"Few dolls can be found to complete speech dialogues after running all over the capital's children's toy market", which is the embarrassment that the reporter encountered during the eve of children's day.
    Coincidentally, before that, the Chinese Academy of Sciences costing 310,000 yuan for making a special robot named "Beiqi", which can complete a portrait according to the human voice commands for Sichuan Science and Technology Museum was just meant to entertain and display to audiences.
    Insider analyzes that speech recognition technology has two directions: First, the large vocabulary continuous speech recognition system is mainly used in the computer dictation machines, as well as telephone network or the Internet combining to voice information inquiry service system, these systems are implemented on a computer platform; Another important direction is small, portable voice applications such as mobile phones dial, auto equipment voice controlling, intelligent toys, remote control of household appliances and so on, these applications generally use a particular hardware system to achieve. In recent years the rapid development of speech signal processing chips and the emergence of     systems on speech recognition chips created favorable conditions for the extensive application of speech recognition.
    At present, the China Mobile Value-added Service has made some breakthroughs through the speech recognition technology. before this, Jiangsu Mobile, Shandong Mobile and other operators, has allowed the user to find a song or a ringtone through "Humph" a melody, without knowing the name of the song, broadcasting and downloading are also provided.
    "This large-scale commercial applications, has entered a mature phase." Related people in Speech Acoustics Laboratory of the Chinese Academy of Sciences said that in the commercial applications, the most critical part of the system is to solve the charging problem, telecommunications operators and the Institute of Acoustics can charge pro-rata by the amount of downloaded quantity of each search.
    "China has more than over 800 million mobile users, accounting for the world's largest mobile phone market." Chambers is full of expectations to China's 3G market. According to him, Nuance has reached an agreement with six banks, including Bank of China. In the future the speech systems will boom in Chinese market.
    Speech Recognition Technical Barriers still exist
    According to media reports, in 2006, China Telecom launched a new business called "Numbers can do everything", which is based on an integration of the user number information and providing new products through the demand in information service such as users' fuzzy query. Unfortunately, results of fuzzy query are not accurate the present Business "Numbers can do everything" needs to be achieved by manually answering the call, which will necessarily require to invest a huge amount of funding to build and continuously improve the call center.
    Chambers said it will step up efforts to develop the Chinese market and investment in China. This year, employees in Nuance will be doubled, and the research and development will be strengthened, to improve the accuracy of Chinese language recognition.
    At present, Google and Microsoft are competitive in speech recognition technology. There is only a few companies in the world can form the core technology, Acoustics Institute of Chinese Academy of Sciences is the only unit with autonomous core technology, the major competitor of it is the Nuance Corporation of United States.
    The global financial crisis coincides with the first year of China's 3G business. The applications of speech recognition technology will be a golden opportunity. It will also be a long time to break the barriers.