Frequently Asked Questions
What are the CPU and RAM requirements for speech recognition?
|Dragon NaturallySpeaking Version 11|
|Pentium 4 2.5 GHz or faster reccomended; if multiple cores, 2.0 GHz or better)|
|2 GB; 4 GB for Windows 7|
|Windows XP (32-bit), Vista (32 or 64 bit), 7 (32 or 64 bit)|
|SoundBlaster MP3+ or USB Microphone|
Zephyr-TEC recommends dual core (or more) computers since speech recognition software is now optimized for these platforms.
What is the dictation rate and accuracy?
A trained user can dictate at approximately 100 to 150 words per minute with 97% or higher recognition.
How long will it take to become proficient with the software?
This is strictly user dependent. If the user is familiar with the Windows environment and applications used, becoming proficient with speech recognition software will take significantly less time than an individual who has never worked in Windows applications or used a computer. Allow 15-30 hours of using the application to establish proficiency.
Why do I need Training?
Without expert training most users fail and quit. Zephyr-TEC's experienced trainers teach you the proper skills and techniques to get you up and running effectively, successfully and productively. Training helps ease the transition into a new way of using the computer and helps you to become an expert as quickly as possible. This is not just one more piece of software to learn--this is a new and innovative way to work.
How much training is needed?
Zephyr-TEC recommends a minimum of eight hours of training for all types of users desiring a reduction in hand use. Even highly advanced users such as programmers will need to learn the process of inputting information by voice as opposed to typing. The software will recognize the user's speech much faster than the user will learn how to use the software. There are multitudes of commands that need to be memorized in order to control the software. Additionally, a dictation style must be learned because the software depends on the cadence and rhythm of the user's voice to translate sounds into text. Training becomes essential if a person has a hand injury, especially if they are on the job. Initially, a person with a hand injury will find it difficult to complete computer tasks as quickly as when using their hands.
What is the difference between the Premium and the Professional versions?
The difference between the Preferred version of Dragon NaturallySpeaking and the Professional version of Dragon NaturallySpeaking is that the Preferred version of the software has a much more limited ability to create macros. This becomes very important for a person with a hand injury because macros can be easily written to create shortcuts for any repetitive tasks. As part of our training service, Zephyr-TEC will install proprietary macro packs that allow NaturallySpeaking to work in applications such as Excel and Outlook.
Which speech software program should I use?
Zephyr-TEC recommends using Dragon NaturallySpeaking Professional for anyone with a hand injury. There are three reasons for this recommendation. First, Dragon NaturallySpeaking Professional has the highest recognition accuracy of any speech software product on the market. Second, Dragon NaturallySpeaking Professional has a number of hands-free options that the other software products do not have. Third, the Professional version has a scripting language, which allows the user to assign voice commands to a series of actions. The macro writing capability of the professional version of NaturallySpeaking expands the functionality of the software. For example, navigating a dialog box with speech software can be very tedious, however, if the macro is written the desired action can happen instantaneously. It is in this way that the injured worker can save time. With the Preferred version of NaturallySpeaking much more hand use is necessary to accomplish some basic tasks.
Will background noise interfere with my dictation?
Dragon NaturallySpeaking comes with a noise-canceling microphone. The software also has a utility that allows the user to calibrate the microphone for the background noise. However, in some instances if the noise level of the environment of the user suddenly increases, i.e. an intercom system or fire alarm, the noise may sufficiently distract the software so that it would interfere with the usage of the computer. Sudden loud noises are the most problematic when using the software.
How big is the vocabulary?
NaturallySpeaking's vocabulary dictionary is in excess of 300,000 words. The user can say virtually any English word. Jargon and commonly used phrases can be quickly added to the vocabulary of words in the dictionary. Professional vocabularies are also available for the Medical, Legal, and Safety industries. Contact us for more information. (Link Contact Us)
Is it best to do my training on my own computer?
Due to the nature of speech software, each user must create their own voice profile. With proper correction the user's speech profile will improve over time. Recognition accuracy will be greatly improved on the system where the user has spent quality time developing the profile. It is preferred that a profile is established on the machine that remains in the possession of the user. Sometimes, due to the internal noise (which is individual to each computer) a voice profile established on one system cannot effectively be moved to another system.
Who do I call for technical support?
Nuance (formerly Scansoft) offers technical support, but charges $29.95 per incident. Zephyr-TEC offers 3-hour telephone support contracts as well as on-site and Web-based support.
With what applications will the speech recognition integrate?
Dragon NaturallySpeaking Version 11 allows continuous dictation into Office 2003/2007/2010 and virtually any Windows application. With the Professional Version or higher, custom commands can often be created to voice-enable any other program, though this is highighly dependent on the program.
Can I access on-line services, E-mail and the Internet using speech recognition?
Yes, you can use Internet Explorer and Firefox with a number of voice commands in version 11. However, some web pages work poorly by voice because of their design.
Can speech recognition software be used to run accounting applications?
Yes. Continuous numbering (no pausing between numbers) makes data entry of numbers very fast.
Why do I need macro writing capability?
Included in Zephyr-TEC's introductory course in NaturallySpeaking the user learns how to use NaturallySpeaking in a wide variety of environments. Additionally, the user is given and instructed on how to use Zephyr-TEC proprietary macros that allow the user to use NaturallySpeaking in Microsoft Excel, PowerPoint, Outlook, Netscape and other applications. Time is then spent creating custom macros so that the user can perform frequent computer tasks with one command.
The productivity of the user is also enhanced with the help of the macros. To use the computer hands-free or as close to hands-free as possible, the user must expect to work slower than if they were able to use their hands. However, given the features of the Professional version, the user can expect an increase in productivity due to the increased functionality macros provide. When considering the user's ability to compete with other workers in terms of performance, speech recognition software and training is a tool that workers with upper extremity injuries need to compete in today’s employment market.
Why 2 GB RAM?
Zephyr-TEC recommends 2 GB or more of RAM (4 GB is preferred) for running NaturallySpeaking with Microsoft Office 2000/XP/2003. Because Microsoft Office and Windows use a significant amount of RAM, we have experienced many clients who required an upgrade to 4 GB RAM (or more) to use Dragon NaturallySpeaking Professional. The increased RAM can improve recognition, both in terms of speed and accuracy.
Why a dual core (or more) machine?
The most recent versions of Dragon are now optimized to take advantadge of multiple processors. Running Dragon on a single processor is still possible, but it will take longer to recognize utterances and be less accurate.
What type of sound card is required?
Speech recognition software can run on SoundBlaster Soundcards or higher and most compatible soundcards. However, we often find that a USB device gives better sound quality than an onboard sound chip, and typically reccomend USB microphones.
Will speech recognition software work on a laptop computer?
Yes, provided you have a SoundBlaster 16 or compatible sound chip or PCMCIA sound card. NOTE: Speech recognition does not always work on all laptops. Approved laptops are listed at the website of Nuance. We reccomend using a USB pod with most laptops.
What if I’m not sure it will work on my system?
The most effective way to integrate speech recognition software into a mainframe environment or with proprietary applications would be to have Zephyr-TEC perform an on-site Speech Feasibility Study.
The Zephyr-TEC Speech Feasibility Study includes: installing NaturallySpeaking on the local machine; testing the capabilities of NaturallySpeaking in all applications; enable Select-and-Say in all applications where it is not available; and look into any conflicts, errors, or disruptions in the software. From this testing we would then make a determination as to how effective the user can expect to be with NaturallySpeaking in these applications. Zephyr-TEC will then supply a written report that includes: a statement of the problem, statement of proposed solution, general information, productivity issues involved, training recommended, additional considerations, evaluation overview, job duties overview, applications utilized, telephone equipment, computer equipment required, evaluation of the environment, implementation suggestions, a list or recommendations, a time frame recommended for implementation, projection of expectations after implementation and ongoing technical support recommendations.
Should an evaluation or test program be performed to determine the feasibility of using Speech Recognition?
Yes. An evaluation or pilot program is a cost-effective way to identify potential problems involved in implementing speech recognition. It will address the equipment, environment, applications used and provides a recommendation as to which software package will be best as well as an estimate of the amount of training time required.
What’s involved with a Pilot Program?
There are many steps to a complete and successful Pilot Program. First, an evaluation of your equipment is necessary to determine if hardware is compatible or needs upgrading. We recommend at least 3-8 individuals be selected for participation in the program. A skills/needs/use assessment on the participants helps determine what type of training program will be implemented. All participants must be allowed time to achieve maximum productivity. Goals, objectives and milestones must be clearly defined and tracked to measure progress and success. For an initial consultation, please call us at 1-877-473-7497!
What if I use a system that has less than your minimum recommended specifications?
You will experience considerably more system slow downs and crashes without adequate RAM and system specifications.
Can others use the computer while your voice profile is loaded?
Yes, speech recognition does nothing to the computer to prohibit the use of the mouse or keyboard by non-voice users.
How many voice profiles can I create on my computer?
You can create as many voice profiles as you have hard disk space. Each profile is approximately 45-60 MB. Any saved .wav files will need approximately 1 MB for each recorded minute. However, a license is require for each individual who wants to have a voice profile.
What happens if I have a cold?
If you are sick, stay home! If you go to work, be careful you drink plenty of fluids and minimize the use of your voice. Speech recognition software may require additional corrections during use because your voice is not normal. It will adapt within a short time. Zephyr-TEC recommends not saving your profile after a session when you have a cold.
Can speech recognition be used on a network?
Yes. The speech software will work on standard Windows networks or with Windows emulators (such as RUMBA, EXTRA and REFLECTIONS) to allow access to mainframes and other operating systems.
Can the speech recognition be used on a UNIX workstation?
No, there is no UNIX compatible speech recognition software. However, clients that use speech recognition can access UNIX using a personal computer with an emulator to UNIX.
Will the Dragon NaturallySpeaking speech recognition software run on a Macintosh?
Yes. Dragon Dictate is available for the Mac and currently on version 2.5. It is based on the same recognition engine as the PC version, but offers less in the way of command and control. It remains hard to use a Mac computer hands-free.
Do you have any information about programming with speech recognition software?
Zephyr-TEC has trained some programmers to use NaturallySpeaking to considerably cut back their keystrokes and mousing. Since programming languages are not verbal languages using NaturallySpeaking to program requires writing many macros for frequently used code lines.
Contact us for more information regarding programming with Dragon NaturallySpeaking, but alot of the effort will need to be made on the part of the user.