In this paper, we describe an adaptive speech rate control technology for ultrafast listening that is equivalent to skimming. Nowadays, listening to audio books on mobile devices is quite common. Therefore, in the future we will obtain ever-increasing amounts of information through speech instead of conventional printed materials. People read books at various levels of detail from close reading to skimming. Although a similar feature to skimming is required to efficiently obtain information from audio sources, there is no tool equivalent to skimming for audio playback. Thus, we have developed a new speech rate conversion method to efficiently obtain information from audio sources with very fast replay. This algorithm will help not only sighted people to enjoy audio books but also visually impaired people because almost all of their information is obtained from speech. Thus, the implementation of this technology on special audio players for visually impaired people as a new replay function is expected to be useful. Moreover, this technology should be useful for all audio book listeners, not only people with limited sight.¹