I’ve just lowered my very soft sounds (35db) by 3db as I found creaking of the floorboards very annoying.
After looking up about soft sounds, it mentions ‘threshold knee-point’ will be raised.
What does ‘threshold knee-point’ actually do?
I think the knee-point stops you hearing the HA noise.
TK (Threshold) Control
A way to adjust the AGC (compression) of an aid
This controls gain for soft input sounds only by adjusting the compression kneepoint over a ralatively low input level range (typically from about 40 to 55 dB).
As such, it is a gain booster for soft sounds. So, if there is too much circuit noise heard for soft sounds, move the kneepoint to a higher value, which will result in less gain for soft sounds. However, it will not affect the level of the MPO.
I believe that instead of compression, something called expansion is applied to soft sounds. Expansion is the opposite of compression.
A compressor with ratio of 2:1 will cut volumes above the threshold level to half. So if the compressor kicks in at 30 dB (threshold level) with a ratio of 2:1, a 40 dB signal would lead to a 35 dB output. This is because the 40 dB signal exceeds the threshold of 30 dB by 10 dB, but the 2:1 compression compresses this 10 dB to 5 dB. So the louder the sound, the more of the gain is cut by the compressor. One way to cause more compression at a certain level is to start compressing earlier (lower threshold), or less compression by starting the compression later.
When the signal is expanded, the most gain is applied to the signal at the threshold and least to the softest signals. Threshold is now the point where the expansion stops, so higher input levels are not expanded. Let’s say we set the expansion threshold to 30 dB and let’s say that a 20 dB signal is expanded to 25 dB. If we want to expand it to 22 dB decibels, one way to do this is to move the threshold higher. If we want a higher output, we move the threshold down.
I think this is essentially what you do with TK/Gain 35dB. If the threshold is set to 35 dB, a 20 dB input signal will get less gain than if the threshold was set to 30 dB. So essentially the lower the threshold, the more gain really soft sounds get. Raising the gain for 35 dB inputs means that also the gain for softer sounds is increased, but they increase more for the 35 dB sounds than softer sounds, effectively making 35 dB sounds stand out better.
1 Like
In Phonak Target we can’t directly affect the kneepoint, only the G35. However, following the manual as quoted by @Zebras suggests lowering the G35 would increase the kneepoint and thus following @bungybanana effectively amplify the 20dB range more.
This is something we can test, albeit a bit hard, since I for one probably haven’t heard sounds in the 20dB range for years. They are different sounds than we’re used to, not just softer sounds. Probably won’t know what to do with those cues.
We’re having a heat wave in Holland and we are not used to heat waves. (Usually only AC in cars and offices.) So I won’t be testing this until I am not feeling this hot. If someone does: Please post!
The threshold (knee-point) of the expander will probably be somewhere in the 45-55 dB range. 35 dB input will get more gain than 20 dB input, making the difference of output volumes more than 15 dB. What the difference in output is, is determined by the ratio that is probably fixed.
Normally we want to hear the soft speech at around 35 dB, but might find lower level sounds like noises from the fridge disturbing. Expansion allows us to put more gain to soft speech than quieter lower volume noise. Expansion has the same function in music, where the aim is to make meaningful soft input level sounds louder compared to background noise.
So if we raise the gain for 35 dB inputs, we also make those unwanted background sounds louder (although less so than without the expansion). The goal is to set the G35 high enough to hear soft speech naturally compared to normal 65 dB speech without the background noises coming too loud.
Too bad! I was hoping that the knee-point would be below 35dB. It is already maxed out in Phonak Target. So no extra optimization is possible there.
What are you trying to achieve if I may ask? 