ABSTRACT
Whispering has been studied with the generally aim of recognizing words in conversations of interest such as in the forensic field [1]. Also, whispered voice has been ex- amined mostly to attempt the reconstruction of normal* speech [2–4], yet it has never been the other way around. In this project, the pursued goal is to get a whispered ver- sion from a normal voice recording.
∗ Normal voice/speech is considered speech voice when talking at a normal conversational level.