This is probably the relationship that most heli-flyers find the most difficult to wrap their brain around. It really is not that difficult if you think of pitch and rotor speed as the load, and throttle opening as the power tap.
First let us clarify the "load", which will define the amount of power required to accommodate it. The three items that define this load are:
1) the amount of pitch being used - both collective and cyclic
2) the rotor speed at which one wishes to turn the main rotor
3) the drag ratio of the rotor blade at that pitch setting and at that head speed.
As we cannot change rotor blades in flight, we will take this item out of this discussion also. The power absorbed by the tail rotor is also a factor, but for this discussion we will assume that it is a constant so will also be left out.
More pitch or rotor speed will require more power, consequently less pitch or less rotor speed will require less power. So you can see that one can increase the pitch and decrease the rotor speed a certain amount to require the same power as that which would be required for lesser pitch and a higher rotor speed. This is why a heli can be set up to hover at a higher pitch and a lower rotor speed or at a lower pitch and a higher rotor speed. The load in the hover is the same (the weight of the helicopter), it is just how we set up the heli to absorb the load that is different.
So let us use a good sport/training head speed of 1600 that will be comfortable and less intimidating for training, sport and entry level aerobatic flying with most sizes of helicopters. Why did I choose this rotor speed? It will keep most engines somewhere in their power band. It is low enough to keep the noise and intimidation level relatively low. It is low enough to limit damage in case of a boom strike or other incident. It is high enough to work with the damping in most helis so that there will be no "nodding". It is high enough to give a good amount of tail rotor power. It is low enough to keep the cyclic from being too aggressive (the higher the rotor speed, the faster will be the cyclic response) yet high enough to give acceptably fast cyclic for most aerobatics.
Let us also assume that at this rotor speed, and with the symmetrical blades installed, we take it out to the field and find that it will hover at 5.5 degrees of pitch, and the throttle barrel will be 50% opened. The easiest load setting to find next would be the one for maximum power available, as the throttle setting is predetermined by having the throttle barrel on the carburator to be wide open at 100% throttle on the TX. I assume your radio throttle function is set up symmetrically so that it has the same ATV setting at the top as at the bottom, and that at 100% on the throttle curve setting, the carb is fully opened . At the bottom of the stick movement, with the trim in the middle, the engine just idles.
To determine this load, one just has to set the top end pitch so that the engine maintains the same 1600 rpm (does not drop off or speed up) at the full up stick setting. Let us assume that this is 10 degrees. So we now know that the maximum pitch that this power system (engine, exhaust system, fuel, and glow plug) can accommodate while maintaining a 1600 rotor speed is 10 degrees. Let us not forget that this is all while working against gravity in the upright position. Needless to say, if I flipped the heli inverted, pushed the throttle/collective stick forward to the full positive ten degrees, the blades would now be less loaded, as they would be now pulling down with gravity instead of pulling upwards against gravity. Yet my throttle would still be at the fully opened position, developing full power. Full power and a much smaller load would result in a very noticeable increase in RPM as the heli is driven towards terra firma. (And there would be a noticeable increase in sphincter muscle tension in the pilot and spectators). Therefore, different settings would be required for the part of the curves that would be used for any inverted flying.
Moving on to the other, more complex throttle settings for loads while upright, we must now look at the loads at the bottom end of the stick. These may be loads that occur with the heli is set up for hover training, for normal descents while flying right side up and on to loads resulting from the need to climb at full power while inverted.
From the explanations above, one can see that when right side up and descending, the blades will be unloading and take less of a throttle opening to stay at 1600 RPM. As an example, let us assume one is in Idle Up-1 and using a low stick pitch setting of -3 degrees. As we know, it took +5.5 degrees to hover, so a -3 setting is 8.5 degrees less, so it will certainly cause the helicopter to descend. It will also represent a much smaller load on the engine. So here we would set the throttle at a point that will keep the rotor speed at 1600 through the descent. If through the descent the head speed increases, then lower the low stick throttle setting. If it decreases, then bump up the lower end of the throttle setting. Set this way, rather then set so the engine comes to idle, will mean that your engine will not have to use any of its' power to come up to speed when you try to stop the descent, as it will already be at speed. All that will be required is a larger throttle opening to allow more fuel into the engine in order to deliver the power required to stop and hover…and this was already set for the hover. Set this way, one will find it very easy to stop a descent instantly with no mushiness and no drop in rotor speed..
Similarly, if one wants to fly inverted, then one must adjust the lower half of the pitch curve to reflect the pitch and throttle values that will hover inverted and climb inverted at full power. If one is using symmetrical blades, he will find these settings to be very similar to those set for right side up flying. All he will have to do is set the mid stick pitch to zero degrees and adjust the throttle at that point to maintain the 1600 rpm rotor speed.
Once the throttle curves are set to also absorb the changing collective loads, without changing rotor speeds, the heli will fly a lot crisper and with a lot more collective precision. Because the rotor speed stays constant, the torque loading on the tail rotor will be a lot more constant, meaning that the workload on the gyro and tail rotor servo will be a lot less. Consequently, getting this right is the first step in setting up a sweet flying helicopter.
Of course, just as such a set up will make collective inputs a lot crisper, setting up the throttle to supply adequate power for the extra loads from changing cyclic commands and tail rotor commands, without changes in rotor speed, will also render these commands a lot crisper and more precise. But that is another discussion. (see the "Faster Cyclic" article here)
Of course a Futaba GV-1 governor (or any other) can also be used to look after all of this for you. But learning to set up a good collective to throttle curve relationship can save a lot of money if one has more then one helicopter. Buying a GV-1 for every helicopter in the stable can put a lot of stress on ones finances.