Audiology Refresher 2
Presented in November, 2002
by Joe O'Donnell
Different Hearing Aid Technologies
Analogue to digital
"Analogue", "Digitally Programmable", "Digital"
What do we understand by these labels?
Three Hearing Aid technologies
- Conventional analogue
- Digitally programmable (DP, Programmable, Hybrid)
- Digital (DSP, Fully digital)
What is Analogue signal processing?
- Analogue signal processing is the representation of a continually changing variable, such as sound, by another physical variable, such as an electric current.
- These signals vary according to the pattern of the sound.
- SPL at the microphone is continually changing and voltage in the microphone is continually changing.
How does a conventional analogue hearing aid work?
- Analogue hearing aids, like all hearing aids, have a microphone which picks up sound in the environment and converts it to small electrical signals.
- The electrical analogue signal is then amplified and shaped by transistors and circuits in the signal processing part of the aid (amplifier).
- Shaping and output limiting are performed by screwdriver potentiometer adjustment.
- The amplified and shaped electrical signal is then sent to the receiver (loudspeaker) to be changed back to an acoustic analogue signal and sent to the ear of the hearing aid wearer.
- Analogue hearing aids can have many complex circuits at the signal processing stage but there is a limit to the number and tends to cause interference.
What are digitally programmable aids?
- These aids use analogue signal processing at the amplifier stage.
- However, they do have a digital chip at the amplifier stage that allows shaping of the frequency response and output limiting to be altered by a computer rather than a screwdriver.
- Research has suggested that use of a computer allows more precise adjustment of the output signal.
- This aid, unlike conventional analogue, allows for the choice of different programmes from the memory to suit different listening conditions.
- Some of these aids also have a remote control hand unit which allows the user to adjust the volume and operation mode.
- Volume control and other switches can be made inoperable.
How does a Digital hearing aid work?
- The microphone turns sound into an analogue electric signal.
- Ant-aliasing filter removes inaudible frequencies, which can confuse the converter.
- The signal is then ?sampled? a given number of times per second (10,000 times/sec, or greater).
- Sampled signal allows analogue to digital (A/D) converter to turn electric signal into digital equivalent.
- The microprocessor operates on the digital signal.
- Digital to analogue (D/A) converter converts the digital signal back into an analogue electric signal.
- Anti-imaging filter smoothes the signal so it sounds natural again.
- The receiver converts the analogue electric signal back to sound.
What is different about digital aids?
- Digital aid transforms the analogue info from the mic into a different form.
- It takes the analogue signal from the microphone and transforms it into 'bits' of data - numbers that can be manipulated by a tiny computer in the hearing aid.
- Digital signal processing of the speech signal makes it possible to tailor and process the signal very precisely and in ways that are impossible with analogue aids.
- The better digital aids can be very finely adjusted to suit an individuals hearing loss very precisely.
- Many now have noise canceling systems, speech recognition and feedback suppression systems.
- However, just because an aid uses DSP it is not necessarily the best aid for an individual. May be audiological reason for fitting an analogue aid or a digital aid that acts like a linear analogue aid.
- In near future all aids will be digital. Therefore, the fact that the aid is digital does not matter. What matters is what digital chip does with the speech signal.
What does research say?
Research falls into two groups:
analogue v digital
digital v digital
- Most use objective ( speech test results) and subjective (preference ratings) data.
- Most of research shows subjective data favouring DSP but objective data unclear as to which aids are best.
- Need for "blinding" in research.
Therefore lot of ambiguity as to benefits and lot of hype
Advantages of Processing in Digital
- Low noise - more pleasant for wearer, sound is clearer, more amplification of soft sounds is possible.
- High Fidelity - less distortion, best possible sound quality.
- Unlimited Processing - flexibility in fitting, specialised functions, ?intelligent? compensation.
- New capabilities via simple software updates - update fitting as the user's ability changes over time, add new capabilities as users identify problems, implement new, unforseen compensation techniques as they are developed.
- Bands and Channels
One line of development for DSP technology is to use its processing power to refine and enhance the 'conventional' processing strategies of analogue hearing aids, such as amplification, frequency shaping and compression. However, the DSP chip has the power to manipulate the signal in completely new ways, in an attempt to meet individual hearing needs. These manipulations are called algorithms The binary digits produced by A/D conversion are the raw data to be used by the processing chip. This raw data can be manipulated in a multitude of ways by the CPU or microprocessor. Manufacturers decide how to manipulate the raw data in the CPU to achieve particular outcomes. These manipulations or algorithms determine exactly how the speech signal is to be processed. Therefore, algorithms are mathematical computations which perform specific tasks such as controlling the frequency response of the instrument, controlling loudness growth, or enhancing the speech signal in background noise. Algorithms can be said to determine the modifications that will be made to the sound signals before they are delivered to the ears of a listener. Each manufacturer will determine the particular algorithms used in their hearing aid and each manufacturer may well have different algorithms to perform the same task. This is why one digital hearing aid can be so different from another.
Frequency Bands and Channels of compression
Digital hearing aid design demands choices, precisely because digital processing is so flexible. The designer can decide to incorporate high level algorithms in the hearing aid or alternatively use the processing power to provide more flexible processing of amplification, frequency shaping and compression. Clearly, most digital hearing aids on the market use a combination of the two approaches which results in one company's digital hearing aid being completely different from another manufacturer's digital hearing aid. Therefore, each digital hearing aid is truly unique.
In order to have more control and flexibility over amplification, frequency shaping and compression, the digital hearing aid has to split the digital signal into a number of bands and channels covering the speech frequency range. Bands are used for frequency shaping of the signal. The greater the number of bands then the greater the degree of flexibility in manipulation of the hearing aid's frequency response. Channels are used for compression. Channels divide the frequency response of the hearing aid into a number of areas where compression can be applied independently from other areas. For example, the Widex Senso hearing aid is a one band, three channel hearing aid; one band for frequency shaping and three channels for compression. Unfortunately, manufacturers often use the terms interchangeably. The Oticon Digifocus DSP hearing aid has a 7-band system . See figure
Although digitally programmable hearing aids do have adjustable responses it is suggested that the use of digital filtering increases this flexibility while maintaining absolute stability and linear phase (Murry and Hanson 1992). Linear phase characteristics are important in order to avoid ripple when the skirts of two adjacent bands overlap ( Levitt et al., 1990). Therefore, by using the flexibility of the digital processing system to split the digital signal into a number of different frequency bands, it is suggested that the frequency response of the digital hearing aid can be specified with greater precision than a hearing aid using analogue signal processing. The present study will seek to investigate this claim.
Main advantages of Digital technology due to.....
The Real Issue
Don't think about Digital versus Analogue
The real issue is ......
Linear versus Non-Linear
It is what you do with the digital chip that matters not the fact that it is digital
- Investigation into the fitting flexibility of the three hearing aid technologies
- Looked at moderate and severe hearing losses
- Also compared NHS to commercial analogue
- Used DSL amplification targets
Hearing aids fitted to moderate hearing loss audiograms
The graph shows that the DSP hearing aid had a closer fit to targets than any of the other hearing aids. The digitally programmable hearing aid had a generally closer fit to targets than the two analogue hearing aids and the two analogue hearing aids deviated around similarly poorer fit to targets.
Box-plots of the hearing aids' goodness of fit to targets generated from moderate hearing loss audiograms. Each box indicates the data spread and corresponds to the interquartile range. The horizontal line inside the box is the median. The highest and lowest scores, excluding any outliers, are indicated by the lines extending beyond the box. An outlier is denoted by a O. It has a value more than 1.5 box-widths away from the box. The number next to the circle is the case number.
Hearing aids fitted to severe hearing loss audiograms
The DSP hearing aid again had a closer fit to targets than any of the other hearing aids, with the digitally programmable hearing aid showing an overall closer fit to targets than either of the analogue hearing aids. The analogue aids displayed the poorest fit to targets.
Boxplots of each hearing aid's goodness of fit to targets generated from severe hearing loss audiograms
Comparison of moderate Vs severe audiograms' goodness of fit to targets