Speech privacy invasion occurs in our every life, especially when you are with other people in the offices, houses or vehicles. People with their personal drivers faces the situation when talking on their mobiles in vehicles. There has been many trials to avoid such a situation among people. Placing a barrier between a person at the back seat and a driver is one of the ways in real. However, in acoustical point of view, playing a masking sound over speeches has been proposed as a way of keeping the speech privacy. In this case, the level of additional masking sound should be higher than the speech level or be around the same level com-pared to the speech level. The high level of masking sound might disturb drivers and distract them from safe driving.
The level of masking sound needed to reduce the speech intelligibility depends on the level of speech delivered from the back seat. As a way of reducing the required level of masking sound, active noise control has been proposed throughout this paper. Active noise control is an acoustical approach to reduce the noise level by emitting a sound out of phase with the noise. It is expected by first lowering the speech level at the region nearby the driver’s ear, low level masking sound will be enough to keep his/her speech privacy.
Active noise control, however, can bring noticeable noise reduction performance in low frequency range since the size of quite zone with 10 dB reduction is proportional to the target noise’s wavelength. This implies that active noise control can be regarded efficient only if the masking sound with low frequency range shows better performance than the masking sound with higher frequency range with the same level. Also, the performance of active noise control can vary with some factors including change of gender and sentences. Throughout this paper, it is shown that the masking sound with low frequency range brings larger speech intel-ligibility reduction than the one with higher frequency range and gender or sentence change with the fixed active noise control filter has little influence on the amount of speech intelligibility reduction.
In conclusion, if masking sound will be used for reducing the speech intelligibility in vehicles, it is better to utilize the active noise control first so that lower level of masking sound will be needed than the case without using the active noise control. The change of speaker and sentence has an effect of 1 or 2 dB per-formance degradation in SNRdB perspective of masking sound.
대화의 사생활 침해는 생각보다 많은 곳에서 발생한다. 특히 타인과 공간을 공유하는 사무실 혹은 차량 내부에서 빈번하게 발생한다. 본 연구에서는 많은 상황들 중 차량 내부 탑승자들이 겪는 통화 내용의 사생활 침해 문제를 음향학적인 방법으로 해결하고자 함에 그 목표를 둔다. 이때의 상황은 VIP 좌석 탑승자의 통화 내용이 운전자에 의하여 침해되는 것을 가정하였다. 이러한 상황을 해결하기 위한 방법으로는 뒷좌석으로부터 전달되는 음성의 인지도를 저감하는 것이다.
음성의 인지도를 저감하기 위한 방법 중 하나로 대표적인 방법은 차폐음을 청취자에게 음성과 함께 전달하는 것이다. 그러나 이러한 방법은 음성의 크기보다 큰 혹은 그에 상응하는 크기의 차폐음을 들려주어야한다는 점에서 운전자에게 불쾌감을 줄 수 있으며 나아가서는 주행에 방해가 될 것으로 여겨진다. 따라서 본 연구에서는 차폐음을 사용함과 동시에 능동 소음 제어 기술을 이용하여 뒷좌석으로부터 전해지는 음성의 크기를 운전자 좌석에서 줄이고 이렇게 줄어든 음성의 크기게 맞도록 차폐음을 전달하는 것을 목적으로한다.
