Forum Discussion
Motion sickness mostly solved in Hydra walking environment
Yesterday I had my first walk around in my 3D environment without motion sickness: a wonderful experience! In this post I want to share the way I have done this.
My setup is as follows:
- Oculus Rift Developer kit
- Razer Hydra
- Unity3D
- realistic 3D model of a historic town.
The Oculus Rift tracker is driving the head movements, the body and arms movements are driven by the Hydra.
Looking around in the model was not unpleasant, but walking around got me (slightly) sick. I was using the left analog stick for moving around: up/down for forward/backward motion, left/right for turning left/right. Walking speed is matched to normal walking speed, rotation speed is set slow to about 15 degrees/second.
Then the following idea came to me: in most cases you hands (or actually, the Hydra hand devices) are pointing in the direction your body is facing... Would it work when the orientation of the body matched the average orientation of the hydra's?
So I took the Y rotation of both Hydra 'hands', averaged them and assigned this orientation to the body (in short). Now you hands are driving the rotation of the body, so no need for the left/right axis any more on the analog stick.
My first tests worked out wonderful: the body is almost every time facing the direction I expect. Better even: slight deviations do not seem to be a problem: body orientation does not prove to be as critical as head orientation. I can walk around in the direction my physical body is facing. To turn around I have to turn on the spot.
The forward motion (still with the analog stick) still seemed a bit strange but then I started to 'walk on the spot': just move the legs up and down in a walking motion and everything fit together wonderfully! (besides my wife entering the room asking about the sound my feet made on the wooden floor...)
Then the downsides:
* The Razer Hydra is wired, so I must be weary not to turn more than 360 degrees. This will be solve in time when the STEM system will be released.
* The Rift is wired. Guess we need a wireless Rift too then, in order not to strangle myself...
* Physical drift: the walking on the spot and turning on the spot did not prove to be completely 'on the spot'. When I took off my Rift I saw I drifted away about half a meter. I am planning to see if I can solve this by placing a disk with a rim on the floor. When my feet feel the rim I will realize that I am drifting and will be able to correct my position.
Of course the Virtuix Omni may give a better experience, but as I don't have it, this seems to be the next best thing.
Pascal.
My setup is as follows:
- Oculus Rift Developer kit
- Razer Hydra
- Unity3D
- realistic 3D model of a historic town.
The Oculus Rift tracker is driving the head movements, the body and arms movements are driven by the Hydra.
Looking around in the model was not unpleasant, but walking around got me (slightly) sick. I was using the left analog stick for moving around: up/down for forward/backward motion, left/right for turning left/right. Walking speed is matched to normal walking speed, rotation speed is set slow to about 15 degrees/second.
Then the following idea came to me: in most cases you hands (or actually, the Hydra hand devices) are pointing in the direction your body is facing... Would it work when the orientation of the body matched the average orientation of the hydra's?
So I took the Y rotation of both Hydra 'hands', averaged them and assigned this orientation to the body (in short). Now you hands are driving the rotation of the body, so no need for the left/right axis any more on the analog stick.
My first tests worked out wonderful: the body is almost every time facing the direction I expect. Better even: slight deviations do not seem to be a problem: body orientation does not prove to be as critical as head orientation. I can walk around in the direction my physical body is facing. To turn around I have to turn on the spot.
The forward motion (still with the analog stick) still seemed a bit strange but then I started to 'walk on the spot': just move the legs up and down in a walking motion and everything fit together wonderfully! (besides my wife entering the room asking about the sound my feet made on the wooden floor...)
Then the downsides:
* The Razer Hydra is wired, so I must be weary not to turn more than 360 degrees. This will be solve in time when the STEM system will be released.
* The Rift is wired. Guess we need a wireless Rift too then, in order not to strangle myself...
* Physical drift: the walking on the spot and turning on the spot did not prove to be completely 'on the spot'. When I took off my Rift I saw I drifted away about half a meter. I am planning to see if I can solve this by placing a disk with a rim on the floor. When my feet feel the rim I will realize that I am drifting and will be able to correct my position.
Of course the Virtuix Omni may give a better experience, but as I don't have it, this seems to be the next best thing.
Pascal.
5 Replies
- I'm going to take a guess at this.
We have "sensors" in our whole body that allows to detect changes in momentum. When you move your arms in any direction, the momentum of your arms changes until they stop moving and the body can detect that. Perhaps if the on screen acceleration can be mapped to the acceleration of the hydra controllers, then the body might "detect" enough change of momentum, to highly limit motion sickness. Maybe once the acceleration of the hydra controllers stops, the velocity of the onscreen world would then be comfortable for the user if the frame rate is high enough. - I think the solution is more straightforward.
In the old situation the physical body was always oriented the same, but the virtual body rotated by means of the analog stick. As the virtual head is attached to the virtual body, the virtual head (and hence, the image in the rift) rotates when the virtual body rotates. This causes motion sickness: the physical body stands still (including head) and the virtual body an head rotated.
In the new situation, the physical body and head rotate. An approximation of the body rotation is copied to the virtual body, the physical head rotation is copied to the virtual head rotation by the regular Rift tracking function. Thus: the physical body and head rotate and the virtual body and head rotate with it: no mismatch and no motion sickness!
The essential difference:that the head rotation is now only driven by the Rift orientation, and no longer by the analog stick.
Pascal. - Ahh I see.
- For me, it's the speed of the left/right (turning) motion that determines motion sickness. If I move left/right either very quickly or very slowly (via mouse), I get little to no motion sickness. If I move using the keyboard and it's in-between speed, I get motion sick as my eyes try to track the environment. If it's a very fast movement, my eyes don't have a chance to track the movement and don't get sick, and if it's very slow, they have no issue tracking. Kind of like if you're in a car or train looking out the side window and watching things close to you fly by vs. things in the distance.
You're hydra setup may just give you more natural turning left/right to where your eyes can track it without getting you sick. - I mostly agree, that is why I set the rotation speed to a slow speed. But I think it is the acceleration/deceleration which causes the most problems. Letting the stick jump back to the resting position (what I often tended to do...) gives a rapid deceleration causing the sickness.
On the other hand: I can now turn very quickly, but now it is consistent with all my senses.
Pascal.
