Why feedback control should be disabled for music programs – Part II

Marshall Chasin
May 15, 2012

Depending on the three ways that acoustic feedback management is implemented, one should think twice about enabling such a feature.  The three ways that have been used are: (1) phase cancellation, (2) overall reduction of gain, and (3) notch filtering.

As discussed in Part I of this blog last week, phase cancellation is based on the generation of a sound that is 180 degrees out of phase with the offending feedback signal.  The two sounds (feedback signal and the generated out of phase signal) add up destructively with the minimization of the energy at that frequency region.  This is nothing new and the original article came out in the Journal of the Acoustical Society of America in 1933.  A drawback of this approach with music is that, unlike speech, instrumental musical signals can stop abruptly such that the generated signal is heard as a brief chirp.

The overall reduction of gain equally for all frequencies is another approach.  It does have the benefit of maintaining the shape of the frequency response, but can result in overly reduced gain.  This approach can actually “turn off” the hearing aid if the input is intense.

And finally, last (and probably also “least”), notch filtering, as the name suggests reduces the gain in the general vicinity of the feedback signal  and depending on its implementation, can “hop around” as the feedback frequency changes.  Some claim that this “frequency hopping” phenomenon can be perceived as blurriness but personally I have never heard this.  A major problem with this approach for music is that it may alter the important balance between the lower frequency fundamental energy and the higher frequency harmonic structure.  Unlike speech, it is really the balance between the lower frequency sounds and the higher frequency harmonics that is crucial for good music fidelity.

But does music as an input to a hearing aid need to be treated in the same way as speech?  Music is generally much more intense than speech, is typically a broader bandwidth than speech, and as such probably doesn’t require as much amplification in order to obtain a desired output.

A person with a 50 dB sensori-neural hearing loss at 1000 Hz may require 20-25 dB of gain for speech (depending on the fitting formula), but that same person may only require 0-5 dB of amplification for music.  The output in the ear canal will be the same but the reduced gain that is imparted to the signal will improve the stability of the output.  Feedback, even sub-audible feedback, will be less of a problem.

The output of a musical instrument, being a hard walled structure, tends to have larger crest factors than are typically found for sounds emanating from the human vocal tract.  The peaks are higher relative to the average of the signal for musical instruments versus those of a typical speech signal.  It turns out that the typical musical harmonic looks more like acoustic feedback than do the harmonics of the vocal tract.   Feedback management systems will tend to engage and attempt to delete or minimize musical instrument harmonics when they should not be deleted.

Having said this, there are some rather ingenious algorithms such as those based on a comparison between the input to the front microphone and the rear microphone ports in directional hearing aids- these can help discern whether a narrow and peaky spike of energy is a musical harmonic or the result of acoustic feedback.

My clinical suggestion is to disable the feedback management system whenever possible when it comes to music.

But, what if your favorite hearing aid does not have the capability to have a disabled feedback management system?  What are the alternatives?

If the music is rock, jazz, or anything modern, my clinical rule of thumb is to go with phase cancellation as the feedback manager.  These types of music rarely have staccato or rapid on/off changes so any “chirping” caused by a rogue generated feedback signal would be masked by the music.

If however, the music is more in the classical vein, or even some of the more modern film scores for horror movies, that has sudden stops, staccato notes, then chirping may be heard.  I never watch horror movies anymore- I actually have beautiful straight hair until I watched Alien, but if I did watch horror movies and was wearing hearing aids, I would not use phase cancellation- I suspect that the best approach would be to use an overall reduction in the gain, since this would maintain the balance of the lower frequency fundamental components of the music and the higher frequency harmonics.

So, disable the feedback management system whenever possible for music, and don’t go to horror movies.

Leave a Reply