If the airplane is demonstrating coupling in knife-edge flight, then the same coupling will be present with rudder deflections in horizontal flight. Try a flat rudder turn without mixing. Does the airplane pitch and try and roll out. Turn on your knife-edge mix and fly at about the same throttle setting as you do in knife-edge. Most likely the airplane will behave better in the rudder turn. Rudder coupling has the same effect regarding coupling regardless of what the airplane's attitude is, level, climbing, inverted, or knife-edge. When trimming an aerobatic aircraft, the object is to remove that coupling in all flight regimes.

To answer your question directly, when you have mixed out the coupling in your knife-edge, you have mixed out rudder coupling in all flight regimes at the same rudder/throttle/airspeed settings which is a very good thing. As the airplane hopefully has some amount of linearity in it's behavior, the mix should be effective at a wide variety of rudder deflections and airspeeds. The mix, however, loses it's effectiveness at excessive rudder throws such as in knife-edge loops, where strange angular and aerodynamic forces come in and the fuselage becomes stalled and buffeted. This is where having a mutipoint curve in the mixing becomes handy. Difficult to program and takes lots of trial and error testing, but can really open the flight envelope by creating a very linearly flying airplane.

F3A airplanes are designed with as little coupling as possible, but most still need some mixing to get them perfectly de-coupled. In an ideal situation, the airplane will only yaw with rudder, pitch with elevator, and roll with aileron.

Normally, the mixes are small enough that at slower speeds, the slight mixed surfaces will not have much if any response, such as in 3D flight. In the 3D flght envelope, control effectiveness requires very large amounts of throw and a few % of mixing is not going to have any effect.

The one mix which I feel needs to be switchable is if I have a rudder-slaved mix (rudder slaved to the throttle or elevator). In 3D manuvers such as waterfalls, torque rolls, etc. I want the rudder to remain perfectly centered to keep the slow maneuvers straight. In precision flight, slaving the rudder to elevator or throttle as required can remove an airplanes tendency to yaw on pushes or pulls, and to correct any slight thrust offsets at different throttle settings.

A carefully trimmed out airplane, where any negative tendency is removed, creates a wonderful flying machine and the pilot gets to concentrate on placement of maneuvers, geometry, and style.

I know I have moved a little beyond your question, but I hope the additional information gives some food for thought. Most folks use very little of the radio's capacity. Computer radios are designed to allow the pilot to regulate any decent airplane to fly as perfectly as possible, not just to double the throws and go crazy.

Mark