It is easy to record speech, but it is not easy to refer to audio recordings. If it is enable to index or summarize audio recordings, referring audio recordings will become easier. In this paper, we aim automatic extracting summarization of spoken lectures. For this purpose, at first we compared results of extracting summarization by human subjects. Then we investigate relations between linguistic surface information and human results and we got the useful surface information. Next, we made summarized audios based on this information, and we compared them with human results. Additionally, we focused on prosodic features: F0 and power. We did same experiments on them. Lastly, we combine linguistic surface information and prosodic information. As the result, we got better K-value. And those were compara- ble with human results.