Taqāsīm perception: Head-accelerometry of listeners standing still in Amman and Oslo

Accelerometer recordings of listeners' head motion while standing and listening to an Arabic maqām improvisation (taqāsīm, maqām nahāwand, 7:42) played over headphones, collected from two audiences: one in Amman, Jordan and one in Oslo, Norway. Motion was recorded using an Axivity AX3 worn atop the headphones at about 100 Hz. The record contains 64 raw recordings (33 Amman, 31 Oslo) as ts, x, y, z in g; and 63 sets of questionnaire answers. After exclusion, we included 60 participants (30 from each site). The following data are included:

• quantity of motion in mm/s, band-limited to 0.2-5 Hz and not normalised, so that listeners and sites can be compared
• per-listener processed matrices used in the original analysis, which are normalised per listener and therefore carry shape rather than amplitude
• the music as MP3, MIDI and a note matrix
• section boundaries of the music and a per-recording offset locating it inside each raw file
• a data dictionary for every table
• study documents.

The data was captured for a study presented in Chapter 4 of Fadi Al-Ghawanmeh's PhD thesis (2025). Ṭarab and AI: Maqām music generation through machine translation and motion capture, University of Oslo and University of Lorraine.

Formats. tab-separated text (.tsv). Column dictionaries named *_dictionary.tsv define the columns, and the README describes any table that has none.