DJI Spark 360 photosphere feature hits and misses


The ‘toy’ selfie drone has big aspirations

 

Jensen Nature Preserve from 150 ft.

Last month, Chinese remote aircraft manufacturer DJI announced a feature for its mini-drone that I thought was quite interesting: The ability to capture a 360 photosphere. Considering the Spark is an inexpensive entry-level drone that produces great aerial footage in both the video and photo side of the house, this initiative has the recipe for disruption in the consumer 360 camera market. Mid-range 360 cameras sit at about $400 from my experience. The Ricoh Theta S and it’s successor Theta V are good 360 cameras that shoot outstanding 360 photospheric stills and pretty good 360 video. While the Spark will never shoot 360 video, neither Theta will be able to shoot a sphere at 450 ft…unless it’s mounted to a drone.

Yesterday, DJI rolled out the firmware that enabled the new Spark 360 pano feature, and I was pretty stoked to see what the result would be. It was obvious that the process was going to take a while, because the Spark currently has a 180 degree pano mode that shoots two dozen photos, then stitches them together for a pretty decent result if it’s not too windy and preflight calibration is good. The image above is a testament to how well a 180 pano can look with some altitude. However, the process is not reliable.  I have had more failed 180 shots due to stitching fails, interruptions in the shooting process and of course…issues with the environment that impact a good shot. My concern was that the Spark is going to have to spend twice as long taking twice as many shots to capture the ‘entire’ 360 volumetric spectrum – (spoiler: the zenith is not covered but it doesn’t look bad because unlike 360 photos taken on the ground you spend more time looking down than up on aerial 360 panos) – and that means 2X the opportunity and 2X the time to fail at just one point in the workflow. The Spark doesn’t have much room for in-flight inefficiencies due to its short flight time per battery.

360 Telemetry

Lets have a look at the telemetry from my last flight, and It will visually explain the experience:

This is a time lapsed tactical of the flight telemetry. You can see that after launch I ascended to about 100 feet then later 150 feet…executing the 360 and 180 pano modes. On the right you can see the gimbal pitch indicator moving up and down while the Sparks heading (represented by the triangle in the middle) moves clock-wise while shooting the sub component images that will eventually yield a full 360 sphere. There were multiple failures on this mission. But I did get one good 360 pano and 2 great 180s on this 10 minute flight. Unfortunately, the output file is small. Too small for inclusion on Google Street view and Maps without some manipulation…which is against Trusted Photographer quality guidelines.

But I did it anyway.

I manipulated the file, changed some of the EXIF metadata so the Ricoh Theta app would accept it and allow an upload and a share to some social media platforms, then imported the file to the Street View app for iOS on my account then built a little constellation using some real photospheres shot with a real 360 camera.

If you are interested in seeing how it looks on Maps…you can view it here.

Caveats

New feature, new firmware and new experiences on a platform that isn’t even close to maturity – that is a tough spot to be in. DJI is by far the best remote-operated aircraft manufacture in the UAV space, and despite the early problems with the Spark, it is by far the best sub-$500 quad you can get today. That’s saying a lot for this price point as the field is crowded with aircraft that have higher flight time between batteries, more speed, better camera hardware and perhaps a little better at customer service than DJI.

Because of these factors, and the reality that if you want a better aircraft, you will need another $500 to step up to the Mavik Pro – also from DJI. You get 3X the fligh time, more range, and a better gimbal to mount a better camera. Unfortunately you also get a bigger bird. Many don’t realize this yet but a larger aircraft is a liability.  Weight and speed produce higher kinetic energy on impact and a larger frame means encumbrance in an environment poses natural and man-made hazards. Having flown bigger aircraft I understand this quite well. Form factor is key, and has a big impact on capability and no limitations other than what is engineered or perhaps not…engineered…into the functionality of the platform.

Hopefully DJI will get the Spark up to operating in its full potential sooner rather than later, and choose to create an upgrade path to existing platforms rather than spawning new generations that leave the later versions of vehicles with shorter life cycle support. It just seems they have a little more vision than they do execution when it comes to the Spark.

My spark flight history


Also published on Medium.