Hmd Calibration: Verify if any solution exist from SteamVR

[bastianilso]

In the prism effect issue #97 we discovered that the hmd calibration interferes with the camera rotation. In addition the Hmd calibration also interferes with other modifiers like controller offset. It would be better if we could "teleport the player to 0,0,0" for example via a SteamVR API call or similar, rather than by moving a parent object of the camera rig.

@Xav1204 if you could investigate whether that is possible, or whether others have asked for this before from SteamVR that would be great.

[Xav1204]

@bastianilso The function that I used to define the position and the orientation of the virtual world is the following function :
Valve.VR.OpenVR.System.ResetSeatedZeroPose();

As the definition of the function that I've found on the OpenVr github say, this function define our position as the position 0 in the virtual world.

In order to enable the function, we need to have the following parameter in steamVR settings :

So, when we close the instruction panel, our position and where we are looking will be the current parameters of our position in the virtual world.

[bastianilso]

@Xav1204 Thanks. What implications does "Tracking Universe Seated", beyond allowing us to call this function? Does it have implications for VR users' experience?

Does an equivalent ResetStandingZeroPose() function exist in the OpenVR API? (Just curious)

[Xav1204]

@bastianilso For the tracking space, if we use "Standing" case, the environment assumes 0,0,0 to be on the floor in the center of the Play Area set during Room Setup and in the "Seated" case, the environment assumes 0,0,0 to be the player's head. As in our case, the game is played from a seated position, use Tracking Universe Seated makes more sense.

Moreover, I tried to find the function ResetStandingZeroPose() but it doesn't exist.

[bastianilso]

@Xav1204 assuming the "Tracking Universe Seated" option, what does it do? Can VR players still walk around in the scene?

[Xav1204]

@bastianilso I've tested and yes, we can move in the vr environment.

[Xav1204]

@bastianilso You have said last Friday that we need to test some features with the camerarig rotation and especially something about the fact that we can play the game in a standing position. Is it right ?

if yes, could you please list me all the test that I need to do to be sure that we can use the function "ResetSeatedZeroPose()".

[bastianilso]

hey @Xav1204, yes, here is a list of tests using the new ResetSeatedZeroPose:

• Start Whack-A-Mole with someone in VR, verify that after calling ResetSeatedZeroPose, the therapist UI shows that head rotation and position is 0,0,0 or close to 0.
• Play the TestRetest treatment program. Save the data locally to disk. Open the Sample CSV file and verify that headPosition XYZ and headRotation Y are fairly close to 0,0,0 / 0 degrees throughout the game.
• Rotate real-world chair 90 degrees and restart Whack-A-Mole. Play the TestRetest treatment program again and save data locally. Compare the data in MoleID and MoleWorldPositionXYZ between both Playthroughs, the numbers should be the same (for example. MoleID 0905 should have same XYZ position in game 1 as in game 2).

@Xav1204 could you also please watch, havent seen it, but wanted to verify whether the guy is suggesting to use the same function as you found or what he is doing.
https://www.youtube.com/watch?v=EmjBonbATS0

[Xav1204]

@bastianilso, when I made tests for the first time, I saw that the value didn't equal to the expected values.
So, I tried to compare the values with the current CameraRig rotation that we have in constraint script.
1.

With the current implementation, we have y close to 1 and X/Z have an offset and the rotation are not good because, when we turn our head on the left, we don't go in negative value rotation but in 350 - 360°.

However, with ResetSeatedZeroPose, X/Z are close to 0 and we have similar value for Y. Moreover, we have the same problem with the rotation on the left with Y.
2. As we have nonsense values in the therapist UI, we have also nonsense values in the csv files.
Current :

ResetSeatedZeroPose :

3. For the mole position, we don't have any problem.

[bastianilso]

@Xav1204 Rotation values cannot be negative. Instead they wrap around when they are negative (3, 2, 1, 0, 359, 358, 357). So if the values are around those values, its good.

[Xav1204]

@bastianilso as we change parameters in SteamVR settings, isn't it better to upgrade SteamVr plugin before doing this issue ? Like that, we are sure that the function ResetZeroPose() will work with the new version of SteamVR plugin.

[bastianilso]

@Xav1204 that's a good idea. Let's do that as a first point.

[Xav1204]

@bastianilso I've made a lot of tests with the fonction ResetZeroPose() in different program sessions and with the different modifiers and it looks good.
Can you just test on the dashboard with the datas to be sure that it's all right?
CSV.zip

[bastianilso]

@Xav1204 data looks like what I would expect, so I think it's good.