TMCnet News
Working at the Speed of Your Voice [Searcher](Searcher Via Acquire Media NewsEdge) Nuance's Dragon NaturallySpeaking Imagine having a personal assistant in your busy life, one to whom you can dictate your emails and manuscripts. One who can take your spoken words and create spreadsheets, reports, and documents on command. One who can even allow you to send emails or submit search commands over the web without touching a keyboard. Sound good? In this busy world, we could all use a break - some assistance in getting things done quickly and efficiently, even when we aren't able to sit in front of a keyboard. For me, the true test was my recovery from hand surgery. At the urging of Searcher's editor, I tried something completely different as a way to meet my editorial deadlines while waiting for my thumb to recover. What I found was an amazing product that clearly demonstrates just how far speech recognition has come in the past few years. Dragon NaturallySpeaking 1 1.5 is the latest version of this bestselling voice recognition software from Nuance Communications, the company which - through both innovation and acquisition - has become the market leader in both consumer and mobile markets. Dragon 1 1.5 represents not only an effective productivity advance, but the ability to see the future of voice- driven information input, web searching, and control systems. More Than a Nuanced Change Nuance has the motto, "The experience speaks for itself." Founded in 1992, this publicly traded company has taken over the market for consumer speech products through both acquisition and product innovation. The company has amassed nearly 1,000 patents as well as a library of speech data adding up to hundreds of thousands of hours. The company focuses on consumer markets, mobile solution, enterprise customer care, healthcare dictation, and professional desktop productivity areas. (For more on the company's policies, you might want to read my interview, "The Future of Search Is ... Speech: An Interview With Nuance's Vlad Sejnoha" [pp. 24-28], in my article, "Voice Recognition Arrives!" from the November 201 1 issue of Searcher [Vol. 19. No. 9, pp. 20-29; 46]). Today, the company boasts more than 22 million registered desktop application users and has products embedded in more than 5 billion phones, cars, and navigation systems. Dragon, which was initially acquired through corporate acquisition, has evolved into a productivity tool with unparalleled accuracy and speed even at the consumer level of free mobile apps and reasonably priced desktop software. Product Details Dragon 11.5 comes in multiple versions to serve different markets. The Professional version is intended for enterpriselevel applications requiring custom commands and implementations. The Premium version is intended for professional applications. A Home version is also available for the casual user. Nuance also offers two free iPhone-based products (Dragon Go! and Dragon Search) , available for free from the iTunes App store. For this review, Dragon Premium was used. Pricing is reasonable: The Dragon 11.5 Home edition is currently selling for $74.99, and Dragon 11.5 Premium for $149.99. The Mac version, Dragon Dictate, is available for $199.99. Professional ($599.99), Legal ($799.99), and educational licenses are also available. Specifications The system requirements aren't onerous for anyone with a recent model computer system: A minimum IGHz Intel Pentium or equivalent AMD processor or 1.66GHz Intel Atom processor. A 1.8GHz Intel Dual Core or equivalent AMD processor is recommended. 2.5GB hard disk space (2.8GB for localized non -English versions) is needed. Currently supported OS are Microsoft Windows 7, Vista, XSP or Windows Server versions. The system works with Microsoft Internet Explorer 6 or Mozilla Firefox. It requires a Creative Labs SoundBlaster 16 or equivalent sound card supporting 16bit recording. Version 11.5 is optimized to work on the company's own DragonPad word processor or versions of Word, WordPad, OpenOffice Writer, and WordPerfect. For email, 11.5 supports Mozilla Thunderbird as well as Outlook, Outlook Express, Windows Man, and Windows Live Mail. One hopes that, given the popularity of Google's Gmail, this will be included soon. The software comes with a stereo communication headset - comfortable and lightweight - making voice input easier and of consistent quality. Setup is easy, requiring a DVD-ROM drive for the software installation and other user information. Installation was fast. In less than 5 minutes, I was able to use the microphone to help the system 'learn' my speech inflections and accent. Dragon's software has been designed to become more accurate over time "as it learns your word choices and writing style" as well. Key Features Information and commands are entered into your computer by speaking into a microphone. As you speak, the words or numbers appear on-screen - far faster than writing - and with almost 100% accuracy. You can fool the system if you try, but for most applications, the system appears almost flawless in English. You can play back what you've entered for easy proofreading, again without any typing or checking the screen for verification. If the system doesn't recognize your words, an on-screen prompt asks you to help the system learn - and actually remembers your keyboard or voice edits to improve future transcription. If you tend to do a lot of "ahs" or "urns" when you talk or think out loud, you can use the Acoustic and Language Model Optimizer to filter these out, selecting "Nothing but Speech," which automatically removes these pauses. You can use vocal commands to edit your text as well, using commands such as delete, bold, underline, italicize, copy, cut text, or capitalize. You can "personalize Dragon with custom word lists and voice commands that reflect the kind of work you do; set options and formatting preferences, too." This is certainly a boon for anyone using a keyboard who has any type of physical difficulty. It is also a wonderful opportunity for those needing to send an email or text message while driving or otherwise unable to use other inputs. For access on-the-go, you can record information using your iPhone or iPod as a wireless mike to transcribe later into Dragon. This doesn't just work with text files. You can work with a wide variety of software programs - from spreadsheets such as Excel, to posting on calendar systems such as Outlook, to word processors including Word and WordPerfect. As the product literature notes, you can "create reports, spreadsheets or presentations; send email, create tasks, or schedule meetings; surf the Web; post to Facebook or Twitter or download your favorite music - using just your voice." You aren't limited to simple commands in terms of web searching either: "Send email to bq and Paula Hane about Dragon" or "Search maps for Mexican restaurants in Minneapolis, Minn." will each work well with the system - pulling information from your email accounts, web searches, or other reports. You can also customize commands to help "automate routine tasks." Are you a pacer? Do you tend to think while pacing or moving about the room? No problem. As long as you have your iPod, iPhone, or wear the microphone headset, you can move freely. Do you like to hear your thoughts spoken back to you? A text to speech feature, with a synthesized voice, will read out loud text that you designate. Taking the System for a Spin Dragon clearly earns its keep with even simple word processing. The process works extremely well. Even when you try to confuse it, it generally figures out what you want. However, Dragon offers much more. Dragon allows you to control your desktop, with commands such as these: * Start (or switch to) OpenOffice Writer or some other application. * Click "Start." * Shut down computer. * Open pictures or another desktop icon. * Minimize /maximize window. For email, you can do the following: * Create new messages. * Go to body field. * Attach a file. * Check spelling. * Forward. * Check for new man. * View by sender. * Reply to all. Web searching is actually fun, using commands such as these: * "Search the web for . . . " * "Find a site about..." * "Search (information type) for (topic/keywords)." The system uses your default search engine to initiate your search. Want to post something to Facebook or Twitter? Try "Post to Facebook...(text) ..." or "Tweet ... (text) ..." Editing or further commands are easy to do after input as well. For example, I spoke a brief message and then told Dragon, "Post text to Facebook," and that last text was posted to Facebook. You can also use this by designating texts by file name. I tried mightily to fool the system, imputing text with sing-song or other exaggerated diction. I even tried off-key singing, but Dragon still figured out what I intended. The only errors I had were with a statistical report (percentage data from population studies) or a quote from Elizabethan England. Since that doesn't come up often, I would agree with its claim of 99% error-free input. The system was amazing. For anyone wanting to get away from keyboard-driven computing or having physical limitations, this is a dream. The added advantage is the amazement of watching a computer translating your words into computer search results, correcting texts, or posting to Facebook or Twitter. In fact, the final version of this article was created using Dragon - dictated from my drafts using voice commands. Easy, fast, neat, virtually error-free. My definition of successful. Ending Today Dragon NaturallySpeaking is available in six languages: English, Dutch, French, German, Italian, and Spanish. This isn't just a single version of these, either. For English, this includes standard accents, Australian-accented English, British-accented English, Indian-accented English, Inland Northern U.S. (Great Lakes area), Southern Asian -accented English, Southern U.S., and Spanish-accented English. My husband, raised in Venezuela, had no difficulty using the system despite his sometimes fractured English. Dragon is offering an amazing window into the current stateof-the-art for voice recognition and speech input for computing. As wonderful as all of this is, even more tantalizing is the future potential for translation between languages. Today the web is moving rapidly from being English-centric to seeing huge growth in Spanish, Chinese, Arabic, and other languages as the web lives up to its World Wide promise. That, however, is beyond today's Dragon products. However, Nuance is working with IBM's Watson to find ways to move into far deeper information systems that, hopefully, will someday soon cross cultural and language barriers: Software that can not only dictate or transcribe, but also translate ... but that's a story that will have to wait for the next issue of Searcher. For now, check out some of the videos on Dragon and be ready to be amazed. by Nancy K. Herther Sociology/Anthropology Librarian University of Minnesota Libraries (c) 2011 Information Today, Inc. |