The first phase of the IEEE Signal Processing Cup 2019 included a bonus task: participants could send their own recordings of drone egonoise to obtain extra points. In robotics, egonoise refers to the noise self-produced by a robot at its acoustic sensors, e.g., by its actuators.
This task was a success, as 10 teams sent very valuable egonoise data. This allowed us to gather a unique database of drone-embedded recordings, which is made publically available here for research and education purpose, as agreed by participating teams.
Please add the following citation to any research paper referring to these data:
Audio-Based Search and Rescue With a Drone: Highlights From the IEEE Signal Processing Cup 2019 Student Competition Antoine Deleforge, Diego Di Carlo, Martin Strauss, Romain Serizel, Lucio Marcenaro, IEEE Signal Processing Magazine 36 (5), 2019. 
Note (May 7th 2021): For some drone egonoise recording on Planet Mars, check out this amazing footage by the NASA :-).
Team Idea!_ssu
Drone Model: Quadrotor UAV (Phantom 4 GL300C from DJI)
Number of audio channels: 1
Team Maverick
Drone Model: Quadcopter UAV (YH-19HW from https://amzn.to/2T1bSl3)
Number of audio channels: 4
Team Diagonal_Unloading
Drone Model: Quadrotor UAV (Phantom 4 PRO by DJI)
Number of audio channels: 16
Team ChuMS
Drone Model: Custom built UAV (by Dotterel Technologies)
Number of audio channels: 8
Team LEADS_UAV
Drone Model: Quadrotor UAV (Phantom 3 Advanced by DJI)
Number of audio channels: 1
Team NSS_Chellamma
Drone Model: Self-assembled UAV (see documentation)
Number of audio channels: 1
Team KumamoTech
Drone Model: Quadrotor UAV (enRoute Zion PG560)
Number of audio channels: 16
Team AGH
Drone Model: Quadrotor UAV (unknown)
Number of audio channels: 8
KU Leuven Team 1 & Team 2
Drone Model: Quadrotor UAV (MK EASY Quadro V3)
Number of audio channels: 8
Team Shout COOEE!
Drone Model: Intel Aero Ready-To-Fly Drone
Number of audio channels: 8