Can a computer predict our feelings while listening to music to the point that it can predict our liking and emotional response? If it can predict these feelings, will there still be measurable differences between a participant-chosen piece of music and a similar piece chosen by computer software? This experiment was designed to answer these questions. College student participants chose an instrumental classical piece of music that they enjoyed and computer search engine chose 3 â€œsimilarâ€ and 3 â€œdifferentâ€ pieces based upon predicted aesthetic responses. They then listened to all 7 pieces and continuously reported their emotions on a scale of pleasantness and activation on a two-dimensional emotion space in LabVIEW-based software. After each song the participants recorded their mood in terms of pleasantness and activation using the Self-Assessment Manekin (Bradley & Lang, 1994. Immediately after this rating the participants rated their liking of each piece. Participants rated similarity of all 6 computer chosen pieces compared to the participant-chosen song at the conclusion of the session. A contrast based upon a repeated measures analysis of variance of the 3 â€œsimilarâ€ and the 3 â€œdifferentâ€ pieces showed a significant difference in similarity (p<0.001). Differences between â€œsimilarâ€ and â€œdifferentâ€ pieces were also significant for ratings of pleasantness emotion, pleasantness mood, and liking (all p<0.05). In addition the chosen song was more liked than the â€œsimilarâ€ or â€œdifferentâ€ songs. This suggests that choosing the music has an effect on how much a person seems to like it. These results show that our computer search engine is able to predict similar aesthetic responses, but responses in liking were higher in the participant-chosen music.