The Fourth of Many Technical Innovations- part 6 of 7

Marshall Chasin
May 26, 2015

Post 16 bit architectures

There is a great formula for the calculation of the dynamic range based on bits.  It is 20nlog2.  You can use this to impress your friends, and even (if done properly), to make you the hit of the party.  Where else can you hear “log2” and “please pass the shrimp” in the same sentence?

The hearing aid industry had selected a 16 bit architecture for their hearing aids way back in the infancy of digital hearing aids almost 25 years ago.  This made a lot of sense because the quantization error (the difference between the original analog signal and the digitized one) was so low that people with normal hearing would not be able to hear the noise. Decreasing this low end of the dynamic range even further (say with a 17 or 18 bit system) would be a waste of time- once something is not audible, making it less audible still, is silly. But, the other side of the dynamic range that is predicated on the selection of the number of bits is the highest sound level that can be digitized with an analog-to-digital (A/D) converter.  In theory, with a 16 bit system (20 x 16 x log2), it is 96 dB.  And of course this 96 dB has no suffix after the dB.  It is a difference measure between the quietest and highest level that can be digitized.

In reality, because of any number of engineering and design decisions and issues, the effective dynamic range is on the order of 80-90 dB. This limitation for the upper end of the dynamic range has been the subject of this 7-part  series. Recently a number of hearing aid manufacturers and third party IC manufacturers (such as ON Semiconductor and Motorola) have come out with post-16 bit chips that have allowed a larger dynamic range to be obtained.

There is nothing magic about  a 24 bit system, but for a number of reasons (some of them marketing) the hearing aid industry appears to be settling into a 24 bit architecture.  Again, depending on the engineering and design decisions, the effective dynamic range may limit the highest level that can be digitized to somewhere in the mid-100 dB range. Also in the past while a new series of hearing aid microphones have become available.  In the distant past (last year), the limit of modern hearing aid microphones was 115 dB SPL with about 3%-4% distortion.

Newer hearing aid microphones can reliably transduce 119 dB SPL with about 3%-4% distortion.  In the near future we may see marketing ads reflecting this by stating that a certain hearing aid can handle inputs of 119 dB SPL, and, barring any design decisions that would need to be made, ads such as these would probably be fairly accurate. So, let’s do some math- Dynamic Range = 20nlog2 = 20 x 24 x 0.3 = 144 dB.

Given that modern hearing aid microphones can even now transduce only 119 dB SPL, 144 dB may not sound as if it’s really required, but it does give the hearing aid design engineer a bit more flexibility to play with various design formats and algorithms that may effectively reduce the maximum sound level that can be digitized to somewhere in the mid-100 or even mid-110 dB SPL range. So even a 24 bit system may really only function as if it were an 18-20 bit system. But this is better than starting out with a 16 bit architecture and ending up with the equivalent of only 12 or 13 bits.

Some examples of post-16 bit architecture are the North platform from Unitron, the Venture Platform from Phonak (actually these two are the same platforms), and the Inium Sense from Oticon. As stated in earlier parts of this series, Bernafon, Resound, and Widex have their own technologies to  handle the higher levels of music. As always, it’s best to contact your manufacturer’s local representative (or in-house audiologist or engineer) to get the most up-to-date information.  Any absence of manufacturer name or product here is due to my lack of knowledge, and in some cases the manufacturers have simply not returned my calls … yet.

Leave a Reply