Critical gain (feedback) management

The Connexx software for Rexton/Siemens HAs can calculate and apply a critical gain measurement to manage feedback. However, when the critical gain measurement is applied or “optimized” to a particular program, this results in a decrease in the gain being applied. My HIS at Costco hates the way Connexx deals with feedback management by “basically chopping off the gain curve”, and therefore usually does not even use it. He noted that ReSound does a much better job with feedback management without compromising the level of gain. The HIS demonstrated this by performing a critical gain measurement, and asked me to watch the gain curve as the measurement was “optimized”. In my case, gain was reduced by at least 5 dB in critical areas. Since the optimized gain (based on a real ear test), would be decreased to manage feedback and would affect my speech recognition, the HIS recommended not using it unless the feedback was a real problem. This seems like a major compromise in an attempt to manage feedback. Would calculating critical gain, but not optimizing to a program do anything to manage feedback, or must the gain be optimized to have any effect?


there are different ways to deal with feedback. One is described by you: Make a measurement and then cut off the frequencies that produce feedback. This, of course, takes away some gain. If this affects your speech recognition, you will have to go a different route.

You can turn on the feedback suppression. This is some automatic detection if there is feedback present. Then the aid does one of several things:

  • suppress only the frequency that is detected
  • turn the phase of the amplified sound
  • alter the frequency of the amplified sound

The last one has the advantage that it really works and doesn´t affect the amplification, but it has the disadvantage that it sounds terrible. Just play one note on a piano and it will sound like a honky-tonk!

With feedback-killers, you can get more gain (5 to 10 dB, I think).

If that´s not enough, you´ll have to go to a more occluded fitting. If your ear is totally closed, in theory there cannot be feedback. In fact: I sometimes use power-domes for hearing protection, and if they sit correctly in the ear, I can cup my ear and do whatever I want, there is no feedback.

Seeing your audiogram, I guess that you have a more or less closed fitting anyway. Maybe you can close it a little bit more and then get rid of the feedback without any feedback-suppression.


Thanks for the help. The “4C feedback preventer” is enabled and set on “fast” for my Trax42. This is likely the automatic detection that you were referring to. I believe it was initially set to “slow”, but then the HIS changed it to “fast” when I mentioned the feedback issue. I did not notice any difference between slow and fast, but perhaps it is working to some degree with it turned on. The feedback is really only an issue on the left side, and is by no means severe. I am using tulip type domes. I also tried soft molds, but they did not work and my speech recognition with them was terrible for some unknown reason. My HIS probably has it right then by using the feedback preventer and not reducing the gain by measurement and chopping off the frequencies producing feedback. I thought the feedback (critical gain) measurement would suppress frequencies causing feedback that were at a level above the gain curve without affecting the actual gain. When I was using the KS6 (Resound), I had no feedback issues using the same type of dome, but overall the Trax worked much better for me.


With those numbers custom molds would be best

I’ve managed to avoid custom molds so far. I am using the double click domes. I had feedback issues after my initial fitting, but when I returned, two things were done. The feedback suppression was set to Fast and my basic fitting formula was changed to the Rexton OEM Power formula, which in my case, pulled my curves just outside of the feedback range. The result pf these two changes was more useful gain, better bass and mid-range and no feedback. YMMV.

The only artifacts that I have noticed is that high frequency beeps (microwave etc.) are distorted a bit, but the effect is very minor and not annoying to me, and I can easily hear the beeps and locate them directionally.

You aren’t get your money worth out of those aids and that audio-gram without custom molds.

Doc Jake, aside from feedback management, what would I expect to find different if I get the custom molds?

My custom molds are the soft kind, and have the adjust-a-vent, which is a large vent hole with a set of inserts of different sizes, so the effective vent size can be adjusted by just changing the insert.

They were fit very deep and do not move during the day. They are very comfortable and cannot be seen from the outside. You just see the wire going in, much like the look when using domes.

Some advantages are 1) they don’t back out, 2) cannot be seen when wearing, 3) are very comfortable, 4) reduce feedback.

I would ask for the very deep fit. It is probably not automatic that they fit them that way. I had tulip domes, which I like, but I would not go back.

Use high power receivers, change to the NAL-NL2 fitting formula, make sure you are set to experienced user profile, get custom molds with a .8mm vent (small) and recalculate the programming to prescription. Then get hooked up to a real ear verification machine and make sure you are actually hitting NL2 targets. You should be able to hear quite well… assuming your WRS is decent. That would be your only limiting factor.