You opened the door for anyone, so I will add my two cents.
I have given this much thought as I will be trying to counter balance the knee of a machine similar to Chich's. Please, lets not allow this thread to digress into a "knee vs. quill" thread as it is not the appropriate forum.
First is the thought that air does not add mass to what needs to be moved so acceleration issues -- which are going to be there already simply based on the mass of the knee or the head -- are not exasperated.
The second is that with a suitable air regulator the system can be easily adjusted to null out the weight of whatever is put on the table. True, the same could be said for a granular selection of counterweights but the thought that I could simply turn a knob to set the appropriate pressure seems easier than going behind the machine to add weight. I am assuming this is not an issue for an RF-45 clone as the weight of the head never changes other than minor differences based on tooling.
The down sides of air appear to be the cost and complexity of an air system and the constant need for air. I dismiss the former out of a naïve sense that once it is set up and running it should be fine until I pass on and I dismiss the second because I already have a robust air system in the shop so it is not a big deal. However in your case the added complexity only mitigates the mass issue as you do not have a 'changing weight' issue.
In summary I think you are going in the correct direction to use a counterweight system and then try to resolve the acceleration issue within the software controling the system.
The only modifications I will be exploring from the implementation presented by Dave would be the use of bicycle chain and sprockets rather than cable and pulleys.
Arvid