Most people can’t type as fast as they can talk, and even virtuosic typists can’t simultaneously do another manual task. And so voice-recognition software can seriously boost productivity. Helen Bradley looks at the new-breed software, and how it is expected to develop in the future.
In a memorable episode of the popular The Games TV show of a few years ago 1999, John Clarke and John Dawes try frantically to open Gina Riley’s computer which has been locked by a voice password. Their antics combined with the haphazard typed results that Gina gets from her efforts at dictation make for a humorous take on voice-recognition software.
All this prompts the question of just where is voice-recognition in the scheme of technological innovations. Is it still the butt of jokes or is the joke on those of us who haven’t yet discovered its benefits?
Since before computers hit the mainstream, the promise of voice-recognition has been with us. The idea behind it is that you can talk to your computer and it will respond to commands and type what you say. The reality has been some time coming to fruition but everything is pointing to the current voice-recognition software offerings as being the next big thing in data entry.
Voice-recognition software has two characteristics. On the one hand it offers the ability to run your computer using voice commands. Say “open Internet Explorer” and it will launch your browser and you can surf the web by voice rather than keyboard or mouse. On the other hand, voice-recognition software converts the spoken voice into editable type, so you can dictate and it types into the application of your choice.
The key players in the voice-recognition software market are Nuance with its Dragon Naturally Speaking, IBM’s Via Voice, and Microsoft’s speech tools. However, this list can be pared down to just two players—IBM awarded Nuance’s predecessor, ScanSoft, global distribution rights to IBM ViaVoice in 2003, and while the program is still available it hasn’t been updated since then and is unlikely to be in the future.
Dragon Naturally Speaking has, by most estimates, around 90 percent of the voice-recognition market. The latest release of the software, Version 9, boasts, for the first time, training-free installation which means that out of the box you can expect to achieve around a 90 percent recognition rate. Within a short time this increases as Dragon becomes more familiar with your voice and the patterns of your speech.
To help train the program you can point Dragon Naturally Speaking to things like your email software and it will learn not only the names of the people to whom you send messages but also your speech patterns from your outgoing messages. According to distributor Neal Gerber, regional manager for Voice Perfect Systems, success with the software can be enhanced with suitable hardware. The program requires a quality sound card, and, while this sounds like a small ask, in the business world unless a user is creating multimedia files, sound cards are low on the list of upgrades when purchasing a new office computer.
“But you can work around the limitations of a poor sound card by using a USB headset with an external sound card,” Gerber suggests. “This bypasses the computer’s sound card to render a high quality voice stream to the Dragon Naturally Speaking software, increasing the quality of the speech-recognition process.”
Additional hardware
It’s also important to have a good quality microphone. A noise cancelling headset will filter out surrounding and ambient noise giving a clearer voice stream and allowing for optimal recognition rates.
Ian Wilson, executive director of Boyden Global Search executive recruiters, cites independence as one of the main benefits of using this software. Having used Dragon Naturally Speaking software for some 14 years, he has seen it grow and develop. His radio microphone lets him walk around the office as he dictates. And when he’s away from his computer he dictates into the mobile handheld recording device and then plugs it in and lets the program handle the transcription process.
And you might be surprised how much you can use voice recognition programs day-to-day. Asked how often he uses the program, Wilson quips, “every 10 minutes”. Much of his work involves the written word and he uses the software not only to type his reports and interview notes but also to answer emails. “I use Microsoft Outlook as well as Lotus Notes for email, and one of the pluses of the new version of Dragon Naturally Speaking Professional is that it not only works in any software but all the correction tools are available when you’re working with them too,” he says.
These correction tools include the ability to highlight a mistyped phrase or sentence and have the software play back what you said so you can refresh your memory. “I like the ability to have the program speak back the finished typing so I can listen to this as I work on something else, like packing my bags to head home after an interview,” he adds. Wilson has a Scottish accent, and he points out that the program he uses “relies more on clarity of diction than actual accent, so if you speak reasonably clearly it will work for anyone”. He reports out-of-the-box accuracy of about 90 percent, and after about 20 minutes of speaking to it around 95 percent accuracy. “Then it takes you the rest of your life to get to 100 percent,” he adds wryly.
If you’re considering voice-recognition software, it’s important to understand that it requires you to dictate, and not simply speak into the microphone. You must form the sentence in your mind and then dictate it to the computer, including dictating punctuation marks such as commas and full stops. For many new users this is an unfamiliar process and, as Gerber suggests, unfamiliarity with the dictation process is often the reason why newcomers give up on the program.
The Dragon Naturally Speaking product is provided in two packages for business: Professional and Preferred for Home Office. The Home Office version starts at $399, ranging up to $1,585 for the Professional version. The difference between the versions is in the extras, such as the ability to use Dragon Naturally Speaking in programs other than Microsoft Word and Outlook Express, and the ability for the program to ‘speak’ your text back to you. The Professional version can be used in any Windows software, it supports multiple custom vocabularies and offers network tools for centralised management so you can move from one computer to another without having to retrain the software.
Voice recognition applications
While voice-recognition software obviously has a big place in the working lives of those with repetitive strain injury (RSI) and who can’t type or use a keyboard for various reasons, it has a much wider potential audience than this. Most of us talk at around 160 words per minute but type at only one third of that speed. By dictating to the computer a user can typically increase their productivity simply by increasing the speed at which they produce content.
A business can also reduce or eliminate its transcription costs whether these are paid to an outside service or to an in-house secretary or typist. And voice-recognition software can free up support personnel from being dedicated to transcription tasks for proofreading and editing tasks instead.
Dragon Naturally Speaking, in particular, has a large client base in both the medical and legal fields where the program is used by doctors and radiologists to dictate reports, and by lawyers in their daily work. There are specialist medical and legal dictionaries which provide words typically used by different medical specialists and lawyers. In addition, the program is available in geographic region versions, with dictionaries compatible with the country and which are suited to the accented English spoken. Gerber’s Voice
Perfect Systems business offers additional services to users such as government or vertical market businesses in customising the program and supplying dictionaries and other tools specific to that particular industry or organisation.
Norman Isenberg, principal of Gibsons Lawyers, uses Dragon Naturally Speaking Legal version largely when creating documents from scratch. “I find it useful to see the document on the screen as I work, so I can see what I’m creating rather than sending it to be typed and having to wait until it comes back,” he explains. “I also use it to write and reply to emails.”
Isenberg had tried voice-recognition many years ago but, at the time, was unimpressed with it. However, more recently he trialled Dragon Naturally Speaking, found it useful and now two other principals in his firm and most of the junior lawyers use it too.
While Dragon Naturally Speaking has the lion’s share of the market today, the might of Microsoft is not to be underestimated. Until the release of Office XP, Microsoft had no visible consumer product in the voice-recognition market and even the speech tools in this and the later version, Office 2003, could best be described as lacklustre.
However, with its new operating system, Windows Vista, Microsoft is making a more significant move into this market. By all reports the voice-recognition features in Windows Vista are more sophisticated. Not only will it type dictated text, but you can also use it to control the computer interface, including browsing the web. While Vista lacks the functionality of the Dragon Naturally Speaking Mobile version, in that you can’t pre-record your dictation, it does come built into the operating system, and is therefore free and available to anyone using it. It is expected that Vista’s release in early 2007 will bring voice-recognition to a wider audience than has been exposed to it in the past.
Voice-recognition software already has a secure foothold in many offices and its uptake is likely to increase in the next few years. For any organisation with staff who spend a lot of their day typing, voice-recognition is a worthwhile investment not only because it reduces the amount of typing and thus lessens the likelihood of RSI injuries but also because of the productivity increases it offers.
