Reconstruction of natural sounding speech from whispers

A training-based speech regeneration approach with cascading mapping models

“…Whispered speech is useful for quiet and private communications in daily life [2,3,4]. Unimpaired speakers occasionally use whispers to communicate in the public locations such as libraries, cinema theatres, or during lectures and meetings.…”

“…In recent years, various techniques have been proposed for converting whispers to normal speech [19,20,21,22,23]. The driving idea of all these methods is based on the assumption of whispers are missing some acoustic and spectral features comparing with normal speech; hence, the problem of converting whispers to normal speech is formalised as a reconstruction issue [4,24]. Through this approach, these methods aim to add or enhance missing or modified features and increase the signal similarity of whispers to normal speech.…”

“…Through this approach, these methods aim to add or enhance missing or modified features and increase the signal similarity of whispers to normal speech. In general, these reconstruction methods can be classified into two major groups of train- These reconstruction methods (either training-based or non-training) suffer from range of disadvantages including problems in converting continuous speech (due to using phoneme switching) [20], being computationally expensive (due to using highly overlapped frames for spectral enhancement, or using jump Markov linear system for pitch and voicing parameters) [19,4], and more importantly lack of naturalness in regenerated output (due to simplified time alignment and spectral features assumptions) [21,23]. In this paper, we focus on a trainingbased approach, and propose a novel reconstruction algorithm to improve the efficiency in phonated speech regeneration.…”