Welcome DJI Spark Pilot!
Jump in and join our free Spark community today!
Sign up

Litchi orbit mode problem

Canalsman

Member
Join
Jul 27, 2019
Messages
21
Age
72
Today I have been trying orbit mode.

It works well apart from inability to keep the target centred in the image. When the orbit starts the target is on the left side of the frame and as the orbit progresses it moves towards the centre of the frame. As the orbit reaches the end the target moves to the left once more.

I have tried setting up an orbit using waypoints with the waypoints focusing on the POI and the same behaviour occurs.

Has anyone else experienced this behaviour?

A Google search has not provided any clues ...
 
Today I have been trying orbit mode.

It works well apart from inability to keep the target centred in the image. When the orbit starts the target is on the left side of the frame and as the orbit progresses it moves towards the centre of the frame. As the orbit reaches the end the target moves to the left once more.

I have tried setting up an orbit using waypoints with the waypoints focusing on the POI and the same behaviour occurs.

Has anyone else experienced this behaviour?

A Google search has not provided any clues ...
I've had the same problem in the past. I think it has something to do with only being a 2 axis gimbal. Maybe try slowing it down.
 
I have been in touch with Litchi support.

Several interesting points have been made about Litchi functionality with the Spark.

Orbit mode is not supported officially by Spark firmware:

When I mean not officially supported, I mean that DJI advertises that Spark is not compatible with Orbit.

In our tests, it did work, however the accuracy is not very good so perhaps DJI did not tune it for the Spark.


And about my pseudo orbit mission using waypoints that displayed the same problem:

For waypoint mode, if you want improved poi focusing, you need to make sure there is only one poi in the mission.

The DJI waypoint engine has improved accuracy for poi but only when there is a single poi in the mission.

With that being said, similar to orbit mode, DJI does not officially say that Spark supports waypoint mode. So it is possible you will still have accuracy problems even with one poi.


And about using a lower speed as suggested above:

I believe that if the speed is slower, it will mean the problem is less accentuated and that may fit your needs, but if there is a problem then it will also be there at slow speed most likely just less visible

Regarding the first couple of points it makes it appear misleading to state that Litchi offers support for orbit and waypoint given it is not implemented using DJI approved functionality.

I have asked Litchi to comment ...
 
I guess the DJI SDK can only take 3rd party developers so far. Would be nice if DJI were develop orbit and waypoint functions for the Spark, but at this late stage it is likely they would hold such features in reserve for the next generation of drones for marketing purposes. Perhaps another developer can come up with a workaround in time. Far less capable and less expensive drones can orbit and follow waypoints.
 
The response from Litchi:

I don't believe that we actively advertise support of waypoint/orbit for Spark specifically, and it does work, just does not have the same level of polish as for other drone models. The Spark itself is relatively limited in terms of its capacities too which does not help.
 
The Spark's limitation while using POI or Orbit can also be attributed to the EIS (electronic image stabilization) it uses due to the absence of the 3rd axis (yaw axis) in its gimbal. The EIS system crops from a larger field of view and adjusts the position of the crop to keep the view/footage stable. With such a system, it becomes difficult to keep the object of your choice centered all the time, when the camera itself (drone in this case) is moving continuously. I think this is the only reason why DJI never gave these features in their official app as they knew users would never be happy with the output.

eis-explained.jpg
 
I agree that the two axis gimbal might cause an object centred in the image to be displaced from centre by small amounts.

However this is not the problem I described.

The focal point of the orbit was completely incorrect with the chosen focal point being placed at the left side of the frame then gradually moving to the centre of frame as the orbit progressed. As the orbit reached completion the focal point moved back to left of frame.

The issue is that the Spark does not face the focal point correctly during the orbit. It appears from what Litchi have said that the unsupported DJI firmware function is the cause of the behaviour.

I'm disappointed that there is no mention of orbit and waypoint deficiencies on their website.
 
Good topic, I have seen some comparison videos with greater (dji) drones and results are quite varied... doing waypoints POI and orbits with Litchi vs Go4 ... Have a look at this video...

My only complaint about this video is that they insist on speeding up the footage, which accentuates the glitches, hard to guage the reality sometimes. The other variable is the video processing - was it done correctly?
 
My only complaint about this video is that they insist on speeding up the footage, which accentuates the glitches, hard to guage the reality sometimes. The other variable is the video processing - was it done correctly?
I would've been nice if they video had shown real time progress vs sped up, but even with it, the Spark wowed in comparison with drones costing magnitudes more. Even the drones with 3 axis gimbals didn't put it to shame.
 
Good topic, I have seen some comparison videos with greater (dji) drones and results are quite varied... doing waypoints POI and orbits with Litchi vs Go4 ... Have a look at this video...

My only complaint about this video is that they insist on speeding up the footage, which accentuates the glitches, hard to guage the reality sometimes. The other variable is the video processing - was it done correctly?

Thanks for the link. The video was interesting but as you say lessened in value by speeding it up.

The Spark footage displayed the problem I have described, especially at the greater distance, with the focal point drifting off to the left.

This is useful because it points the finger at the implementation rather than my Spark.

Litchi have now updated their help page to state that waypoint and orbit implementation for the Spark is 'experimental'.

I was offered a refund but I'm happy to keep the app wishing to experiment further to see if I can work round the problem.

I have a theory I wish to try.

If I define an orbit using waypoints it is possible to adjust the heading at the individual waypoints after defining them as focused on the POI. This I think will allow me to fine tune a mission after an initial flight to get the desired result.

I'll update this thread once I've tried this solution.
 
Right ...

I have tested my theory and it works. Adjustment to the heading for waypoints that incorrectly align the Spark with the POI allows you to achieve a pretty good orbit.

I set up a series of waypoints, 22 in all, in a circle round the POI. To ensure that it was close to a true circle I cut a piece of paper to create a circle of the correct radius to fit my tablet screen. In the centre I cut a small circular hole to place over the POI.

I then taped the template over the screen with the POI in the centre and using a Bluetooth mouse, to improve accuracy, created the waypoints round the outside of the template. I found that if you try to create a waypoint too close to the adjacent waypoint you get the property box displayed, so you have to be careful.

Zooming the base map as required allows you to choose the radius of the orbit.

With a steady hand you could use the drawing tool I imagine.

A rough and ready way to create the desired orbit but it works.

I flew the mission a number of times, the last time in completely still conditions. Interestingly the panning round the POI is notably smoother when there's no wind. This confirms the limitations of the two axis gimbal, but also suggests that panning in general is likely to be significantly better in light wind conditions.

It's a bit laborious but may be worth doing for that 'must have' footage.
 
Right ...

I have tested my theory and it works. Adjustment to the heading for waypoints that incorrectly align the Spark with the POI allows you to achieve a pretty good orbit.

I set up a series of waypoints, 22 in all, in a circle round the POI. To ensure that it was close to a true circle I cut a piece of paper to create a circle of the correct radius to fit my tablet screen. In the centre I cut a small circular hole to place over the POI.

I then taped the template over the screen with the POI in the centre and using a Bluetooth mouse, to improve accuracy, created the waypoints round the outside of the template. I found that if you try to create a waypoint too close to the adjacent waypoint you get the property box displayed, so you have to be careful.

Zooming the base map as required allows you to choose the radius of the orbit.

With a steady hand you could use the drawing tool I imagine.

A rough and ready way to create the desired orbit but it works.

I flew the mission a number of times, the last time in completely still conditions. Interestingly the panning round the POI is notably smoother when there's no wind. This confirms the limitations of the two axis gimbal, but also suggests that panning in general is likely to be significantly better in light wind conditions.

It's a bit laborious but may be worth doing for that 'must have' footage.

Thank you for this great explanation. Were you in photo mode or video mode when running the mission? Would like to see the video if you recorded any. And just for your own understanding, you can try running the same mission with the live view in photo mode next time. Then you can clearly see how shaky the live view is without any stabilization, especially when doing pans, orbits and also in windy conditions.
 
Last edited:
I'm glad it makes sense.

I was using video mode and recorded the mission. I hadn't considered using photo mode as a comparator, interesting idea.

I don't at present have anywhere to upload the footage. Do you have any suggestions about upload areas? I imagine YouTube is one possibility but it's an area I've never investigated.
 
Do you have any suggestions about upload areas? I imagine YouTube is one possibility but it's an area I've never investigated.

Yes, YouTube is the most widely used video sharing platform and the best part is that it's free. You can simply login with your Google account and start uploading.
 
Last edited:
  • Like
Reactions: Canalsman
Thanks...

I'll look into it tomorrow. I don't have access to the internet other than through my smart phone, but I have unlimited data allowance, so it's a case of being in a location with 4g connection...
 
I decided to try Orbit mode again today.

The previous trials all used anti-clockwise rotation and were poor.

So I tried clockwise followed by anti-clockwise at the same location. It was the only parameter I changed.

Surprisingly clockwise works much better.

I plan to run a series of tests starting the orbit at North, South, East and West, both clockwise and anti-clockwise and see how it performs ...

I really didn't expect this difference.

I will also try a waypoint orbit. This too I have only tried anti-clockwise so I'll see if clockwise works better.
 
  • Like
Reactions: nilanjan118
Tried orbit mode today.

I used the same orbit entering at the four points of the compass, with clockwise and anti-clockwise rotation.

Anti-clockwise is poor, clockwise fairly good.

It makes no difference whereabouts you start the orbit. The POI drifts to the left slightly when rotating clockwise as you approach the southernmost point, therefore facing northwards, and drifts slightly to the right as you approach the northernmost point, therefore facing southwards.

This behaviour is much more exaggerated when flying anti-clockwise. I can't think why.

I think that by moving the POI slightly to the west of the intended point and flying clockwise you should get an acceptable result irrespective of whereabouts you start the orbit.

I'll try this next time and report back.

Annoyingly the remote controller had dropped out of FCC mode which I didn't realise which caused lots of problems with these tests with lost connection and gimbal control problems.

Lesson learned... Check before flying!
 

Members online

No members online now.

Forum statistics

Threads
14,593
Messages
118,799
Members
17,987
Latest member
csdisme