On a recording method for ambient sounds with a confidential speech

Abstract

There are many reports that environmental noise in healthcare facilities affects patients and healthcare workers both physically and psychologically. To study the sound environment in hospital wards, it is necessary to analyze live sounds (conversations, sounds of daily life, alarms, etc.) generated by human activities. Recording in hospital wards, as in all healthcare facilities, requires the protection of personal information. Sound recorded in healthcare facilities may include conversations that people do not want to hear. Therefore, it is necessary to develop a recording method that solves confidentiality issues. We have already confirmed that by fragmenting recorded sounds at short time intervals, the time-averaged sound pressure level of 1/3 octave band frequency of the sounds does not change. This time we developed a new PC application that fragments and then stores the data without storing original sound data. We analyzed the time-averaged sound pressure level of 1/3 octave band frequency of fragmented speech and alarm sounds using this new application. The results showed that the time-averaged sound pressure levels of fragmented and non-fragmented sounds were almost the same.