Trevor Eddolls Arcati Mainframe Yearbook Virtual IMS and CICS user groups IBM champion Hypnotherapy
         



Pin It

Monday, 28 October 2013

Speakwrite technology today

The novel, Nineteen Eighty-Four, was written in 1948 by a then very ill George Orwell. The novel has given us “Big Brother”, “Room 101”, “thoughtcrime”, and other frequently quoted names and ideas. One that hasn’t caught on is the speakwrite machine that Winston Smith and others use to enter text rather than typing. Although software that understands human speech is available, there seems a reluctance to use it. Nuance, the people who produce Dragon NaturallySpeaking, have recently addressed some of these ‘myths’.

The first myth they address is people saying: “I don’t like talking to myself”. Their response is to suggest that when voicemail was introduced, people used to comment that they disliked using the system. This was reflected in the brief messages they left. As they got more familiar with voicemail and more relaxed using it, they accepted it as a useful and natural extension to the phone service. This pattern of acceptance was also reflected with the emergence of mobile phones. In the early days, when used in a public or confined space like a train, the mobile user often spoke quietly and terminated the call promptly. Today, with the ubiquity of the mobile phone, users are more relaxed, and often chat at length – sometimes using a Bluetooth headset – with their conversation forming a natural part of our environment and surroundings. Similarly, with desktop speech, you’re not talking to yourself; you’re writing a report, story, invoice, or inventory list, booking a table, dictating a memo or a text. In short, you’re getting on with your daily life using the most natural interface available – your voice!

The second myth they address is: “Speech recognition doesn’t work with my accent”. They answer that by saying that regions are often distinguished by their local accents as much as by their geography. To allow for that, speech recognition is engineered to understand speech even delivered with a heavy regional accent. Desktop speech recognition offers users the option to select the accent type that most accurately matches their own to ensure consistent transcription accuracy when a user is speaking fluidly. Speech solutions from Nuance recognise and understand regional accents and its applications can be optimised for a particular region to improve their accuracy. It programmes its solutions to identify multiple pronunciations of certain words to improve understanding – for example, the Irish accent sometimes pronounces the word “three” as “tree”. Nuance currently recognises 52 different pronunciations of the word “Heathrow” for British Airways to cater for those whose first language isn’t English. Nuance also takes into account that languages and the meanings of words develop over time – while the word “cool” refers to a temperature, it is also a synonym for “yes” – the solutions need to be updated to accommodate the changes. This means the power of desktop speech recognition can be enjoyed across the country, both now and in the future.

Their third myth is from people saying: “I don’t feel comfortable with a headset on”. Nuance suggests that no-one likes to feel like they’re trapped at their desk, whether at home or work. Aesthetically, some people associate cables with clutter, a blight in an otherwise tidy environment. That’s why Dragon NaturallySpeaking and Dragon Dictate for Mac are available with a wireless headset. This helps users unleash their creativity because they can wander around the office, home, or garden and lose no time capturing the creative thoughts that have been encouraged by their mobility. Not being tethered to the PC also means that anyone who suffers discomfort from sitting too long, can stand or move around and exercise without interrupting what they were doing, or losing their chain of thought or ideas. To extend this convenience and flexibility further still, they even provide the option to use your iPhone or Android-based phone as an input device to Dragon.

The fourth myth is one about taking too long to correct. Nuance proudly tells us that Dragon has accuracy rates of up to 99%, which many first time users achieve close to after a brief training process. The longer you use Dragon, the more accurate it becomes, by learning how you speak and optimising its recognition to match your pronunciation. Therefore, you won’t be correcting as many errors compared to when typing. Simple commands like ‘scratch that’ or ‘forget that’, or correct a specified text, makes editing quick and easy for the few times you’ll need it. Furthermore, Dragon will also read a document out loud, to help you further ensure it reads just as you intended.

And the final myth they try to dispel is from people who say: “I tried it a long time ago and I don’t think I need it today”. Nuance responds to this by suggesting that how we live and work has changed, so there’s a greater need for the speed and convenience delivered by speech recognition. Whether multitasking at work or at home, we’re all trying to cram in more of the things we need to do. Dragon frees you up to do these and the things you want to do. It gives students more time to study and research because it takes away the erroneous task of writing lengthy pieces of study or coursework. Busy parents can create essential to-do or shopping lists quickly, leaving them with more family time. For senior users, it opens up a door to all the information, entertainment, and communication capabilities of the Internet-connected PC, without reliance on the keyboard. In the modern competitive workplace, where speed and information exchange are essential, it gives users a powerful way to stay ahead!

I’m not suggesting that you rush out and buy Nuance products for one moment – but we’re getting used to automated telephone systems that recognize what we’re saying, so it might be interesting to revisit using voice to input data to a computer, whether that’s just jotting down some notes as you think of them or writing a full-blown report.

Maybe George Orwell’s speakwrite machine will become more of a reality in 2014.

If you need anything written, contact Trevor Eddolls at iTech-Ed.
Telephone number and street address are shown here.

 

Valid XHTML 1.0 Transitional Valid CSS!