It is well known that tempo affects the emotions perceived in musical clips. Generally, it is assumed that high tempo music is associated with high arousal emotions and low tempo music is associated with low arousal emotions. In this work, first we re-examine these assumptions on a large benchmark dataset for emotional analysis of music. Then, we use the ambiguous clips of this dataset to study the range of tempi and emotions perceived in these clips. We find that some of these assumptions hold true for this dataset—calm and content clips have low tempo. But, some anomalies are also observed, where high arousal clips have low tempo and vice versa. Ambiguous clips are found to have low-medium tempo predominantly, and are identified as sad and depressed mostly. In the second part of this work we use 21 Hindustani classical music (HCM) clips of 8 selected Ragas but with varied tempo (laya) to examine whether different tempo have any influence on the perceived emotions of these clips. We find that this is indeed true and identify that a change in tempo may lead to (a) a change in dominant emotions, or (b) change in intensity of dominant emotions or (c) change in secondary emotion perceived in a clip.