I have already designed my filters for 50R impedance, but the simulation shows a much different response on say 1K and even on 300R. I could include resistive impedance matching pi networks at inputs and outputs of the filter to 50R, but I need to have an estimation of the impedance of the sa612an to be able to match correctly. Attenuation of these networks wouldn't matter too much as the whole thing is used as a local oscillator and balanced modulator, so the signal levels are higher.