Why filter bank design is critical for effective hearing aids
Modern hearing aids are a medical technology marvel, specifically in the area of digital audio processing. They capture sound waves, convert them to digital signals, process them through the device, then deliver a new sound wave to the ear automatically optimized for the setting and wearer. Digital audio processing is not new, but when it’s done through a medical device like a hearing aid, it can change lives for the better.
One in every eight Americans aged 12 and older has hearing loss in both ears, and a majority would benefit from treatment. If left unaddressed, hearing loss can contribute to significant cognitive decline, in addition to depression, anxiety and isolation. Meanwhile, studies suggest that socialization and daily participation in everyday life — often enabled by hearing aids — boost cognitive function, helping to keep the brain stimulated. Basically, the worse you hear, the less you participate. The less you participate, the worse for your cognitive health.
The importance of signal processing: volume versus clarity
For hearing aid wearers, the challenge isn’t just about hearing more (volume). It’s about hearing better (clarity). For example, if you’re in a noisy restaurant, you don’t want the noise to be louder. The best hearing aids help wearers limit background sounds so they can clearly understand the words people say. And hearing aid technology sometimes gets in the way of that clarity.
Hearing aids sound “tinny” and artificial when there’s a delay in sound processing. Most hearing aids today take 5-8 milliseconds to process sound. That may not seem like a long time, but when coupled with direct, immediate sound that enters the ear canal by passing through or around the earpiece, the resulting sound is distorted and unnatural. That’s why signal processing is so important. How long it takes and the techniques manufacturers employ can make all the difference.
Overall, most digital hearing aids process sound almost the same. A microphone picks up sound waves and converts them into a digital signal. An analysis filter bank splits the signal into multiple frequency bands, which can then be compressed, amplified and processed depending on the listening situation. After that, the bands are brought back together in a synthesis filter bank and that’s what’s heard through the hearing aid ear tip.
This approach underscores how important filter banks are for digital signal processing. Choosing the right filter bank is key to natural-sounding hearing aids.
Choosing a filter bank
Frequency-domain filter banks are the popular choice for hearing aid makers, and there are some convincing reasons why. But to deliver the most natural sound, those in the hearing industry should consider time-domain filter banks.
The obvious question is: What’s the difference between the two? The answer lies in time and frequency resolution. In any filter bank, the wider the frequency band, the greater the time resolution. Therefore, a bandwidth that is five times wider has a time resolution five times higher (faster). Conversely, a narrower bandwidth means poorer time resolution (slower).
With frequency-domain filter banks, the width of the frequency bands is always the same. This means all bands are kept relatively narrow because bandwidth is set based on what’s needed for the lowest frequencies, where the ear’s sensitivity is highest. Because of the correlation between time and frequency, all bands operate with the same, relatively poor time resolution.
Time-domain filter banks allow manufacturers to use filters of varying widths. Therefore, product designers can set bandwidths any way they like. With the bands varying in width, so does the time resolution.
Recreating real hearing
What do adjustable widths mean in practice? Time-domain filter banks allow hearing aid designers to narrow bands at lower frequencies and broaden them at higher frequencies, resulting in the same trade-off between time and frequency as humans have in their own ears. A healthy human ear has higher frequency sensitivity at lower frequencies and lower sensitivity at higher frequencies.
Recreating human speech is a complicated endeavor. Spoken consonants are short in duration, but high and wide in frequency. Spoken vowels are longer in duration but lower and narrower in frequency. This means the sound signals would most effectively be processed by filters where the high frequencies have high time resolution while the low frequencies have better frequency resolution — something only time-domain can achieve.
Staying away from down-sampling
One more thing to consider when designing a hearing aid is down-sampling. Hearing aids are vital medical devices for those who wear them. Ideally, they last all day before needing to be recharged, which means power-consuming functions like signal processing need to be kept in check. Down-sampling can help conserve battery, but at the price of sound quality.
Frequency-domain filter banks make down-sampling straightforward because of their narrow filters. With time-domain filter banks, the benefits of down-sampling are minor and don’t outweigh the risk of diminished sound quality. Therefore, a time-domain filter bank requires other available solutions for keeping power consumption low. But by avoiding down-sampling and preserving the original signal content, manufacturers achieve higher fidelity, lower the risk of artifacts and improve natural sound quality.
When designing a hearing aid and refining signal processing, frequency-domain filter banks may have their benefits, but a time-domain filter bank will deliver the best result for the wearer.
Article source: Medical Design & Outsourcing