SSSS Recording Notes May 4, 2013, updated July 9, 2013 Recording Conditions and Equipment Dialogs were recorded with the conversants in different rooms seeing each other through a window. Acoustic separation between the rooms was good. There was a small amount of across-track bleeding, but not enough to be noticeable to human listeners at most volumes. Each participant wore a Sony DR-200 headset. The inputs were fed into an Olympus DS-2 digital voice recorder. Audio from monitoring jack of the recorder was split, to 1) sent it back to the participants' headsets so they could hear each other 2) feed the audio input of our main recording device, an iMac. The recordings were created with Quicktime on the iMac, with the video coming from the built-in camera and the audio from the voice recorder. Subsequently the recordings were edited with iMovie to remove the times where subjects were being set up and when the recording was being ended. (Recording was done using Quicktime since it has better audio-video synchronization than iMovie, despite the fact that its narrow aspect ratio wasn't ideal for our through-the-window setup.) Notes on Specific Dialogs Participants wishing to get an intuitive feel for the corpus are recommended to start by listening to dialogs 004, 006, 008, and 012; thse are rich in interesting and diverse audio regions that one might want find as results to a search. This release includes all 20 dialogs collected, interesting or not. It also includes some with flaws and some not strictly comparable to the others: - Dialogs 000-002: recorded using a MacBook (whereas the body of the corpus was recording using an iMac). - Dialogs 000,002: include participants knowledgeable about the intended uses of the recordings. - Dialogs 014, 016-018: include one non-CS major (although he/she was taking a CS class). - Dialog 000-002, 004, 007: audio recorded separately from the video and later imperfectly aligned. - Dialogs 000, 004, 005, 010, 011, 013-015: video incomplete. - Dialogs 001 and 007: topic suggestions were given strongly and/or participants took them literally. in /home/users/nigel/papers/mediaeval/recording-notes.txt