seems the data im getting is in ‘tracking space’
ok, I think I fixed it by using the original blueprint style that sent, mine was a hybrid with the way it mentions in the camera refactor docs online
testing now
hmmm, cubeHead only has a visibility track now after baking.
Using a Vive, I have a VRPawn set up with a camera component and two motion controllers. Each component has a child mesh for visualisation.
To record:
- Open Sequence Recorder, Add a recording and select the VRPawn as the actor (no need to create another blueprint receiving transforms via Event Tick)
- Click Record
- Open VR Preview and do your thing
(I don’t think opening the VR Preview before recording is necessary any longer, but try that if things don’t work as it should)
After the recording is done, open it and select the camera button on the first track to lock the viewport to what you would have seen through the headset.
Usually, for room scale experiences, the VRPawn should be placed on somewhere on the floor. There is also a base eye height setting on its camera component that you can set to 0.
In 4.12 there was a bug which resulted in incorrect camera transforms being recorded, but this seems to be fixed in 4.13, so no additional steps are required to fix it.
heres my setup in 4.13 and following your method, Im not getting any keyframes, we’d expect to see them on either the camera transform or the ‘Vive head’ mesh transforms or somewhere?
thanks heaps for your help
today Ive tried the ‘tick’ method as well , see attached for slightly altered network. Using print string Im pretty sure its executing and getting the right values, but they are not being baked (im trying to bake onto 2 different meshes here). yesterday they were at least baking the wrong values so not sure what has changed.
ok, I’ve just tried this in a fresh scene using 4.13 default VR setup and all is working super nice (using your ‘just works’ setup). so Im blaming this on a 4.12 scene setup, now going to see if this will merge nicely into my current architectural scene. thanks for all your help guys…
OK! so Ive been messing for a few hours.
if I add a cube parented to the camera in the pawn blueprint in the VR template scene, or if I make a new pawn in this VR template, Im able to see the cube following my head camera, and therefore things seem to bake out.
BUT, if I use an old architectural scene (I didnt create it, its an evermotion scene we bought online) from pre-4.13, and make a new pawn blueprint with the same setup, something in there is stopping the cube from moving, it remains static where the pawn spawns
its not really behaving like it should in the docs
https://docs.unrealengine.com/latest/INT/Platforms/VR/VRCameraRefactor/
Im guessing its something related to my scene, any ideas what in a scene could stop the cube following the camera? I’ve compared each element of the pawn blueprint between the 2 scenes and I cant see whats different
Ok, its not specifically something in the arch-vis scenes doing this
If I create a new scene with VR template it works fine
if I create a new scene ‘blank’ no template the cube wont follow the HMD
so rather its something in the VR template that makes the cube follow the camera
Maybe something in the pawn that ships with the VR template in unreal has something ? Or maybe a world/scene setting hidden somewhere
To frame this question another way, would there be an easy way to import my arch-vis scene into the VR template, I’ve tried migrating all content over, but the ‘levels’ functionality gets in the way, what Id like to do is import my arch-vis level into the VR template level, then just delete the VR template geo etc…
You could try the reverse and migrate the template VRPawn to your project. In the template, find the VRPawn blueprint in the content viewer, right-click->asset actions->migrate and copy all the required bits to your project.
Failing that, a few things to look out for is that the components in your VRPawn is set to Moveable (the mobility setting underneath the transform category in the details panel). Also make sure your pawn’s “Auto Posess Player” setting is set to Player 0
Thanks, I’ll check out those things in a few hours when Im back.
Thanks, I’ll check out those things in a few hours when Im back.
I tried all these, it wasnt looking good until I tried setting the auto posess player thing and it worked. so everything is all good now. Strange how it was the same pawn as from the VR template, must have been something to do with the player controller (I need to learn what that is now…) thanks everyone for your help on this journey. Now I can make lots of videos so very happy!
Can you do the seqeunce recording outside of Unreal Editor? For example, inside the VR game? Have you tried to use startSequenceRecording function inside blueprint?