This Zhihu problem is quite old. It seems that it was put forward in April of 20 1 1. Answering this question today is very different from answering this question two years ago, because the recognition rate, application scope and market structure of speech recognition at 20 1 1 are different from today. In today's 20 13 years, Yunzhisheng and Iflytek are the most active and outstanding voice recognition in China. Yunzhisheng was founded on 20 12. Although it only has a history of 1 years, its core team of speech recognition technology has been in business for more than ten years, which is why it can make great progress in the field of speech recognition in just one year. Micro Voice Plug-in, sogou Voice Assistant, LeTV super tv, Lao Luo Hammer Operating System, TouchPal Input Method and Yixin launched by Netease+Telecom all use Yunzhisheng's voice recognition. 20 13, Yunzhisheng is highly recognized in the capital market and highly sought after. Iflytek is an enterprise within the system hatched by China University of Science and Technology. Engaged in speech synthesis for many years. Since 20 1 1, he has been engaged in large-scale speech recognition, including Iflytek dialect and Iflytek input method. Speech recognition technology has a threshold. At present, Yunzhisheng and Iflytek have released public cloud and provided industrialized services. In addition, all three BAT companies have begun to do research and development of speech recognition technology, and there are no good products yet.