Best practice for hand detection in unity?

[oliverhamilton]

Hi folks,

I've been learning Unity and how to integrate Leap Motion. I've got a basic demo working, a block that hovers between two hands. My question is what's the best practice way to detect the hands?

I'm currently using

public HandController HC;
HandModel[] HMList;
HMList = HC.GetAllPhysicsHands();
for (int i=0; i<HMList.Length; i++){
    if (HMList [i].GetLeapHand().IsLeft)
        LeftPos = HMList [i].GetPalmPosition();
    } else { 
         .....etc for right...
    }
}

But I see there's another way using...

Frame f = HC.GetFrame ();
HandList HL = f.Hands;
for i ... {
    if (HL [i].IsLeft) {
       Vector PP = HL [i].PalmPosition;
       LeftPos = PP.ToUnityScaled();
    } else {
     ....etc for right...
    }
}

The GetPalmPosition() scale works better than ToUnityScaled() which still needs rescaling. Which method is prefered? Is one better for performance? Reliability?

[Joe_Ward]

In Unity you are (potentially) dealing with two different APIs. Your first example uses the first -- provided by the scripts that are part of the Unity asset package. Your second example uses the second API -- that of the underlying C# wrapper. Our asset scripts are built on the C# wrapper API.

As to which is better, it depends on what you are doing. If you are using the assets, including the hands, you are probably better off sticking to the type of code in your first example. That way the coordinates you get are consistent with the coordinates used by the assets.

If you aren't really using the assets and just using the tracking data, it might be more straightforward to use the tracking data from the underlying C# API directly. But you will also have to do things like map Leap Motion coordinates to your application that the asset scripts do for you.

[oliverhamilton]

That makes sense, thanks very much for clearing that up.