1 (edited by eobet 2023-08-07 21:03:18)

Topic: Killing all frequencies above 11khz?

I have tinnitus and there's something called "notch" therapy where you filter all the frequencies you hear in your head and then play white noise. My tinnitus tones vary from 10khz to 13khz so I thought I'd try to use the ADI-2 to just kill everything above 10-11khz but I'm having trouble doing so:

https://i.postimg.cc/x1z2cYBV/Untitled.jpg

I guess I'm confused as to why, on the highest band I have to go as low as 6.5khz on the display, in order to get quite a sharp cut off at around 10khz, and I'm still seeing the EQ display sound at above 10khz?

(EDIT: I'm also using the treble "band" to try to make the shelf sharper, and that creates a huge peak in the graph that I'm using the cyan band at 5khz to compensate and level it out with visually... I must confess I feel like I'm using quite extreme values to try to make the graph look good so I wonder what in the end affects the music... the extreme values in the bands or the look of the graph?)

2 (edited by KaiS 2023-08-08 12:41:59)

Re: Killing all frequencies above 11khz?

Although not a medical doctor, I could tell you a lot about successful and failed tinnitus therapies - but this doesn’t belong here.

In short, this cut-out and replace variant misses the important part of pulsation of the replacement signal to be effective.
PM if you like more info.



According the filter:
ADI-2’s filters have a steepness of 12 dB / Octave.
You seem to expect much steeper filters, which could be achieved with cascading more then one, like you did.
But- for what you’re heading this is not induced, a steep filter will likely trigger tinnitus.


On the other hand, the Analyser has a 1/3 Octave band to band separation of ca. 14 dB, and 34 dB / Octave (3 bands apart).
So, a strong signal at 8 kHz will show with -14 dB in the 10 kHz band, even if there is no real signal present at 10 kHz.

Re: Killing all frequencies above 11khz?

Thank you for the reply and information!

I have a MiniDSP Ears which I used with REW to test, and the difference between what I thought was my pretty extreme EQ and no EQ at all made a surprisingly small impact in the resulting test graphs in that application... (although I must admit I'm very new at using both and do not understand everything yet!)