Text To Speech: Talking Computers and Why We Love them

When a computer "reads" a document and turns it into spoken words, we call it "text to speech".  The artificial production of human speech as has been possible for many years.  Only recently has voice quality improved to the point where it is difficult to tell between a talking computer and human speech.  This has led to many new applications like Text Speaker for text to speech software.

How does text to speech work?

There are two components to every text to speech tool.  First, a "speech engine" program analyzes the text (including the punctuation), and synthesizes the sounds of the words, pauses, and intonation.

The speech engine program turns words like "DeskShare" into what are called "phonemes".  A phoneme is one of the basic sounds of speech, like the "d" sound at the beginning of "doctor" or "DeskShare".  The other component is called the "voice font" or "voice." You can think of the voice as the "language pronunciation file," which is used to create the audio.

Combining the speech engine program with a good quality voice produces speech that is both easy to understand and natural-sounding.  The speech engine translates text into phonemes and the Voice turns the phonemes into audio.



People can say the same word differently.  There are regional differences, called accents or dialects.  Men and women speak in different tones.  And of course, words are not pronounced the same way in different languages.  Because of these variations, text to speech programs use various voices to speak text.  These voices can speak in many of the most common accents, and pronounce words according to the rules for different languages.  You should listen to samples of several voices, and pick the ones that sound best to you.

There are several voice technologies available.  We recommend using SAPI 5 voices if they are available.  Read Voice Software: Selecting a Text to Speech Voice for more information about voices.

Since the speech engine actually "sounds out" words by knowing how the letters are pronounced, there is no limit to the program's vocabulary.  It can read any word.

Modern text to speech software sounds amazingly life-like, and people are using windows software like Text Speaker for more and more applications:

• Education: A computer that can read aloud is an excellent tool to teach reading.  Beginning readers can follow along as the computer reads.  Text Speaker highlights the word currently being read, helping the student learn it.  Reading speed is adjustable and can be adapted to a learner at any skill level.
  Text to speech is also used in teaching foreign languages.  By having a native language voice read text, the student can correct pronunciation at anytime, anywhere.
• Audio Narration: There are many times that a script has be turned into spoken audio.  For instance, you may have a phone menu system, or a video-based tutorial that needs a voice-over.  Text to speech software can turn the typed script into MP3 files quickly and easily.  This can be especially helpful when the script changes often.  No need to wait for a human actor to be available-just change the script and use Text Speaker.
• Accessibility: One of the most important uses for text to speech systems is to assist people who have difficulty reading.  This technology can be useful to a wide range of people:
  a) The visually impaired
  b) People with reading difficulties such as dyslexia
  c) Children too young to read
• Proofreading: You'll be surprised how useful Text Speaker is in your daily work.  For instance, listening to a document read aloud is often the best way to proofread.  Certain errors, like a a double word (did you catch that?) are hard to spot when reading, but impossible to miss if you're listening.

Modern technology has made it possible for anyone to use text to speech.  Whether you're learning a new language, creating a phone voice menu, or have trouble with reading, Text Speaker can make your life easier and better.