[time-nuts] Is the practical quality of a 10 Mhz reference determined by the quality of the fundamental or by the quality of the zero crossings?

[Erik Kaashoek]

For one of my projects I was requested to make a presentation about 
measuring phase and frequency
Part of the presentation is about measuring phase and frequency for 
which I could use a lot of excellent material from various sources.
I did run into one small problem when trying to explain why the 
PhaseStation phase measurement method (decimated I/Q down mix to zero 
Hz) is ok compared to previously zero crossing methods such as used in 
interpolating reciprocal counter.
When using 10 MHz reference in a modern measurement device, is the lock 
on the reference done by direct conversion to a square wave (some simple 
digital circuit like a limiting amplifier) or are more advanced clock 
recovery approaches used that look only at the fundamental and use all 
information in the 10 MHz fundamental, just like the Phase Station is doing?
In what category would a PLL for clock recovery fall? Is the PLL looking 
to the fundamental and ignoring noise on the zero crossings by using all 
available information or is it plagued by the same problems as a zero 
crossing clock recovery?
I hope someone with knowledge on clock recovery could help out here. 
Many thanks in advance.
Erik.