Having these be separate implementations sounded awesome at the time,
but it only ever led to confusion. There's no practical difference if,
for example, catch sees hitobjects with `IHasPosition` instead of
`IHasXPosition`.
The equality of samples is generally used to compare the sample
equality, not its full properties. For instance, we don't compare
`Volume` in the base implementation.
Having `IsLayered` here breaks actual usages of equality, ie. for
pooling purposes.
This is quite a breaking change, but I think it is beneficial due to the large amount of usage of this class.
I originally intended just to remove the allocations of the two delegates handling the `Changed` flow internally, but as nothing was really using the bindables for anything more than a general "point has changed" case, this felt like a better direction.