As you know, in most cars it is not easy to turn the intake or exhaust camshaft by hand. You probably need to use both hands, grab the cam sprocket and apply a lot of force to rotate them even with the timing belt removed. This is due to valve springs resisting against compression. However, when you rotate the camshaft a few degrees, the springs should start to decompress which makes it much easier to continue turning the camshaft but is this what happens next? I think the camshaft will be hard to rotate through the full 360 degrees of its rotation because even if some springs are decompressing, the other springs of other valves have started to compress. Correct me if I'm wrong.
Generally, your assertion that as one spring compresses another spring decompresses is correct. There are some caveats to this, though. This mainly has to do with the camshaft setup (ie: OHV v. SOHC v. DOHC) and the number of cylinders. Here is why each makes a difference ...
In a four stroke engine, there are intake and exhaust lobes on the cam which actuate their corresponding valves. The more cam lobes you have on any given cam, the more acculative tension you'll have on those lobes due to how the valvetrain works. If this is a single cam as in an OHV configuration, the tension is fairly equalized on the cam as there will be some lobes compressing valve springs while other springs are decompressing. The more cylinders, the more equalized it will be. Due to this, you most likely will not get the snapping effect you were asking about here.
If this is a V engine in a SOHC configuration (ie: Ford mod motor), the cam lobes are spread across the two camshafts. As long as the two cams are tied together via the timing chain, you should again, not see the snapping you're asking about. If the cams are not tied together, you have a lot higher chance of it happening. I've personally seen the snapping you're talking about on smaller, SOHC engines (ie: inline 4 cyl). There aren't enough cam lobes for there to have equal compression/decompression forces throughout the spin of the camshaft, so it will snap (or pop or spontaneously turn) due to the forces.
On a DOHC configuration, this becomes even greater, because the lobes are now spread out over four cams (in a V configuration) or two cams in a four cylinder. Again, if the cams are tied together via the timing chain (or belt), you'll see a lot less of it, but the fewer the cylinders, the higher the probability.
Something you didn't ask about, but I wanted to point out. There is a basic conservation of energy going on with the camshaft and spring compression/decompression. Yes it takes energy to open the valve as it comes up on the lobe. The energy is then returned back into the system when the valve is closing. There are some energy losses due to friction as the cam rotates in the cam bores as well as the lifters against the cam, but all-in-all, the system gives back to itself for the most part.
First of all you should NOT do this. Rotating the cams while the crank is stationary can damage the valves and/or the pistons.
But assuming the head is removed from the engine and there is clearance below for the valves to move, you should be able to rotate the cams.
That being said, the needed force is not constant but varies depending on what valves/springs are being compressed or decompressed at that time. You will feel that it's "jerky" to some degree.