jAER sample recordings

These files are short event-camera recordings you can play in jAER without a camera plugged in. Use them to try playback, rendering, and filters.

Get the files

In jAER: Help > Sample data > Download jAER sample data, or Show jAER sample data folder and README if recordings are already present.

Permanent link to the zip on the latest GitHub release: https://github.com/SensorsINI/jaer/releases/latest/download/jaer-sample-data.zip

That downloads about 995 MB. You choose the folder: default is this sampleData directory next to dist/ and lib/ (or the installer sampleData folder). If that location is not writable (typical for C:\Program Files\jAER), the chooser offers jaerSampleData in your home directory. The chosen folder is added to the File > Recent Files and Folders list.

If this folder is already here with recordings, Help > Sample data > Show jAER sample data folder and README opens the folder and the GitHub README.

You can also use File > Open and choose a file here.

How to play one

1. File > Open (or drop a file onto the jAER window).
2. jAER usually selects the matching camera. If the picture looks wrong, set Sensor (AEChip) to the camera in the table below.
3. Play / pause with the player controls. Scrub the timeline; F1 shows you quick help on keyboard shortcuts.

These samples are AEDAT-4 (.aedat4). To replay one as a live OpenCV / DNN / ROS camera, or to load it in Python, see docs/README-DNN-OpenCV-ROS.md.

What is in each file

Previews are 5-second loops of the rendered recording (240 px wide).

Preview	File	Size	Camera to select	What you will see
	DAVIS240C 2016 Tobi juggling.aedat4	6.4 MB	DAVIS240C	Tobi Delbruck juggling. DAVIS 240x180: events plus gray frames. Small file, good first try.
	DAVIS240C CapoCaccia 2016 Hotel dei Pini bar-export.aedat4	12.5 MB	DAVIS240C	Hotel bar at CapoCaccia Neuromorphic Workshop: people moving, APS frames with a brightness histogram. Medium length. Useful for tracking and trying out Flextime playback modes; also advanced frame-event fusion.
	Davis240C DVSFLOW16 Rotating Fan.aedat4	4.9 MB	DAVIS240C	Optical-flow test: rotating fan (DVSFLOW16). Events, APS frames, IMU.
	Davis240C DVSFLOW16 Translating Boxes.aedat4	2.3 MB	DAVIS240C	Optical-flow test: translating boxes (DVSFLOW16). Events, APS frames, IMU.
	Davis346 DAVIS24 2016 Telluride mountain biking.aedat4	110.5 MB	Davis346blue	Telluride 2016 mountain biking; Tobi Delbruck following Alex Zhu through the forest. Try stabilizing and frame event fusion. Enjoy the flow.
	Davis346blue 2026 Steadicam test 6mm lens.aedat4	140.8 MB	Davis346blue	Handheld / test Steadicam IMU derotation by setting 6 mm lens on a DAVIS346 (346x260). Larger file.
	Davis346blue DAVIS24 2019 fast spinning dot ramp up.aedat4	32.2 MB	Davis346blue	Dark spinning dot speeding up (DAVIS24 / 2019). Event rate climbs; useful for tracking and playback and frame-event fusion. Used in S.-C. Liu, et al., "Event-Driven Sensing for Efficient Perception: Vision and Audition Algorithms," IEEE Signal Process. Mag., 2019.
	Davis346redColor MISTLab RoboCup soccer ball approaching from air.aedat4	14.6 MB	Davis346redColor	Color DAVIS346: a soccer ball coming toward the camera (RoboCup / MISTLab).
	DDD20 rec1501953155 San Marino drive clipped.aedat4	70.6 MB	Davis346blue (or Davis346red)	Clip from the DDD20 driving set: road, other cars, IMU. Converted to AEDAT-4.
	DVS128 DVS09 2006 mouse behavior over 3 days.aedat4	86.3 MB	DVS128	Lab mouse over 3.5 days. Sparse events; useful for long recordings and playback with Flextime modes, behavior analysis. Try using the activity histogram on the playback slider. Also used in SC Liu paper above.
	DVS128 DVS09 2006 Patrick Lichtstieiner juggling.aedat4	13.9 MB	DVS128	Early DVS128 juggling clip (Patrick Lichtsteiner). Used in seminal DVS128 paper P. Lichtsteiner, et al., "A 128x128 120 dB 15 us latency asynchronous temporal contrast vision sensor," IEEE JSSC, 2008.
	DVS128 2007 robo goalie balls and arm.aedat4	20.5 MB	DVS128	Early DVS128 RoboGoalie (YT video) data. Balls coming at the goal and servo arm blocking them. See goalie paper. Try multiobject/multizone tracking and velocity prediction.
	DVS640 EssacSim Warehouse Quad walk env0_ep0.aedat4	98.2 MB	DVS640	EssacSim (somewhat optimistic) simulated events from a quadruped walking through a warehouse. See paper.
	NRV DELTA01 2026 humming birds squabbling.aedat4	10.4 MB	DELTA01 / NRV S5KRC1S	NRV 960x720 DVS: hummingbirds feeding and squabbling around their feeder. Try Flextime, slow motion, and RectangularClusterTracker with filterSettings/NRV hummingbirds RCT tracking.xml (Filters window: File > Import preferences…).
	PropheseeIMX636HD Metavision driving_sample street.aedat4	370.8 MB	PropheseeIMX636HD	Prophesee / Metavision street driving sample. Largest file (~371 MB); HD traffic.

More public datasets (not in this zip) are linked under Help > Sample data.

Notes

PyPI aedat 2.2.0 opens these files (events and IMU) but cannot decode DAVIS APS frames. The recordings use the valid DV tag OPENCV_16U_C1 (10-bit ADC in 16-bit samples). That library only maps 8-bit Gray/BGR/BGRA and raises RuntimeError: unknown frame format on the first APS packet, which also stops further events. Color DV frames (8-bit RGB, e.g. the RoboCup file) work. This note will be updated when aedat reads 16-bit gray. Until then, use jAER, File → Save As HDF5/CSV, or dv-processing. Python dataloaders and File → Remote (OpenCV, DNN mmap, ROS2): docs/README-DNN-OpenCV-ROS.md.