My hunch is that it's not the speakers that are the problem, it's the way the service providers
manage bandwidth during the calls.
With a landline call, you have dedicated bandwidth for the duration of the call, even if their are
gaps in the conversation.
With a cellular call, bandwidth is dynamically allocated, even during the call.
My impression is that the service providers are eager to relinquish the allocated
bandwidth as soon as there's a gap in the conversation. When one of the parties
begins to speak after a gap, there's a small lag in re-acquiring bandwidth.