中国计算机产业的下一个亮点——汉语语音合成的实用化

The Next Hot Research of China Computer Industry——The Practicability of Chinese Voice Synthesis

现代社会已经进入数字化信息时代 ,网络技术和多媒体技术获得迅猛发展 ,计算机与人之间的交互日益频繁。如何使电脑具有类似于人一样的听、说能力 ,成为自 90年代以来信息产业的研究热点。要建立一个具有听、说能力的计算机语音系统 ,必需的两项关键技术就是语音识别技术与语音合成技术。同语音识别技术相比 ,语音合成技术相对成熟一些 ,是该领域中近期最有希望产生突破性进展并形成产业化的技术 ,而汉语语音合成的实用化更将成为中国计算机产业的下一个亮点。近几十年来国际和国内对于语音合成技术的研究主要集中在按规则进行文语转换 ,即将书面语言转换成口头语言。到目前为止 ,法语、德语、英语、日语等语种的文语转换系统都已经研制成功 ,相对而言 ,中文语音合成技术现在还尚未达到实用化的要求。本文对于当前语音合成中热点的文本分析、韵律生成、语音合成三项关键技术进行了剖析 ,并针对中文的文语特点 ,指出了中文语音合成技术的难点所在。勿庸置疑 ,中文语音合成技术具有非常惊人的市场潜力 ,因而必将成为国内外IT业争夺的重点。虽然国内的语音合成技术起步较晚 ,但是我们拥有其他非汉语言国家所不能相比的优势。汉语对于我们来说是如此熟悉 ,以至于我们可以说

With the coming of the digital information era, network and multimedia technology are developing in a tremendous speed. The interaction between computer and man is increasing greatly. How to make the computer have the same listening and speaking ability as human being has become the focus of research of the information industry since 1990s. To establish a computer system which has listening and speaking ability, Voice Identification and Voice Synthesis are the two key technologies. Comparing with the Voice Identification technology, Voice Synthesis technology is somewhat more mature and is the most promising technology which can bring forth breakthrough development and realize industrialization. Meanwhile, the utilization of Chinese voice synthesis will become the next hotspot of China computer industry. In the past decades, the domestic and international research of voice synthesis technology had mainly focused on the text to speech transition according to the rules, that is, to transform the written language into oral language. Up to now, the text to speech systems on French, German, English and Japanese have come into being. However, Chinese Voice Synthesis technology can not yet meet the requirements of utilization. This paper analyzes Text Analysis, Rhythm Generation and Speech Generation, the three key technologies which are the hotspots of voice synthesis, and points out the difficulties that may come up according to the characteristics of Chinese language. No doubt, Chinese voice synthesis technology will have an amazing market potentiality. It will surely become the key point of competition among domestic and foreign companies. Although out domestic voice synthesis technology was launched comparatively late, we have some matchless advantages that the other non-Chinese speaking countries lack. Chinese is a language that we are so familiar with, we can even say that we have a talents team which can well master the Chinese characters and its voice processing technology. Also, we have outstanding achievements in fields of Chinese input, Chinese output, Chinese composition and Chinese OCR. Therefore, the real practical Chinese voice synthesis system should and will be developed successfully in China.