This is our fourth blog on the digitization of braille music. So, what is new with this project?
First, we are boosting our production by outsourcing the proofreading of scores scanned using Optical Braille Recognition (OBR). Our braille music specialist Gilbert Busch continues to review all the scans completed on DotScan so we can provide accurate scores.
As always, our first priority is patron requests. These are usually time sensitive, so soon after we receive the request, the music is scanned and proofread by a librarian then placed on Gil’s desk for quality assurance.
Every day we scan on both the OBR system and DotScan to digitize our entire braille collection. We use DotScan to digitize one-sided braille scores that have never been circulated. These are our masters; their braille dots are not worn down with use, so the images we scan are clear and well defined.
We are using OBR for mass digitization of the general collection, most of which are two-sided (inter-point) braille. On OBR, the flatbed scanner takes a picture of the page which shows the protruding dots of the front page as light circles, while the depressions caused by the dots on the back page appear as dark circles. OBR is equipped with an algorithm to identify and separate the light and dark dots into proper groups as you see in the original score. This makes it possible to scan both sides of the braille score at the same time.
As you may recall from our last blog on digitization, one of our biggest challenges is capturing accurate and complete images of all the dots in the braille score. The NLS Music Section has been exploring ways to tackle this challenge.
We are excited to report that last week, Katie Rodda and I presented a paper entitled “Digitizing Braille Music: A Case Study” at the 2018 Imaging Science and Technology Archiving Conference at the National Archives in Washington DC. This was a major conference, with archiving professionals and librarians from 17 countries and 27 states attending the three-day event. We heard many interesting presentations that definitely broadened our horizons to topics such as new techniques for digitizing historical photographic archives, 3-D scanning of textured objects, and developing guidelines for government records.
Our presentation was a case study outlining our processes for digitizing braille music and the challenges we face in capturing accurate images of the braille scores. After our presentation, we received some very innovative and interesting suggestions from scientists and archivists who are experienced and knowledgeable about digitization. They were definitely thinking outside the box. Some of the suggestions that we received are:
- Updating the software of the single-sided digitization to digitize inter-point, to output both the front and back pages of a braille sheet like OBR.
- Using a device with sensors that will work like fingers on the braille to detect the dots (some variation of this idea was suggested to us by three different scientists).
- Using laser measurement to detect the protruding dots and depressions made by the dots on the back side of the page.
We conducted an intensive literature search to prepare for our presentation but had not come across these novel ideas for scanning braille. We clearly sense that there are many new technologies that can help us digitize with more accuracy and speed. The Music Section will continue exploring new methods of digitizing.
Please read our past blog posts on braille music digitization:
Comments (2)
Having read your latest post on your digitising project, I was wondering if you could give an indication of how much has been scanned, and how much left is there to do?
Regards,
Roger Firman.
This fiscal year (since October, 2017), we’ve scanned over 10,000 pages of braille music but we still have many more pages to scan. Fortunately, most of the music we are acquiring is born digital so the total number of the pages to scan will not increase much.