[hdersch]

I wonder how this compares to purely vision-based systems which use nothing but the images themselves for stabilization. Here are some quite old results of stabilization using image-based 3d-reconstruction of the scene which I wrote more than 10years ago, compared with other stabilization programs of that time (Deshaker, Adobe After Effects, Youtube). With todays improved hardware and progress in 3d-algorithms you may not need any additional gyroscopic data.

https://www.youtube.com/watch?v=-m3fwhx3Z5g

[_flux]

I think gyroscopic data still can have the edge if it has higher sampling rate than the video: then it could be used for removing blur from individual frames.

I also expect purely gyroscopic approach to be much lighter compute-wise.

[7e]

I expect you can recover most if not all of the gyroscopic data even from a blurred image, possibly at higher accuracy.

[_flux]

I actually don't think that's possible, if you assume the objects in the video can also shake and the intent is to eliminate only the camera shaking.

[mikae1]

In my experience, what you call vision-based systems don't even come close to Gyroflow. After I started using Gyroflow I haven't touched the Ronin gimbal many times.

[7e]

A physical stabilizer like the Ronin is not what they mean. They mean vision based approaches which should be more accurate.

[Gigachad]

It doesn’t seem obvious that vision based stabilisation would be more accurate than a high accuracy gyroscope of the actual motion.

At any rate if gyroflow obsoletes an actual gimbal it’s pretty close to perfect.

[UltraSane]

gyroscope data is much higher time resolution and is the ground truth vs guessing motion from pixel.