A WORD IN YOUR EAR: LIBRARY SERVICES FOR PRINT DISABLED PEOPLE IN THE DIGITAL AGE. Greg Morgan* Mary A. Schnackenberg* Royal New Zealand Foundation for the Blind Private Bag 99941 Newmarket, Auckland 1031 Ph: +64 9 355 6961, Fax +64 9 355 6936 Email: gmorgan@rnzfb.org.nz Introductory questions What s a DAISY when it isn t a flower? What changes are taking place in talking book technology? How are formats set to converge? What opportunities will there be for cooperation between public or other libraries and libraries for the blind? What management challenges are presented by emerging digital technologies? Note: this session will include a digital talking book demonstration. It is not intended that the session will dwell on technical details, but it will give us a chance to share thoughts on the opportunities presented by digital talking book technology. Royal New Zealand Foundation for the Blind The Royal New Zealand Foundation for the Blind (RNZFB) is a statutory authority constituted under the Royal New Zealand Foundation for the Blind Act 1963. (The Royal New Zealand Foundation of the Blind Bill 2002 was among the Bills carried over for the new Parliament; the Bill introduces changes to the constitution of the RNZFB.) As the sole provider of habilitation and rehabilitation services to more than 12,500 blind, deafblind and sight-impaired citizens, RNZFB works to remove social and environmental barriers that hinder the participation of blind, sight-impaired, and deafblind people in all aspects of life. One of RNZFB's key activities is the production of accessible formats, such as braille, audio, electronic text, large print, under the Permitted Acts provision outlined in s.69 of the Copyright Act 1994. This provision gives legal weight to the moral right of print disabled people to read society's written information. S.69 assists certain, timely production and controlled distribution without damage to the interests of copyright owners. Because information is not commercially available in formats which give equity for blind and sight-impaired people, RNZFB spends close to $2 million each year to produce material in accessible formats. Distribution costs are on top of that. The extremely high unit cost of the time-intensive processes involved in accessible format production is a severe constraint on the number of works that are converted from standard print. RNZFB operates two national library services used by over 5,500 members: (1) the talking book and braille library based in Parnell, Auckland and (2) the Homai Library in
Manurewa, which meets the reading needs of blind and sight-impaired children and young people in special education or mainstreamed settings. There are approximately 1,300 New Zealanders aged 0-21 who are known to have a sight impairment. RNZFB holds a contract to produce curriculum material on behalf of the Ministry of Education, and is contracted by a variety of agencies to make their public releases and other information into accessible formats. Examples are: Census 2001 materials produced in braille and audio, Ministry of Health brochures read on the RNZFB Telephone Information Service. Talking books In this country talking books - audio books produced for blind people to use - have a proud history extending back to their introduction in 1937, only a couple of years after Britain and the USA started making them. Between that year and 1986 the RNZFB changed its talking book format four times, from long playing records to succeeding generations of cassette. The current talking book is a standard looking cassette which is in the four track format so that each side plays twice. Half the normal playing speed gets almost six hours of recording on to one cassette and delivers good quality sound to readers who can use the controls on their talking book machine to speed or slow the reading. What is the reading experience? A talking book provides a human voice narration of the printed document. Our talking books are accurate renderings of the author's words - although they should be an engaging read, they are neither abridged nor dramatised. Tone indexing can be used so that different sections of the work are indicated, for example the start of a new chapter. But the audio is not structured. Going to the start of chapter six means fast forwarding past tones and intervening words. So the Bible, which is highly structured into chapter and verse, is hard to navigate in audio. Likewise, a contents page is just an inactive reading of what will follow, and there is a lot of fast forwarding and maybe shifting back to locate a particular section within the document. Digital talking books Since October 2000 the RNZFB recording studios have been recording books and magazines digitally. From the uncompressed master file we can make, as at present, an audio cassette (two track or four track), or any brand of compressed audio file. However, our commitment on behalf of the leadership is to using a brand of digital technology known as DAISY, and we are a member of the Consortium managing its introduction. See http://www.daisy.org Drawing on a proprietary standard developed in Sweden, talking book libraries formed the DAISY consortium in 1996 to lead the worldwide transition from analogue to Digital Talking Books. Some of the member countries are already distributing DAISY books to end-users. DAISY stands for "Digital Accessible Information SYstem". The Consortium pursues the vision "that all published information is available to people with print disabilities, at the same time and at no greater cost, in an accessible, feature-rich, navigable format" (http://www.daisy.org/about_us/default.asp). In 1997 the DAISY Consortium adopted open standards based on file formats under development for the internet. DAISY is consistent with standard ANSI/NISO Z39.86 2002, the American National Standard for digital talking books which came into existence in March 2002. It has been developed by the National Library Service for the Blind and
Physically Handicapped (Library of Congress), other North American organisations, and the DAISY Consortium. See www.loc.gov/nls for more information. The composition of a digital talking book Several unified files comprise a digital talking book. They are: Package file - an XML 1.0 file that contains metadata describing the digital talking book and a statement of the default reading order of the elements in the document. Textual content file - an XML 1.0 file marked up to support words spelt on demand, keyword searching, and navigation. Output can be refreshable braille display, synthetic speech, or large print on screen via magnification. Audio files - human or synthetic speech recordings of the document in one or more of the specified formats. Image files - digital talking books can contain images to be displayed on playback equipment which includes a visual display, e.g. a PC. Synchronisation files - Synchronised Multimedia Integration Language (SMIL 2.0) is used to define the sequence of media events so that audio, text, and any visual elements are presented simultaneously. Navigation control file - the file handles movement through XML applications by structure (e.g. chapters, sections, page numbers, footnotes), allowing the reader to navigate by major divisions down to the small divisions determined by the detail marked up in the document. More precise navigation within the parts of the document is handled by the textual content file or the SMIL file(s). Bookmark/Highlight file - an XML 1.0 file supports bookmarking and highlighting. Resource file - this file records navigational information, such as where the reader has got to in the document. This is the equivalent of typographical clues in a print book. Distribution information file - this file maps the SMIL files to specific media units. Presentation styles - optional style sheets can be used to control the presentation of the elements within the digital talking book. Some of these files are optional in particular cases. Producers have considerable flexibility as to the type of digital talking book they will make to meet reading needs in relation to a certain document. Types are: Full audio only - i.e. the digital talking book is not structured but contains a recorded speech rendering of the contents of the document. SMIL files sequence only audio elements. The reader cannot navigate to points within the digital talking book. Generally speaking, this kind of digital talking book will be created when analogue talking books are converted into digital form.
Full structured audio - i.e. the digital talking book contains a recorded speech rendering of the contents of the document. A navigation control file uses SMIL technology to link to the structural elements in the book such as page numbers and footnotes; the reader can navigate to these. Only audio items are sequenced by the SMIL files. Full or partial audio with structure and partial text - i.e. the recorded speech renders the full or partial contents of the document. Links to structural elements such as page numbers and footnotes are supported. Part of the document is included in a textual content file to enable features such as keyword searching and spelt out words and/or synthetic speech rendering of, e.g., lists. Images may be included. The reader is able to navigate directly to items in the navigation control file and to specially tagged items in the textual control file. Where different media are present, they are synchronised by the SMIL files. Full audio with structure and full text - i.e. the whole document is recorded in speech, structured, and synchronised via SMIL files so that the reader is able to navigate to items in the navigation control file and to tagged items in the textual control file. Full text with structure and partial audio - i.e. only parts of the document are recorded in speech, and those parts are synchronised with the accompanying text. Otherwise the SMIL contains text in sequence, or text and any image elements. This pattern might be used for dictionaries and other reference works in which human speech pronunciation of certain words is required. Full text with structure but no audio - i.e. this is a digital talking book without audio files! The full text of the document is in a textual content file, structured and linked so that the reader can navigate directly to items in the navigation control and textual content files. SMIL files synchronise text elements and any images. ANSI/NISO Z39.86 2002 is on the web at the www.niso.org site. Download is free. The structure and degree of mark-up chosen - or afforded - for a title will depend upon the likely use of the work. Mark-up is time, time is money. Easy demarcations between fiction and non-fiction or genres or categories of reader might not mean the best use of the resource. For us, the ideal is that any digital talking book we produce should offer the reader the same reading experience as provided by the printed work. The question is: how will sighted readers read this book? Just from cover to cover? So we might go for minimal mark-up, say page numbers and chapter divisions. By skipping backwards and forwards? So we might mark-up sections to let the audio reader navigate around the text in the same way. By looking carefully at particular names and working out their significance? So we might embed a letter by letter rendering of names, or link occurrences of a name back to the list of characters at the start of the story. Converging formats Agencies producing talking books for blind readers encounter in DAISY and the digital talking book standard a technology which supports compression and handles multimedia content. The latter is a key benefit. Many people who have a sight impairment will read in more than one format to get the same reading experience as a sighted person. For instance, the sighted person's ability to skim read a print page and omit irrelevant information might
find its equivalent in using audio to read quickly through some of a report, and then enlarged print for another section, perhaps some graphs that are not easily verbalised. Or a person might choose to read braille for study-related purposes because it has all the structure and referencing systems in the print works, and then audio for reading a light action novel. As we have seen, the textual content in a DAISY-type digital talking book will be a valid XML 1.0 file. In addition to the text of the document, the element set can include the mark-up for properly formatted braille that can be read via a refreshable braille display, large print, or synthetic speech. A "showin" attribute controls the display of an illustration which varies from one format to another - e.g. a bar graph in large print might physically vary from, but will give the same information as, the graph in braille. Teachers of sighted children can use a range of exciting print and multimedia options to stimulate the interest of a reluctant or underachieving reader. Digital talking books will provide new scope for encouraging the blind child who has not clicked to the delight of reading, or for whom the unwieldiness of a traditional format is off-putting. This new age of integrated formats promises blind and low vision youngsters more navigable and portable texts as well as the experience of enriching multimedia elements. See http://www.rfbd.org/apb.html Delivery How will blind people get DAISY talking books? Currently our four track audio cassettes are played on a special player known as a talking book machine. Digital talking books technology presents us with some exciting options. Several agencies are working on digital talking book players that take a book on CD. It is not that the book is itself a CD or has to be on one. It is delivered by means of CD. Or computer users can load reading software on a PC and read a digital talking book that way, combining audio and text output for example. The book could be sent out on a disk or emailed. Download devices that work from a PC or via a broadband connection have the potential to by-pass the expensive handling of physical talking book and to give users the convenience of a book slipped into a back pocket. Companies are even exploring players that have a built-in CD burner so that field narrations (i.e. outside the studio environment) can create a DAISY book immediately. Producing agencies will of course ensure that no matter the circumstances under which digital talking books are made, specifications and customary production standards are observed. Players that comply with the standard will be able to decode specific MPEG-A AAC, MP3, and Linear PCM - RIFF WAVE audio formats. Compressed audio (so the biggest Harry Potters on one CD, not a dozen or more) is restricted to specified output sampling rates. Overlapping needs A Here is another aspect of the "converging formats theme". Under s.69 of the Copyright Act prescribed bodies such as RNZFB may produce accessible formats for the use of print disabled persons regardless of whether they use our library. Why has the RNZFB not been supplying public libraries or the National Library of New Zealand's Print Disabilities Collection with a steady flow of well narrated, up to the minute titles, including New
Zealand and non-fiction works? Primarily because in the past libraries were not set up to handle large-scale, restricted access services in non-print media and then because the cost to us of making two track versions of our four track books was prohibitive. The environment is changing. First there is the demand - From 2001 New Zealand Disability Survey Snapshot 6, Sensory Disabilities: An estimated 81,500 New Zealand adults were blind or had a sight limitation that could not be corrected by glasses or contact lenses in 2001. Approximately 7,800 of these adults were completely blind, the rest had some level of seeing limitation that made it difficult for them to see ordinary newspaper print, or see the face of someone across the room (with glasses or contact lenses if they usually wear them). http://www.stats.govt.nz Then there has to be the mechanism - Public libraries do not charge sight-impaired people to borrow from their audio collections, so in effect manage a register of print disabled persons. Under our Copyright Act rights and responsibilities we can supply books to libraries that can restrict their use to readers who have a print disability. Could such books be more widely released? Yes, if someone undertook the necessary copyright clearance processes. What are the incentives? Once it has got through the development phase of producing/converting and archiving digital talking books, RNZFB will begin to think about producing non-daisy formats for other uses that comply with the Copyright Act. Public libraries are very keen to acquire local content and non-fiction for their audio collections. What are the responsibilities? Compliance with legislation and intellectual property rights. Blindness agencies have an impeccable record in working within the exception provisions and audit requirements of their national copyright legislation. We adopt a human rights approach in advocating for permitted exceptions because publicly released information disseminated only in standard print discriminates against those who cannot read it. Accessible format producers need to be able to produce works in a timely way, not held up by copyright clearance hurdles. The responsibility we accept is that if a work were to be produced for any use not covered by a permitted exception, that use would be duly cleared with the copyright owner. B Internationally there are endeavours to create large repositories of documents that are the digital equivalent of printed text documents. Examples are the Elsevier Science archive at the National Library of the Netherlands. At the time of writing this paper, the US Congress has before it the Instructional Materials Accessibility Act 2002. This would require any publisher supplying a text for use in schools to deposit an electronic file copy of the work with a National Instructional Materials Access Center. The deposit of documents in a designated national standard file format would support the subsequent production of accessible formats.
Conclusion Published information that exists in a suitable file format alongside the printed document enables agencies to make significant reductions in the time required to produce accessible formats. An existing XML file becomes the basis of the mark-up which makes the experience of reading a digital book like that of reading a print book - there are page numbers, you can flick backwards and forwards, the tables of contents or the index are active to allow precise navigation to the desired page. For print disabled citizens digital technology offers structured, navigable reading - that is, the functionality sighted readers take for granted when perusing a printed or flat-on-a-screen document.