I guess you are talking about complexity of huffman coding.
1.We have n keys,and we need to merge tree n-1 time to generate final tree.
2.we use min heap to store frequency of every node.
3.In ever iteration we find out top two min frequencies ------->logn time
4.Then we merge those two frequencies and insert in the tree again---->logn
we repeat this process n- 1 time,then total complexity of huffman coding is =(n-1)*(3logn)=O(nlogn)