Assume a pure tone (single frequency) is listen, lets say 2 kHz. If I understand correctly the temporal theory (aka timing theory), in a cochlea neuron the action potentials create a signal that will have one spike for each single tone cycle with the exception of some gaps due to the minimal interval between neuron triggers.
Volley theory extends previous model to a set of neurons. The union of triggers in all neurons of the set will have at least one spike per cycle (with few exceptions). In other words, the base frequency of the signal composed as the union of triggers of all neurons in the set will be the same frequency than the one of the incoming sound.
My question is, if in this model the trigger intervals is used to encode the sound frequency, and taken into account that usual models for neuron activity only take into account intervals between action potentials (all-or-none), how sound amplitude (power) of the sound is translated to the nerve signals. That is, the differences in the nerve activity caused by two tones of same frequency but different amplitudes/powers, assuming frequency/volley model.