wavenet

During training the DeepMind team gave WaveNet real waveforms recorded from human speakers to learn from .

培训期间，DeepMind团队让WaveNet学习了一些真实记录的人类语音波形。

The research that led to WaveNet , along with a number of samples of speech generated by the system , was published online on Thursday .

周四，WaveNet的相关研究以及一系列该系统生成的语音样本被发布到网上。

WaveNet however overcomes this problem , by using its neural network models to build an audio signal from the ground up , one sample at a time .

然而WaveNet克服了这个难关，利用神经元网络模型从头建立一个音频信号，每次生成一个样本。

Named WaveNet , the model works with raw audio waveforms to make our robotic assistants sound , err , less robotic .

这项名为WaveNet的技术通过研究原始音频波形，使机器人助手的声音听起来不那么像机器人。

WaveNet doesn 't control what the computer is saying , instead it uses AI to make it sound more like a person , adding breathing noises , emotion and different emphasis into senteneces .

WaveNet并不会控制计算机的说话内容，它只会应用人工智能技术在句子中添加呼吸声、情感和各种重音，从而使计算机语音听起来更像真人。

The result was that the WaveNet learned the characteristics of different voices , could make non-speech sounds , such as breathing and mouth movements , and say the same thing in different voices .

所以WaveNet学习了不同声音的特点，可以发出非语言声音，比如呼吸声和嘴部活动的声音，并且可以用不同的声音说同样的内容。

One person familiar with the research said the system , called WaveNet , was not being employed in any of Google 's products yet , and that the massive computing power it needed to generate speech made it impractical for real world use for the foreseeable future .

一位知情人士表示，这一名叫WaveNet的系统尚未被用于谷歌任何产品。该技术在生成语音时需要极大的运算能力，令它在可以预期的未来无法实际运用到真实世界。