There is provided a conference system which performs a remote conference by communicating conference data between a transmission terminal and a reception terminal. The transmission terminal receives input of the conference data including audio data, selects a real time mode of sequentially performing transmission of the input audio data and a package mode of performing transmission of the input audio data for each input of a predetermined unit amount, records the input audio data in the predetermined unit amount to generate audio record data when the package mode is selected, and transmits the input audio data to the reception terminal when the real time mode is selected and transmits the audio record data to the reception terminal when the package mode is selected. The reception terminal outputs the received input audio data preferentially over the received audio record data.