초록

AI-Helper

본 연구에서는 남녀 화자 8명의 목소리를 녹음한 후 전신과 얼굴을 촬영하고 100명의 사람들에게 목소리와 사진을 짝짓는 인지실험을 실시하였다. 그 결과, 8명 중 4명 이상의 목소리를 주인공과 올바로 짝을 지었으며, 전신을 보았을 때가 얼굴만 보았을 때보다 정확도가 높았다. 또 좋은 소리는 좋은 인상보다 더 많은 사람들이 동감하는 것을 알 수 있었다.

Abstract ▼ AI-Helper

To investigate the relationship between voice and the image created by the voice, a perception experiment was undertaken in which voices of 8 males and 8 females reading a fairy tale were recorded and presented to 100 subjects along with the speakers' pictures. Two kinds of pictures were taken: one,...

To investigate the relationship between voice and the image created by the voice, a perception experiment was undertaken in which voices of 8 males and 8 females reading a fairy tale were recorded and presented to 100 subjects along with the speakers' pictures. Two kinds of pictures were taken: one, a whole-body picture, and the other, a face-only close-up picture. Subjects were asked to match each voice with its owner. The results showed that out of 8, more than 4 were matched correctly. And the correct rate was higher when the voices were matched against the whole-body pictures than against the face-only pictures. In addition, it was noted that far more people picked a single voice as the most 'favorite' one than they picked one picture as the most 'favorite' image.