Add a second detection mode to calibrated sculk sensors

Problem to solve: One of the challenges with the sculk shrieker is that there are several times as many detectable sounds as there are possible output strengths. This means that any circuits built using one have the potential for false positives. Although this is currently manageable in many use cases, some simply are not (e.g. an Alice in Wonderland-style "eat me" / "drink me" setup where each of the two actions produces different behaviors in the circuit). Additionally, such limitations are likely to become more common as future updates add more detectable sounds, thereby degrading the signal-to-noise ratio.

Proposed solution: give the calibrated sculk sensor a secondary mode that detects the position of a sound within its category, rather than filter on the category itself. In effect, this turns a one-dimensional measurement (which of the existing sound categories contains the triggering sound) into a two-dimensional measurement (the existing functionality is one axis and the proposed additional mode is the other). Using the previous Alice in Wonderland example, the primary mode would be fed a signal strength of 8, and two sensors in secondary mode would be fed signal strengths of 1 and 2 for drinking and eating, respectively.

In the case of amethyst resonance, I would suggest that the behavior match existing behavior (pass along the same signal strength it takes in).