I've just implemented support for live IPD adjustment. Testing with CV1.
As suggested, I am calling ovr_GetRenderDesc every frame. In all cases from the start, regardless of the the slider's position, the values returned indicate an IPD of 64mm (32mm each way).
When the slider is adjusted at some point, the reported offsets go to the true value.
but doesn't the scale snap back to the right value, as soon as you move the slider?
Yes, correct. But of course the user should not be required to move the slider a bit every time they start a new program. At the moment, all users who do not have 64mm IPD will be seeing the scale wrong every time they start a new program. So Oculus should fix the bug instead of denying it. But it doesn't really sound like the more extreme problem you are describing.
I think I have the same problem with the CV1. In released VR Applications like Oculus Home or Chronos, everything looks unnaturally small to me ( I have an IPD of ~59 mm ) and in my own engine, I can not use the HmdToEyeOffset() functionality of the SDK since it always returns two cameras with a distance of 0.064 units. Is this expected behaviour? Are we developers responsible for giving the user an option to set their IPD? Seems kinda pointless with the headset having the nice IPD adjustment functionality...
you can have a little log to help you in your report :smile: (but it is dk2 ) because this plague everyone since the launch right now.
I have none of this files on my computer, but i believe at home, i have the ipd issue while i have the files, i will provide a log from that other computer tonight with a cv1
VRINF: [CAPI] LibOVR module is located at C:\Program Files (x86)\Oculus\Support\oculus-runtime\LibOVRRT64_1.dll11/07 13:37:55.336 {DEBUG} [Client] Connected to the server running version (prod = 1).1.5.0(build = 241049) feature version = 0. Client runs version (prod = 1).1.5.0(build = 241049) feature version = 0
VRDBG: Connected to the server running version (prod = 1).1.5.0(build = 241049) feature version = 0. Client runs version (prod = 1).1.5.0(build = 241049) feature version = 011/07 13:37:55.337 {DEBUG} [Kernel:Default] Failed to open file: C:\Users\nsilvagni\AppData\Local\Oculus/ProfileDB.json
VRDBG: Failed to open file: C:\Users\nsilvagni\AppData\Local\Oculus/ProfileDB.json11/07 13:37:55.337 {DEBUG} [Kernel:Default] Failed to open file: C:\Users\nsilvagni\AppData\Local\Oculus/Profiles.json
VRDBG: Failed to open file: C:\Users\nsilvagni\AppData\Local\Oculus/Profiles.json11/07 13:37:55.337 {DEBUG} [Kernel:Default] Failed to open file: C:\Users\nsilvagni\AppData\Local\Oculus/ProfileDB.json
VRDBG: Failed to open file: C:\Users\nsilvagni\AppData\Local\Oculus/ProfileDB.json11/07 13:37:55.337 {DEBUG} [Kernel:Default] Failed to open file: C:\Users\nsilvagni\AppData\Local\Oculus/Profiles.json
11/07 13:37:55.337 {DEBUG} [Kernel:Default] [HMDState] Using default profile default
VRDBG: Failed to open file: C:\Users\nsilvagni\AppData\Local\Oculus/Profiles.json11/07 13:37:55.337 {INFO} [Kernel:Default] IAD changed to 64.0mm
VRDBG: [HMDState] Using default profile defaultVRINF: IAD changed to 64.0mm11/07 13:37:55.337 {DEBUG} [Kernel:Default] [SharedMemory] Creating factory
Update a couple days ago for anyone following this issue.
Actually, you may be right. I just wrote a quick C++ console app, and the value of HmdToEyeOffset is always +-0.032 regardless of the slider. Let me have an engineer look deeper into this. - @cybereality
Thanks for the info. It is quite hilarious how hard it is to get Oculus developers to look into a bug. I love cyber's comment "Actually, you may be right." as if he had never heard of this before. And @tomf denying the bug four months ago, then ignoring us confirming that the bug really does exist. Exasperating!
I'm really sorry for the delay. I don't think I initially understood how serious this issue was and didn't do a proper investigation. I can confirm that it is happening for me too. I have a bug report filed already, and will be sure to stay on top of the engineering team until this is resolved. Thanks.
