Reconstructing the recorded sessions in Janus
- The output of the Janus (WebRTC) video calls are .mjr files, and for everybody in a session, we have three .mjr files, video, audio, and data. Therefore, we should convert the files to visual audio files that are understandable by humans. The link showed how to convert the raw mjr fils to mp4 and wav for video and audio files respectively, and finally merge them to each other by ffmepg.
- The main challenge of the session reconstruction project is the time that we have multiple persons in a session, and their entrance and exit time to the session is different. Therefore, we should consider their time in the session and create a video-audio file that consists of all the people with their correct attendance time in the session.
- The project was implemented by python and handled different situations for the session.
Due to the business laws, I can't share all the codes here. If you have any questions you can ask me. Please feel free to contact me.