Previous experiments have shown that localization with binaural recordings made with artificial heads is inferior to localization in real life and also to localization with recordings made in the ears of selected humans. These results suggest that artificial heads may be improved. A new experiment was made, employing recordings from two human heads and seven artificial heads some of which had been developed recently. The listening room setup from previous experiments was used and 20 listeners participated. As in the earlier experiments, more directional errors were seen with binaural recordings than in real life. A clear learning effect was seen over five days, emphasizing the need of a balanced experimental design. The new results show that artificial heads are still not as good for recording as a well-selected human head, although some of the new heads come close. The accumulated results from the present and four earlier studies provide sufficient statistics to conclude that there are significant differences between some currently available artificial heads.
(See document for exact affiliation information.)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
