How much do you know about voice recognition?

dragon software

A few years ago, it was a tool reserved for some professions and people for whom the use of a keyboard was difficult. But for a year, voice recognition has appeared in many everyday objects: GPS systems, mobile phones, various applications…Click here for dragon software. 

Help in the development of voice recognition dragon software

One of the keys to this development is the rise of new generation phones. Since its 3GS version, the iPhone integrates, for example, advanced voice control functions. Google deployed its voice search service in France for its Android phones. The company does not report usage figures but estimated in its first tests in the United States that a search on three performed on mobile was done through the voice.Visit this site for dragon software.

Languages in voice recognition

In this process, not all languages and speakers are equal. Some languages have “landmarks” that are more easily identifiable; the way the person articulates also plays an important role. This is why speech recognition applications fall into two broad categories: “generic”, such as Google’s voice search or the mobile version of Dragon, and “learners” like the PC version of Dragon, which create a profile of their user and learn as and when to better recognize the voice and vocabulary.

More specialized applications, such as GPS systems, use another trick to increase their accuracy. The corpus of addresses is not infinite, the software is programmed upstream to focus on a list of predefined words, which makes it easier to refine the search and reduce the error rate.

What about the imperfection?

Because if technology continues to evolve, it remains imperfect. How far? Publishers are very discreet about the error rates of their systems; averages would, in any case, be insignificant, the rate may vary greatly from one speaker to another depending on his or her speech or vocabulary. With a bit of learning, on the machine side as well as the man’s – the programs work much better with a slow and clear diction, which may require a time of adaptation – it becomes possible to dictate complete texts with a recognition rate close to 100%. But some difficulties remain unavoidable, for example, the addition of words in a foreign language in a foreign  text, which will most often require spelling the word in question.

Between the progress of technology and its trivialization in the objects of everyday life, webelieve that the conditions are ripe to address the market of the general public.

Please follow and like us: