Mesure the difference of length by VNA, is it necessary to calibrate ?

I have a question concerning the necessity of calibration in my measurement by VNA.
In fact, the measurement time is very important in my setup, so I would like to ask if the calibration correction is necessary ?

Here is my measurement (it is just an example):

It is to measure the length of a cable (50 Ohm), I don't know the length of this cable,

I have a cable (50 Ohm) with a reference length of 1 m. I will measure S21 by VNA in 2 cables. From the phase of S21 I will know the length of these two cables, maybe the lengths are wrong because I did not calibrate the VNA before. But is the difference of length I calculate from the difference of phase S21 give me the good result ?

Let's suppose the the cable in question is with length from 0.9-1.1 m, so the difference of phase gives the difference of length.

Thanks !

I don't see any problem with this kind of differential test. Just use frequency high enough to get required measurement resolution.

calibration becomes more important the junkier your VNA is. If it is not a good 50 ohm source....and your cables have poor return loss....a proper calibration can cure some of that.