Lose your accent on your own

January 13, 2007 by Pablo | Leave a Comment
Filed Under Software

As a speech instructor, I have met many students who spoke excellent English, but their accent immediately identifies them as non-English speakers. Part of my job is to get them to speak in a ‘neutral’ accent. One of the challenges of teaching correct English pronunciation is that you have to coach each and every student if they are to all get it right. This can be a strain not only on the teacher but on the students who will have to endure long intervals of doing nothing while waiting for his/her turn.

(In the old days, we taught them to speak in an ‘American’ accent, but today, since there is no single American accent, we teach our students to speak in an accent that is at least clear and understandable to native English speakers.)

Through the years, there have been audio tapes and CDs claiming to teach correct pronunciation, but my personal experience has shown that a student will truly believe he is pronouncing a word the way he is asked to, even if he is not. For a long time, technology could not teach a student ‘active listening’, that is, the skill listening to both your teacher and yourself, and realizing when your pronunciations are identical. Similarly, an audio tape may repeat a word endlessly, but if a mispronouncing student believes he is correct, the purpose is defeated.

Videos are little improvement. Like listening to an audio, a student may erroneously believe that his tongue, say, is in the same position that the video shows. It seems that nothing can replace a live teacher who will give you live feedback.

Those who think that, then, have probably never tried Automated Speech recognition.

is not new. However, like other early software, it was unreliable. At best, it would recognize you as long as you spoke slowly and deliberately. Originally, ASR was used as a voice-print ID system and a none-too-successful voice command system.

But now, with improved ASR Technology, a company called Blue Shoe Technology, Inc has added a new facet. BSTI has developed a speech recognition program for the US Army to teach its recruits to speak Spanish with the correct pronunciation.

The lesson begins with you being introduced to Capt. Bill Smith, your computer-generated 3-D instructor. The usual language lessons are there: Introducing others, ordering a meal, and so forth. Capt. Bill asks the students questions in Spanish, which they must respond to by speaking over a microphone.

But Capt. Bill is a tough taskmaster. It is not enough for a student to use the correct words in a response. BSTI’s Automated Speech Recognition analyzes your vocal response and matches it with its (correct) model pronunciation. If the student gets it right, Capt. Bill proceeds to the next question. If your response doesn’t match, then you are asked to repeat.

BSTI’s website (www.blueshoe.com) states that they offer ‘the ability to create training content in several languages’, be it ideographic, right-to-left, or left-to-right. Rather than being served a rigid program, BSTI’s customizable system allows clients to create language programs according to their individual needs. This system is usable over both the internet and the intranet.

While nothing can beat that ‘personal touch’ a live instructor will impart, ASR has the advantage of training a large number of students in the same time as one with almost equal accuracy.

Incoming search terms:

Trackbacks/Pings (Trackback URL)



Name (required)

Email (will not be publihed) (required)


Speak your mind