Looking for a way to limit a cinemachine brain to specific virtual cameras - without using layers.

As the title mentions, I am looking through Cinemachine's code right now in an attempt to limit a Cinemachine brain to specific Cinemachine virtual cameras, without using layers.

The reason I am looking for such a way, would be best explained in the hypothetical scenario of:

• Making a multiplayer game where each player has their own camera view (CinemachineBrain)
• Making multiple items players may own, that differ per player (such as by choice of equipment)
• Making items be spawned by individual prefabs, not always in the scene
  
• Making each item have its own set of necessary Cinemachine virtual cameras - for instances where the item may trigger an animation that puts a closeup on the player's face, focus on a projectile it shoots into the distance, or the likes.

In my search for the answer, I stumbled upon the Cinemachine documentation for using multiple unity cameras here:
https://docs.unity3d.com/Packages/com.unity.cinemachine@2.2/manual/CinemachineMultipleCameras.html

(Worth to mention I am using Unity 2019.2 and Cinemachine Version 2.3.4 from the unity package manager)

Working with layers would not be ideal in my case though - the fact I'll have to start slapping a ton of layer scans and changes for the cameras to separate them per-player instance aside (because I'll likely need some sort of way to filter through what cameras to use even if its not by layer) it means I have to reserve at minimum X layers out of the 32 layer limitations just for every unique view - and the control would feel very hackish.

I tried looking up CinemachineBrain's code and found it is a sealed class so I cannot exactly inherit or override anything in it, and I can't seem to find a way to get direct control. The user of the camera's layer mask is caked through different methods going even into the Cinemachine's Core class calls as well...

I would go ahead and duplicate the CinemachineBrain class and swap out the layermask parts for something like an overridable IFilterAvailableVirtualCameras interface which defaults to some LayerMaskFilterByCameraLayerMask - but because the cinemachine core class is involved I cannot do that.

Am I missing anything regarding the usage of Cinemachine here? Is there a way I can reduce hackiness and avoid layer usage to limit brain to a few specific cameras, please?

Thanks for reading, please help!

 

Yes, I do want to display multiple camera views at the same time (as per the first dot), split screen for 2 players who may be looking at different directions / in a map that may be too big for both of them to share a single view 'properly'.

 

That's still up in the air for what seems practical from several factors (how complex it would be to do was one) but the original goal was up to 8 (worth mentioning that rather than 8 actual real-life players the screens would be for characters the player manages).

Is it reasonable to edit the classes in Cinemachine's package to try and change up the usage of layers to something else? the files seem to be read-only and I want to make sure I'm not committing some big voodoo.