Channels vs. Bands

I have been doing a lot of research on hearing aids. I have read several websites which discuss channels and bands. I have failed to comprehend what the difference is. So far, it seems like a number of channels make up a band and the band can be adjusted by the audiologist. Does anyone want to try explaining the concept for me? Thanks.

Channels usually refers to wide dynamic range compression or noise reduction processing. Bands are like the controls on a graphic equalizer, used to adjust the frequency response. Some companies use the terms interchangeably. Some hearing aids have more channels than bands, or an equal number of both. You’ll also see terms such as fine-tuning channels, fitting bands, gain handles, processing channels, frequency channels, and compression controls.

Thanks. I have seen most of these terms and have not learned enough yet to know they were all related.

I am a newbie wearing Phonak Brio R312Ts with open-dome RICs for about 6 weeks. I have had two adjustments for persistent (1) “white noise” in my left HA that accompanies loud speech or TV and (2) reverberating (echoing?) notes in both HAs when I play the piano. In both situations the unwanted sounds last less than a second after the triggering sound stops. Because he is new to Phonak, my HIS at Costco calls Phonak tech support, who walks him through adjustments. Recently they reduced my number of channels from 20 to 3 and turned off Whistleblock. This has reduced but not eliminated the unwanted sounds. I do not know if the lower number of channels is present when I turn the HAs on or only when I activate the program they set up to lower the noises. Since the default number of channels was 20, what am I giving up with the setting being turned down to 3 (intermediate numbers were available)? For example, my voice might sound a little different now, and I might have lost some speech recognition (am still at 80% of target). Also, can you suggest any other adjustments to eliminate the left-sided white noise? Thanks.

Hi,

for piano-playing, you´ll definitively have to turn off whistle-block. You need a program with everything (I mean every extra feature like noise-reduction, whistle-block, whatever) turned off.

As for the channels, I can only guess: Do you know what compression is? To put it simple: Makes loud sounds softer, soft sounds louder. If you have 20 bands in your compression, it can make deep loud sounds softer while still making high soft sounds louder. This is an advantage when you are in an environment with deep loud noise but still have to understand speech (for instance, people running around making deep noise with their feet on the floor, someone standing near you speaking to you).

A 20 band compression takes your sound apart into 20 slices. Then each slice is compressed according to it´s loudness. Then it´s put together again.

Now, why is this bad for piano playing? A single note on the piano consists of multiple frequencies, the base frequencies and overtones. Now the problem: The overtones are softer than the base frequency, so it might happen that the overtones are amplified more than the base frequency, resulting in an unnatural sound. Even worse: Some overtones are non-harmonic. If those are over-amplified, you get a real nasty sound.

So if you reduce the number of channels, it is less likely that your piano-sound is taken to pieces and becomes unbalanced.

So, as I explained above, you might get less understanding in noise with fewer channels. Now comes the difficult part: If also the bands (not channels) of the equalizer are reduced to three, this would be very bad. Then you cannot adapt the hearing aid exactly to your hearing loss anymore. So it would be good if you have three compression channels and still 20 frequency bands. Ask your audi about that.

As for the white noise, this sounds like an error to me. Of course, hearing aids produce noise, but this should be 20 dB or lower, and with your hearing loss, you shouldn´t be able to hear it. Ask for a replacement.

Best regards,

Musician_72

Very helpful. Thank you!

That’s hilarious, they need to do some better training. He’s not reduced the number of channels at all, he’s just grouped them into three adjustment levels.

Turning of whistleblock is certainly a step worth doing and possibly removing the feedback settings after the initial feedback manager set-up so that the aid is allowed to hit louder notes clearly without the phase canceller trying to jump all over the signal.

The other thing he could probably do is to look at the resonance that he aid is developing on his REM system (you might need to remind him to put it into the REAR Live speech mode) and make the kind of loud noises that cause the problem. While observing this window, he should note the pitches that are causing the issue and make corresponding adjustments to the levels in the Phonak software. Relying on the idealised output graph of the Phonak screen is not the best approach: especially when the customer is reporting your sort of level related issues.

Hi Um bongo,

thank´s for the expert to chime in.

On the phonak site they say that though they use a multiband-compressor, the bands are somehow “linked” because it would be bad if the channels were completely independent. Could you elaborate further?

But is my explanation, at least in parts, correct? I once had then phenomenon that with a siemens aid (also multichannel) the consonants of a choir suddenly went wrong (i heard different consonants from what were really sung) as the choir got louder. I thought that some parts of the signal were compressed differently from other parts.

I now use a channelfree aid, excellent for music :slight_smile:

The reason for grouping and overlapping channels with similar compression effects is fairly obvious if you think about it: You’re squeezing the amplitude level of the sound wave for certain input levels but not others, if you do very different things in adjacent channels you’ll introduce temporal distortion as the sound shifts from one band to another. This is also the real difference between channels and bands: bands allow you to futz around with level, while channels allow you to control gain/compression/noise management: but not anything radically different in terms of performance in adjacent channel. Bands can virtually be set on/off at whim, subject to to the roll-off between each.

The issue with the Siemens aids might be them struggling to deal with singing at high input levels. Could be that at the peak resonance the aids can’t decide whether they are feeding back or not. One of the methods of dealing with feedback involves shifting the peak of the resonance a few hertz away from its actual value. I expect the aid is transitioning between amplifying reasonable level sounds without peak shifting, but when the level goes up the impact on the higher pitched sounds is marked.