From text to speech conversion technology is not regarded as nothing new, but at the moment this technology, there are many imperfections in the section. And now it is experiencing the greatest difficulties, and further iterations breakthrough technology bottlenecks. This means that if you want to simulate real human voice to text through dialogue, then this can be a very difficult task.
However, humans never give up exploring in this area, and even has paid a considerable amount of effort, recently, Microsoft executives announced the news, they seem to have made in this technology area where some small breakthrough. Microsoft made this model in practical application may further simplify the conversion process, even ordinary people can quickly get started using this feature. The AI conversion software is developed by Chinese researchers, based on 200 speech samples, they were able to create a sound more authentic voice. Zhengzhou leukoplakia Hospital: http: //www.zybb120.com/
The key factor is the technological breakthrough of "Transformers", "Transformers" is a deep neural networks, designed to mimic the neurons in our brains. By using the converter, which helps it to process information more efficiently. So far, from the results, scores of new models in the word intelligibility aspects of 99.84%, although there are reports that it still sounds a little mechanical. If you want to see for yourself that you can search for a sample posted on GitHub to audition.
We have to admit that the current sound effect is still quite real, the level basically reached double Google's artificial intelligence technology. Although these technological advances are welcome, but we must also be concerned about the ability of artificial intelligence if fraud is too strong, we must be vigilant and know whether fear it could be misused to propagate error message.