Voice input and recognition for computers promises to revolutionize the way people with disabilities will use technology. Whether the challenge is repetitive stress injury, multiple sclerosis, cerebral palsy, or motor skill impairments, speech input gives these individuals access to the world of computers, e-mail and the Internet.
Speech recognition can be used also for therapy to improve voice quality. With speaker-dependent, discrete-recognition systems, the consumer must say each word separately and distinctly for the system to recognize the spoken word. After several months of using voice recognition software, a marked improvement is often seen with a speech impediment caused by strokes or cerebral palsy.
There are different types of voice recognition systems, also called speech recognition. Voice recognition allows the computer user to speak to the computer instead of using a keyboard or mouse. The programs can be used to create text documents, such as letters or memos and can also be used with e-mail, databases and many other applications. Some voice recognition systems are very sophisticated and offer many options, others are less expensive with less power but may do the trick.
To choose a voice recognition system that will help you or your child with your specific needs you will need to understand the different classifications of the various systems.
- Static vs. Adaptive - A static system must learn each vocabulary word individually. Systems with adaptive voice files get better with use. The user corrects the software until it learns those sounds. Most systems now are adaptive.
- Speaker Dependent vs. Speaker Independent - A dependent system must learn each user's voice in order to operate effectively. Recognition improves after many hours of use as the software learns your voice. An independent system is one that works for everyone with no individual training necessary.
- Discrete vs. Continuous - Discrete speech means the speaker needs to have a small pause between words. Currently there is a rush for the method of continuous speech voice recognition. Continuous speech means that one does not need to pause between words. Testing has shown that the continuous speech model at its current level does not adapt as well for some users with disabilities. Before continuous speech systems, users were required to speak with a short pause between words, which is the way most individuals with challenges speak. This method also provides the user with a chance to correct and further strengthen their voice file
Steps to Assessment
There are several steps that will help determine if this is appropriate technology for a person with disabilities. Anyone who is seriously considering a voice recognition system should be aware of the work and time commitment involved.
- Determine the Goals - What is the need for voice recognition? Is it to navigate the computer completely or to write longer papers and assignments? Concrete goals and realistic expectations are important.
- Readiness - Is the user cognitively ready? Does she understand the operating system and basic spelling techniques? Can she sustain attention and motivation for using voice recognition?
- Environment Factors - Consider the situation in which the consumer will be using voice recognition. If he is using it for several hours a day, voice quality and articulation may be an issue. Will the system be used at both school or work and home? If the technology is not working properly, is there a backup method readily available that he can use?
- Evaluation- Finally, evaluate the usefulness of this method and keep current on the latest developments in the field. There is an ever-increasing number of voice recognition software manufacturers, which will continue to push down prices and further advance this technology.
For a listing of voice recognition manufacturers, contact LINC at 410-659-5462 or visit our website at www.linc.org.