will ultimately enrich the Library’s collections by making them more accessible, searchable and discoverable.”Jenna Bain - Digital Projects Leader, State Library of NSW