I saw a video for start citizen that says they use the same rig/animations for both the 1st and 3rd person view (as opposed to floating arms). I haven't found any details on how they did it though. My initial issues would be a skinny character vs a fat character, where in first person you would have to warp the shoulder and crunch them closer to center of body so they fit in the frame.  I'm very curious how star citizen (or any other games) have implemented this. Most people seem to just chop off the 3rd person somewhere after the legs in first person so that you can see your legs, and then have just the arms in first person floating by the camera.

The Arma games have always done this. The first one was almost 20 years ago and was a lot less polished (obviously) so is a good example of a naive approach. Most of the times the arms on screen simply don't look like traditional FPS gun/arms. Their first person "head bob" is also an emergent feature of putting the camera in the 3rd-person-player-model's head! The later ones allow you to disable head-on, which must be modifying the animations so that the head remains stable. 

The traditional way of drawing FPS arms is horribly unrealistic -- as an exercise, strap a GoPro to your face, grab a toy gun, and try to reproduce some footage that looks like an FPS. You have to contort your arms to achieve this task

So one option is to simply produce realistic poses that *don't* look like traditional FPS poses... Show what it would actually look like, even if that means that the gun is off-screen. 

Alternatively, enter the animation hacks. When drawing someone is first person, add hacks that dampen head/neck movements. Disconnect the shoulders and neck from the skeleton and apply arbitrary offsets that result in the hands being in view. Hand craft a completely different set of animations to use for the first person perspective. And so on... 

6 hours ago, dpadam450 said:

I saw a video for start citizen that says they use the same rig/animations

They are talking about animation re-targeting.

So they have a arm model for first person but it's animations is a re-target of the arms from the 3rd person. So two 3D models but 1 rig to make the animations.

Some engines like Unreal and Unity have this build in. You can also re-target using a your 3D software and that is what I think they did.

Worst case scenario: Your animations are bad, there's a too agressive compression, and your engine does some kind of destructive retargeting... you end up with noisy and awful stuff. But you can fix part of it using IKs and interpolation/springs... in a very aggresive way if it's first person. Changing IKs and parents on the run. In Arma you can see how there's an IK when the hand is fixed to the gun, and this arm moves in strange ways matching the torso.  

Look at the way Mirror's Edge handled this. They hand-tailored animations so that they looked good in first-person (and severely twisted in third, you can see some examples on YouTube). You could do something similar - same rig and model, but a separate set of animations tailored for a good looking first person experience.

Cant you just define one animation for both views, really you just change camera position

9 hours ago, WiredCat said:

Cant you just define one animation for both views, really you just change camera position

The problem with this is that you don't get the same quality you would from making two animation sets, one for First person and 3rd person. But you will save time.

You could also make it from 3rd person. Copy the rig and model. Cut away what you don't need and then fix the animations a bit for first person.

The main problem with having a 3rd person body with a first person view is that the body and camera is in the same place, so you see the inside of the head. If you move the camera out, the arms look shorter.

I also tried it with VR games, placing a camera on they eyes. The forehead and nose blocked out part of the view. In real live our brain filters out things like our noses.