Pick your loss at one frequency. Look at the G65 (amount of gain for 65dB input) then work it out like this:
Loss of 40dB (S/N); Gain circa 13dB + input 65dB = output 78dB
Loss of 40dB (cond.) Gain circa 20dB + input 65dB = output 85dB
Loss of 60dB (S/N) Gain circa 20dB + input 65dB = output 85dB
Loss of 60dB (Cond.) Gain circa 30dB + input 65dB = output 95dB
Loss of 80dB (S/N) Gain circa 27dB + input 65dB = output 92dB
Loss of 80dB (Cond.) Gain circa 40dB + input 65dB = output 105dB
Mixed losses will combine proportionate elements of both. Extrapolate away for figures either side or different input levels.
Actual manufacturers will bend these rules to fit their Hearing Aid strategies and particular algorithms, according to whichever fitting philosophy or dynamics they incorporate.
So you’re unlikely to get an exact implementation of the above whichever way.
You’re ignoring this member’s posts.
