Why do we modulate signals in optics experiments? And how is modulating polarization of an optical beam better than modulating the intensity? I have been working on simple MOKE setup where magnetization of a sample is reflected in the extent to which it modulates polarised light. The intensity and the changes in it (due to rotated polarisation) are measured using lock-in detection to be sensitive to very small changes.
Am I right in understanding that - since a lock-in can 'see' only periodic signal to compare it to a reference frequency, we introduce periodicity in the signal using modulation. 
Is this what 'modulation' means? 
Is the sole reason for modulation to provide a periodically varying signal?
I used two schemes - intensity modulation using a chopper and polarisation modulation using a photo-elastic modulator. I read that polarization modulation is more sensitive than intensity modulation. But never found a reasoning for why. Can anyone please explain why this is the case?
Thanks a lot!
 A: In a sense, all comparison experiments --- where you first do A, then do B, and see whether the results of A and B are different --- are "modulation" experiments.  Once you see this, it's usually a short jump to having the experiment's sensitivity improve if you make many measurements of A and B at high frequency compared to a few measurements at low frequency.
For example, suppose you are measuring a reflectivity from some surface that has a chemical reactivity you're not interested in.  You discover that after you've prepared the surface, it tarnishes measurably over the time it takes you to do your interesting measurement.  You discover this by measuring A, then measuring A again, and discovering that your "control" experiment isn't repeatable.  So you address that problem, perhaps by changing the atmosphere that's causing the tarnishing.  You also change your experimental procedure: you measure A, then you measure B, then you measure A again to make sure that your tarnishing problem is gone.  Or maybe you can't totally get rid of the tarnishing problem, but with a series of measurements ABABABABA you can state with some confidence that it's affecting your A measurements and your B measurements in the same way and subtract out.
That's what your lock-in amplifier is doing.  There are a million reasons your signal could be varying --- maybe your surface isn't tarnishing, but your laser diode is aging, or the laser power is modulated by the alternating-current frequency on the power line, or the like.
You can get rid of a lot of those problems by intentionally modulating your input: if the effect doesn't change in sync with your modulation (or its harmonics) then you plain old don't care about it.
As far as modulating intensity versus modulating polarization: my instinct is that's easier to introduce spatial non-uniformities by doing intensity modulation than by doing polarization modulation.  But someone who does more optics than I do may have more to contribute there.
