Lecture 7: Hearing Aid Features Flashcards
How many channels do current hearing aids have?
10-30
How many channels are needed to accomplish what is needed?
5-8
What is the purpose of a lot of channels in a hearing aid?
- Music processing
- More channels for a sharply sloping hearing loss
What is signal classification?
- Comparison with decision rules that specify actions and time constants of the hearing aid under different circumstances
1) Time between
2) The speed of the execution of the decision rule
3) Time between the detection of the change in the sound scene and the release of the current set of decision rules
4) The speed of releasing the current set of decision rules
What is noise reduction in a hearing aid?
Goal: to reduce hearing aid output in presence of noise
- the goal is not to improve the SNR
- the goal is to make it more comfortable to listen in noise
What is noise?
- the signal that you are not focusing on
What is digital noise reduction (DNR)?
- Varies by hearing aid manufacturer
- Temporal modulation of speech is generally around 3 Hz
- Amplitude fluctuations 30-50 dB
- Many manufacturers establish their gain reduction rules based on some combination of modulation frequency and modulation depth
What is the threshold (SNR) of activation for DNR?
- Setting it too high brings too much noise in
- Setting it too low reduces the audibility
What are the different kinds of Noise Reduction algorithms?
1) Modulation-based
2) Spectral subtraction
3) Transient Noise
4) Wind Noise
What is the modulation-based noise reduction algorithm?
- Reduces listening effort, aversiveness of sounds
- Increases listening comfort and cognitive capacity in background noise
- Uses modulation rates to identify the presence or absence of speech
- Estimates the SNR in the incoming signal
- Estimates the SNR of the incoming signal by tracking the peaks & valleys of the signal
- The info. is compared with a set of decision rules
- If the modulation rate is outside the range of speech, and/or the modulation depth is low, it is assumed to be noise and the gain of that channel is reduced
- If the modulation rate is inside the range of speech, and/or the modulation depth is high, the gain of that channel is minimally altered
What is the modulation rate of speech?
2-10 Hz (slow modulation rate)
- Noise are sounds with modulation rates less than 2 Hz or modulation rates greater than 10 Hz
What is the spectral subtraction algorithm?
- Estimates noise spectrum
- Shape the “speech in noise” spectrum
- Subtract to achieve “clean speech” spectrum
Limitation:
- estimates only, not actual values
What is the transient noise algorithm?
- Designed to reduce the level and annoyance of transient noises (ex. people clapping hands)
- Characterized by rapid rise times and a peak level that is much higher than the RMS level of other sounds
- The amount of gain reduction depends on the peak to RMS ratio
- Time constants are typically in milliseconds
What is the windy noise algorithm?
- Wind blows across the microphone diaphragm, creating the wind noise (bad for bike riders & golfers)
- Identifies the presence of wind noise by computing the correlation between the outputs oft he 2 omnidirectional microphones
- Infer the presence of wind noise if the correlation is low
What are the strategies to reduce wind noise?
- Low frequency or high pass filtering
- Omni directional microphones in all frequency channels or in low frequency channels (omni directional mics have low wind noise levels)
- Microphones with low amplitude outputs
- Summing the output of the two omni mics as the input of the HA increases because when correlated signals are summed the overall level increases by 6 dB, when uncorrelated signals are summed, the overall level only increases by 3 dB
