Accessible Music

Home  Braille Music  Talking Music  Finale Plug-in  Interfaces  Contact  

Talking Music

Reading Braille Music is a complex task and for those people who become impaired beyond the age of twenty, it is extremely difficult to learn Braille. For those who can learn Braille Music, there are often significant delays in receiving music scores, as the highly specialised production process is both time-consuming and expensive. Indeed, in many European countries the service either does not exist, or has been discontinued for reasons of cost or a drop in demand caused by fewer people learning Braille Music.

Talking Book Record from 1939
A Talking Book on vinyl from 1939.

One of the most popular alternatives to Braille music is Talking Music. Talking Music is presented as a played fragment of music, a few bars long, with a spoken description of the example. The musical examples are now produced automatically by software and as far as possible with a non-interpreted version of the music. All notes are played at the same volume, at the same velocity, and all notes of a certain duration have exactly the same duration. The musical example is there to provide the user with a context for the Talking material, making it easier to understand and play. The musical example is non-interpreted, to afford the Talking Music user the same level of subjectivity to the musical content as the sighted user.

The most important part of the information is the name of the note. All the other information should be centred around this. Some of the additional musical information should be placed before the name of the note: some of it should be given afterwards. The guiding principles are the way in which the musical information will be processed; the duration of the musical information; and the importance of the information. For instance, we must indicate which octave the note belongs to before we say the name of the note so that the listener can prepare some motor movements before playing the note. When, as an illustration of the second principle, a dynamic sign and a bow is mentioned at the same position in the score, the one with the longest duration should be mentioned first. Thirdly, not all the information is equally important for everyone. For example, fingering is additional information, and is mentioned after the note and its duration.

Naturally, it is preferable not to describe every facet of every note, as some data are implicit. Having established a basic set of rules it was then necessary to consider ways of limiting the volume of information without losing any critical data. This led to the use of abbreviations. For example, we do not need to include the octave information for every note. If the interval between two consecutive notes is a third or less, then the octave information can be excluded. In the same way, if the following note is a sixth or more above the current note, we have to include octave information. If the following note is a fourth or fifth above the current note, and the notes belong to different octaves, we do need to include octave information. The abbreviations adopted were the same as those employed in Braille Music.

The DAISY Standard

Talking Music scores make use of the DAISY standard for talking books. The DAISY consortium developed the DAISY format (DAISY 3.0/NISO z39.86), which allows data to have an underlying structure for navigation reasons. The main application for this technology is in Talking Books for print impaired people. There are now hardware players for navigating DAISY structures and these are often supplied free of charge to visually impaired users in many countries.

DAISY is a key international standard to incorporate in accessible formats. Not only is it a structure with which impaired users are familiar, but it is also sufficiently well established to be accepted in recognised hardware. A DAISY book contains many files, which are synchronised to provide the information and a means to navigate the information.

Each fragment contains an .mp3 file and a .smil file. The .mp3 file provides the audio information, while the .smil file provides the structural and metadata information for that particular fragment. This essentially describes the content for each fragment. Each fragment has this information, and the .smil files can communicate with each other to provide different levels of abstraction. The collections of files are controlled and synchronised by a master .smil file and an ncc.html file, which provides the hierarchy for the structure, and is the file which is read by the DAISY player or software.

Talking Music output

One of the guiding principles of universal design and Design for All guidelines is that the print disabled should have access to the same information as the sighted, and that only the format in which the information is presented should change. For Talking Music, this means that everything on the page of a music score must be represented in a Talking format. Furthermore, this Talking format must be applicable to all types of music and instruments. By way of illustration:

Bach Excerpt
Bach Excerpt - Minuet in G - BMV 114.

The resulting Talking Music for measure 1 of the simple example above is shown below:

The title is: Minuet in G. Composer: Bach.
This piece contains 8 bars and consists of 2 sections. The piece is read in eighth notes, except where otherwise indicated.
Key Signature: 1 sharp. Time Signature: 3 fourths time.

Section 1.
Bar 1.
second octave g half.
a quarter.

in agreement with:
fourth octave d quarter.
third octave g.

The description above shows the information within the Talking Music book taken from a DAISY Software player. The description starts with a header which delivers the document data about the score (composer, title, production notes), a description of the score (in terms of length and content), and then the key and time signatures. The description then continues to section 1, where it describes the first measure note by note. The lowest voice is described first.

This information exists as audio files, and the text description is used for production use and explanatory reasons. The audio files are structured in such a way that the user can easily navigate from one section of the score to another, in order to build up an idea of what is in the score. The tree structure of a more complicated score taken from a DAISY Software player is shown below:

Talking Music Score structure
Talking Music Score structure

A Talking Music book exists as a collection of files controlled by an .ncc file but to describe the structure it helps to think of a folder hierarchy. The Talking Music Book contains at the first level folders for the header information and folders for each section. The folders for the header information can be navigated through vertically in order to skip to the desired piece of information. There is then a midi representation of the entire score. This is usually played at the standard tempo, but the tempo can be altered in the production process if desired. The sections then contain a folder with the midi for that section (usually slower, depending on the complexity of the piece). The midi is then followed by a folder for each measure which in turn contains a folder for each voice. After the first level the score can be navigated both vertically and horizontally. This means that the user can jump from section to section, measure to measure or voice to voice. This affords the user the luxury of relating the information with its context in a way similar to a sighted user, who can overview the score at several levels of abstraction.

Site content licensed under a Creative Commons Licence.
Software licensed under the GNU General Public Licence.

Creative Commons Licence Logo