Proper cinemachine camera placement on full body first-person model

Hello, first post here, first I’d like to say I really appreciate that Cinemachine was made for Unity and the Free Starter Asset pack is amazing to use for beginners like me.

I know this question has probably been asked but I really cannot understand the solutions posted in other parts of the forums, i thought I might as well just ask so I can understand the concept myself and if there’s a solution.

My question is what is the proper way to get my Camera properly snapped onto my players neck or head bone. I really want it to be locked onto either those locations because I want my arms and legs to show properly when the player looks down without seeing its head cut-off. I’m using the method where you separate the head from the model and cull the layer so it’s not interfering.

This is my current layout:

(Note the playerhead is not a rig component it’s just a head model that I have culled in the ArmCamera component)

And these are my current settings for my PlayerFollowCamera:

One thing I noticed in the Starter Asset Pack is that the trick to doing First Person is to use Third Person Follow and basically put a Tracking Target to follow the PlayerCameraRoot.
I tried setting the Position Control to Hard Lock To Target but that completely ruins my Look input for the camera (basically stops working)

I tried putting the PlayerCameraRoot under the mixamorig:Hips so it basically is attached and follows the entire model, especially when it crouches but it causes so much shaking b.c my model has animations.

You can even tell I added a Look At Target which is the playerhead Transform (the Starter Asset Pack did not do this) just to see if it would help with my camera movement but to no avail.

So far my camera movement is not bad, like there’s no shaking but I want it attached directly onto my playermodels neck acting as a head so it doesnt show the neck being cut-off and it feels properly positioned and really like a head.

I’ve been stuck on this for a couple of days and thought I might ask experts here.
I appreciate any feedback or responses and thank you all.

In general, nobody does that. Not even in VR. :wink:

Okay that’s not entirely true but the games where you can look down to see your body, they are extremely rare, and you see very little of your lower body. In VR that was “fun” for a while but everyone adapted to “floating arms” and no body for the most part.

Without VR this sort of thing makes much less sense. Dead Island uses this because you can also attack with your feet, but even when you look down 90 degrees you still won’t or will only barely see the legs or the feet at most. It really seems to be just another “legs+feet” model that’s under the player and seemingly offset behind the player in such a way that you can’t look inside the legs. Arms are probably separate too.

When the player gets knocked out in Dead Island I believe they just switch to a full body model since the camera movement is pretty fixed in that circumstance. It would be instructive to run Dead Island and try to see what the relation between the body parts and the camera is, and wether that would make sense in real life (ie look down on yourself, use your fingers to pull out a “screen rectangle” to look through, and check what you do see of your body and compare it with what games do).

I don’t think games showing the lower body use a full-body mesh all the time, and they’re heavily “cheating” to make this work.

When we do see the lower body clearly AND it looks great, it’s usually a cutscene of sorts, with minimal interaction. I think Call of Duty uses this to great effect. But in such cases you would have limited control over the camera.

That said, consider if it would work best to split the player into just the first person camera and use a lower-body mesh below, then tweak it so that looking down will never feel odd.

Here, this is Dead Island where the player is trying to pry a door open. You can barely make out one foot and knee even at that angle:


The arms here may even be special because it’s one of those “locked camera” situations.

Full video:

While searching for it, I also found this …

I also looked at gameplay footage from Arma 2 and Escape from Tarkov. It was absurdly rare for the player to even look down at an angle more than 70+ degrees, so 99% of the time none of the body was visible.

So for the most part: fake it until you make it. :wink:
And make it so that you rarely even notice your legs and feet. Being able to see the hips, belly and chest follow a decreasing priority and I’d say, at most, you’ll see the hips when looking straight down.

1 Like

For Cinemachine, it helps to understand what the targets mean. The tracking target is what your camera attaches to or follows. In your case, this should be a GameObjet that you position where you want the head to be. In Starter Assets, there is an invisible object representing the player core, to which the camera attaches. Use this, but move it from the center of the character to the head. Then adjust the ThirdPerson rig so that there is no camera offset (or just enough to position the camera at the eyes).

The invisible object’s rotation is controlled by the character controller, so attaching the camera rigidly to it makes sense.

The LookAt target represents that the camera looks at. The player doesn’t look at its own head, so you should remove that for sure. Generally, a first person camera doesn’t look at anything specific, so there should be no LookAt target.

1 Like