Computer Input: Speech / Voice Recognition

 

Voice input, sometimes referred to as speech or voice recognition, allows you to enter data and issue commands to the computer with spoken words. It is an alternative to typing on a keyboard. The software has been developed to provide a fast method of writing onto a computer and can help people with a variety of disabilities. It is useful for people with physical disabilities who often find typing difficult, painful or impossible. Voice recognition software can also help those with spelling difficulties, including dyslexic users, because recognized words are always correctly spelled. Put simply, you talk to the computer and your words appear on the screen.

If you talk with pauses between your words, computers have been able to understand human speech since the early 1970's. However, no one wants to talk like that, at least for very long. The goal has always been to enable computers to understand our continuous speech. There are no pauses between our words when we talk normally.

Voice recognition promises to be the easiest method for data entry, word processing and conversational computing, since speech is the easiest, most natural means of human communication. Voice input has now become technologically and economically feasible for a variety of applications. Early voice recognition products used discrete speech recognition, where youhad to pause between each spoken word. New continuous speech recognitionsoftware recognizes continuous, conversationally paced speech.

Voice recognition systems analyze and classify speech or vocal tract patterns and convert them into digital codes for entry into a computer system. For a computer to decipher the signal, it must have a digital database, or vocabulary, of words or syllables, and a speedy means of comparing this data with signals. The speech patterns are stored on the hard drive and loaded into memory when the program is run. A comparator checks these stored patterns against the output of the A/D converter.

In practice, the size of a voice-recognition program's effective vocabulary is directly related to the random access memory (RAM) capacity of the computer in which it is installed. A voice-recognition program runs many times faster if the entire vocabulary can be loaded into RAM, as compared with searching the hard drive for some of the matches. Processing speed is critical as well, because it affects how fast the computer can search the RAM for matches.

All voice-recognition systems or programs make errors. Screaming children, barking dogs, and loud external conversations can produce false input. Much of this can be avoided only by using the system in a quiet room. There is also a problem with words that sound alike but are spelled differently and have different meanings – for example, "hear" and "here." This problem might someday be largely overcome using stored contextual information. However, this will require more RAM and faster processors than are currently available in personal computers.

Several programs are available that provide voice recognition. These systems work best on Windows 2000 and Windows XP. A number of voice recognition programs can be used with Windows, including a basic one that is supplied with Microsoft Office XP and 2003. Most special voice applications include a software CD, a microphone headset, a manual and a quick reference card. You connect the microphone to the computer, either into the soundcard (sockets on the back of a computer) or via a USB connection. Then you can begin talking using the following steps.

 

Enrolment

Everybody sounds slightly different, so the first step in using a voice recognition system involves reading an article displayed on the screen. This process, called enrolment, takes less than 10 minutes and results in a set of files being created which tell the software how you speak. The enrolment only has to be done once, after which the software can be started as needed. The new pieces of software claim that the enrolment process is even easier then in previous versions.

Dictating and Correcting

 

When talking, people often hesitate, mumble or slur their words. One of the key skills in using voice recognition software is learning how to talk clearly so that the computer can recognize what you are saying. This means planning what to say and then delivering speech in complete phrases or sentences. The voice recognition software will misunderstand some of the words spoken and it is necessary to proofread and then correct your mistakes. Corrections can be made by using the mouse and keyboard or by using your voice. When corrections are made the voice recognition software will adapt and learn, so that the same mistake will not occur again. Accuracy should improve with careful dictation and correction.

Editing and Formatting Text

 

Text can be changed (edited) very easily. The text to be changed can be selected (highlighted) by using commands like “select line”, “select paragraph” and then the changes can be spoken into the computer. These will then replace the selected text.

Typically, voice recognition systems with large vocabularies require training the computer to recognize your voice in order to achieve a high degree of accuracy. Training such systems involves repeating a variety of words and phrases in a training session and using the system extensively. Trained systems regularly achieve a 95 to 99 percent word recognition rate. Training to 95 percent accuracy takes only a few hours.

Two examples of continuous speech recognition software for word processing are Naturally Speaking by Dragon Systems and Via Voice by IBM.

Dragon Naturally Speaking

This program is distributed by Nuance. NaturallySpeaking is recognised as the market leader and is the alternative most frequently recommended by AbilityNet.

IBM ViaVoice

This is also distributed by Nuance. It offers good accuracy, but is not as easy to use as NaturallySpeaking.

Qpointer

Qpointer provides good command and control facilities, but is not so good for writing tasks as is makes more recognition errors. It operates differently to Naturally Speaking and Via Voice.

Minimum requirements are a 133 MHz Pentium class microprocessor, 32 MB of RAM, an industry standard sound card, and 50 MB of available hard disk capacity. The products have 30,000-word vocabularies expandable to 60,000 words, and sell for less than $200.

 

Comments:

dyslexic дислексія, невміння писати

 

accuracy точність; безпомилковість; чіткість зображення

 

to decipher розшифровувати, дешифрувати

 

to convert перетворювати, конвертувати

 

comparator компаратор, порівнювач, блок порівняння;

 

enrolment реєстрація

 

I. Match words in the text with their definitions.

 

1. Recognition A. Esp. tech. separate; distinct.

2. Feasible B. A person, thing, or form that is an example to copy.

3. Discrete C. To do something in answer.

4. Accuracy D. The fact of knowing someone or something.

5. To respond E. Able to be carried out or done; possible.

6. Pattern F. The quality of being accurate;

exactness or correctness.

 

II. Identify whether the following sentences are true or false. Use the model:

Student A: I’ve heard that IBM ViaVoice offers good accuracy, but is not as easy to use as NaturallySpeaking. – Student B: Yes, that is quite right.

2) S. A: All voice-recognition systems are faultless. – S. B: No, that is false, because all voice-recognition systems make errors. Such noises as screaming children, barking dogs, and loud external conversations can produce false input.

 

1. Early voice recognition products used conversationally paced speech recognition.

2. Training to 95 percent accuracy takes only a few hours.

3. Voice recognition devices in work situations provide faster and more accurate input.

4. The sounds of a sequence of words are like a series of sound bites.

5. Trained systems regularly achieve a 95 to 99 percent word recognition rate.

6. Voice recognition systems work the best only on Windows 2000.