Editing in progress. Nothing to see. Move along please.

A screen reader is a program mainly used by blind, partially sighted or dyslexic individuals. In essence the software converts the written text, graphics and other aspects of the visual screen into synthetic speech output through a hardware or software speech synthesiser. The screen reader software runs 'in the background' whilst the user runs application software such as word processors or databases.

A screen reader lets a blind user select and read the sections of
the screen he/she wants to hear. For example, sighted users
stop reading a menu when they come to the item they want, or stop reading a web page when they reach the link they are after; users of screen readers use them to silence the synthesizer's output when they've heard it. The ability to filter out irrelevant information is often as important as the ability to hear relevant information.

The screen reader software is normally adaptible to a users needs, since the speed of the voice, the pitch, the way it announces punctuation and other aspects of the sound can all be easily changed. For example, the software can be set to announce each key as it is pressed, useful for less confident touch typists, or to announce each word as the space bar is pressed, or when the 'Enter' is key hit. Alternatively many confident touch typists will turn off the speech altogether when typing for prolonged periods so that the speech does not speak the words or letters as they type. The text written on the screen is able to be reviewed using the up and down cursor to read a line at a time, 'Ctrl' and cursor left or right to read a word at a time, and other such functions to read paragraph at a time, or read the entire document straight through.

The screen reader software also contains features to allow the user to perform functions normally done with the mouse via the keyboard. So, pressing the 'Alt' key with the letter 'F' brings up the file menu, whilst 'Alt' key with 'E' brings up the edit menu etc. Simularly, in most screen readers, pressing the tab key will move through a webpage a link at a time, whilst shift with the tab key will move back a link at a time. Similar functions exist to read specific aspects of the screen, such as the title bar, the suggested correction in a spell checker, or the contents of a cell in a spreadsheet.

Screen readers are however not only useful for people who are visually impaired but are increasingly being used in situations where people are unable to look at a computer screen, such as in the dark, or even for in-car computer-systems.

History of Screen Readers

Video displays were first used with computers in the early 1960s but did not become common until the early 1970s. Early terminals used a printer rather than a video screen for output; text was a one-dimensional stream of "scrolling" text which could be fed into a speech synthesizer connected between the computer and terminal. The only problem was that the user had no way to re-read text that had already been spoken.

The real development in screen readers however began in 1982 with the release of IBM computers, these provided a standard interface for connecting external devices, combined with a standard operating system and the possibility of installing much more memory with which to run more powerful software, including screen readers. By the late 1980s, several screen reader programs were commercially available, including Enable Reader, Soft Vert and the Enhanced PC Talking Program.

Speech synthesisers

The speech synthesiser is the piece of equipment that produces sound from text in the form of a synthetic voice. This is either an external synthesiser- a box that plugs into the rear of the computer, a 'software synthesiser' which operates through the PCs soundcard, or a card that plugs into the PCMCIA slot on a laptop PC.

Probably the most famous user of synthetic speech in the world is Dr Stephen Hawkins, Lucindian Professor of Mathematics at Cambridge University. He uses the DECtalk speech format synthesiser which gives him the distictive voice. He tried a newer one a few years back but he didn't like it, as the synthetic DECtalk voice had over the years obviously became his own.

External synthesisers

DECtalk was the best hardware synthesiser on the market during the early 1990s, and is still fairly popular today, with newer versions of it still being availible. At the time the main competitor in the UK market was the Apollo II speech synthesiser produced by the UK company Dolphin systems. Whilst the DECtalk was a relitivly small piece of external equipment, the Apollo II was really rather large, but both the Apollo II and DECtalk are still used today by those that prefer an external device to a software synthesiser running through the PCs soundcard.

Apollo II synthesisers were about £500 in the early to mid 1990's whilst the smaller and some would say better DECtalk synthesiser was nearer £700. Both can still be found hiding in the lofts or even still in use.

Dolphin computer systems who made the Apollo II also produced a PCMCIA synthesiser version of the larger desk top equivelent, called Gemini, which could be plugged into the PCMCIA socket on a laptop. These like the Apollo II had a headphone socket, but because of their very much smaller size did not have a built in speaker, though one could be attached externally to the card.

Software synthesisers

From the late 1990's onwards, software synthesisers became the most popular option for synthetic speech used with Screen Readers, the main reason for this was the expense; rather than having to pay a lot of money for an external device, a standard Windows soundcard was able, with only the addition of a piece of extra software to produce equally, if not better sounding speech than the bulky external devices.

'Eloquence' is a software synthesiser that can be used with several screen readers including JAWS, whilst 'Orpheus' is used by Dolphin systems (replacing the apollo II) with their screen reader, HAL.

Kurzweil scanner software, a self voicing scanning and optical character recognition software package designed specifically for the blind or dyslexic also runs using the Eloquence engine.

Screenreaders

There are several screen reading programmes on the market currently and whilst most individuals have a preference for one of them over the others, this is mainly due to personal choice and largely because once one becomes familiar with one particular package. It is a very steep learning curve to change over to another companys' product.

The two main screen readers on the American market are Job Access With Speech (JAWS) produced by Henter-Joyce, a division of Freedom Scientific, and Window-Eyes produced by GW Micro Inc. These two screen readers are also common in the UK, with a third UK based company Dolphin computer systems screen reader HAL probably being equally as popular as JAWS, with Window Eyes in third place.

A screen reader is a memory resident program that uses a text to speech synthesizer to relay information about what is appearing on the screen. Currently, most Windows screen readers use system hooks to gain access to keyboard input and menu and dialog box messages. It will call the Windows API to determine some information about the current status. It monitors some API functions to catch changes to the cursor, focus, and other elements on the screen. Finally,
it creates an off-screen model that contains information about each point on the screen (e.g. character, font, font size, class, etc.). The user can then use a keystroke to access this information when he or she wants to find out about an element on the screen (e.g. the current title in the title bar.

The screenreader that runs using Orpheous (and can be used with the Eloquence engine too) is a UK manufactured one called Hal. One of the big advantages of HAL using the Orpheous engine is that users find it helpful to buy HAL (the speech bit like) with their magnifier and have both running together. HAL can be used as a DOS screen reader as well, such as with the Apollo II synthesiser and Gemini cards.

JAWS is another popular DOS screen reader made by the American company Henter-Joyce and can be used with a range of synthesisers, such as DECtalk.

Window-Eyes is another screenreader, which magnifies and is compatible with a wide range of speech synthesisers.

Zoomtext is the other main magnifier, but can't be integrated with speech as well, so its use is for magnification of the screen only.

Whilst this article has focussed on software such as JAWS, HAL and Window Eyes, these specialist software packages are extremely expensive, and cheaper alternatives exist on the market. However, these cheaper alternatives, often termed 'Text To Speech' systems, although they may be sufficient for some users, do not enable full functionality of the PC to be accessed by users.

Speech synthesisers on phones

Over the past few years mobile phones have become more and more complex, many now having internal memory that the user can access. Because of this, and because blind/visually impaired mobile phone owners want to be able to use text messaging and access their phone books etc, two pieces of software now exist that can produce synthetic speech through the mobile phone's speaker. These screen readers for mobile phones are limited at the current time to which models and makes of phone they work on, with most Nokia series 60 and some series 80 phones being supported by the software.

The first phone to be supplied with speech was the Nokia series 80 mobile/PDA the Nokia 9210 and the 9210I. The software that allowed these phones to be used by the blind is called TALKS, and TALKS has now extended the number of phones it works with to include most of the Nokia series 60 phones on the Symbian opperating system. TALKS, like JAWS uses the scansoft Eloquence speech engine, and for individuals who have used JAWS, the software has many similar aspects that makes learning to use it very easy indeed.

TALKS is the speech software for PDA/phone (Nokia 9210I). This also runs on the scansoft Eloquence speech engine. Others are available, such as mobilespeak works on series 60 Symbian 6 platforms but most manufactures seem to have gone for TALKS. The series 60 version just went up to 1.4, but for the series
80 (nokia communicator 9290 (USA version) and 9210 and 9210I ) they are still on something like 1.3.6.

The most up to date Communicator from Nokia, the 9500 was recently released in the UK market (it is not expected for some months in the US, but is triband so US citizens can order a European model). The 9500 is not as yet supported by TALKS. It is on the Symbian 7 platorm rather than Symbian 6 and the TALKS software developers are not yet expecting release of a version of the series 80 communicator TALKS which will work with the 9500 until early 2005.

Braille Displays

Some people use a refreshable Braille display rather than listening to the information. A Braille display is not like a computer screen. Usually, it is placed near or under the keyboard. The screen reader sends the information on the screen to the Braille display. Then, a person places his or her fingers on the display and reads the information in Braille as it appears. Braille displays are helpful to people who are deafblind and to people who work at jobs
that require them to talk to customers on the phone.

The two main problems with Braille displays is firstly that the user has to move their hand or hands off the keyboard in order to read the display. However, the main problem is their price, which puts them beyond the means of even relatively wealthy indivduals.