Second Exam Flashcards
(103 cards)
1
Q
when was analog technology popular
A
until the end of the 1980s
2
Q
how were adjustments made to the signal in analog technology
A
potentiometer trimmers (aka trim pots)
3
Q
very basic what is analog signal processing
A
all physical signals are continuous in time and amplitude
4
Q
digitally programmable analog (DPA)
A
- first generation of digital HAs (but did not digitize the signal)
- analog processing at the amplifier stage
- electroacoustic characteristics could be altered via computer programming
- –shaping freq response
- –limit HA output
- signal still stays analog (continuous time) throughout the entire process
5
Q
digital signal processing at the level of the input signal
A
- DSP transforms an analog signal into a digital signal
- how?
- –HA must “sample” the input signal
- –input signal is sampled at discrete points in time
- –only these sampled points are used by the digital signal processor
- –the rest of the input signal is not considered
- –sampling of the signal takes place in the analog-to-digital (A/D) converter
6
Q
digital signal processing after sampling
A
once the input signal has been sampled, quantization occurs
- turned into binary code
- the signal is either on (1) or off (0)
7
Q
what are the specifications for digital signal processing (DSP)
A
- instructions/second (millions/second or MIPS)
- sampling rate
- number of bits
- current consumption
- processing delay
- physical size
8
Q
what advanced signal processing does DSP allow for
A
- adaptive feedback control
- adaptive directional mics
- wind block
- echo block
9
Q
advantages of analog over DSP
A
- established technology
* cost
10
Q
disadvantages of analog over DSP
A
- limited signal processing
- no advanced features
- adjustments made with trim pots
- may not be able to find any for sale
11
Q
advantages of DSP over analog
A
- miniturization
- low power consumption
- stability
- programmability
- complexity
12
Q
disadvantages of
DSP over analog
A
- longer time delay b/t input and output signals than analog
* cost
13
Q
what is an algorithm
A
- a process or set of rules to be followed
* basically, a set of instructions for how the DSP will manipulate the signal
14
Q
what is the sequence HA uses to process sound in terms of algorithms
A
fitting algorithm
programming algorithm
sound-processing algorithm
15
Q
fitting algorithm
A
- defines the parameters of amplification for each pt based on the audio
- manufacturer software will give recommendations for gain/output
- once pt data has been entered into software, “first fit” process will begin
16
Q
programming algorithm
A
- set of instructions to download the fitting algorithm data to memory of HA
- must have temporary connection b/t HA and manufacturer’s software
17
Q
sound processing algorithm
A
- defines how the HA will process the incoming signal
- how the HA actually manipulates the signal
- algorithm is completely downloaded inside the HA
18
Q
the overall process of fitting
A
1) audio date from the pt is measured by fitting system
2) fitting algorithm= algorithm for specific pt
3) programming algorithm= programmer puts algorithm and data into HA memory
4) sound-processing algorithm= programmed DSP HAs
5) final fitting on pt
6) verification of correctness of fitting
19
Q
what are the two goals of fitting HAs?
A
1) provide appropriate gain
2) increase SNR
20
Q
how do we accomplish the two goals of fitting HAs?
A
1) we provide appropriate gain using gain and compression
2) we increase SNR by using directional mics and noise reduction algorithms
21
Q
I/O function
A
a graph of input and output where input SPL is on the x-axis and output SPL is on the Y-axis
22
Q
gain basic definition
A
the difference between input and output intensities
23
Q
Maximum Power Output basic definition
A
(MPO)
*max SPL the hearing aid will produce
24
Q
compression basic definition
A
gain of HA is varied based on the intensity of the incoming signal
25
what are the two static aspects of compression?
* compression threshold (CT)
* compression ratio
*static aspects of compression determine when and how much compression will be applied to the signal
26
what is the compression threshold (CT)
* the input SPL where compression begins
* sometimes referred to as:
- --kneepoint (KP)
- --threshold kneepoint (TK)
27
what is the compression ratio?
the change in input level needed to produce a 1 dB change in output level
*the amount of compression is provided by the hearing aid
28
input compression
* in analog aids: volume control (VC) placement dictates type of compression
* for input compression:
* aka Automatic Gain Control (input) (AGCi)
* VC located between amplifier and receiver
* VC affects both gain and MPO
* used in moderate-power HA's
* signal reaches compressor before it reaches VC, therefore, it does not affect KP
29
output compression
for output compression:
* aka automatic gain control (output) (ACGo)
* VC located between mic and amplifier
* used in high power aids
* placement of VC only affects the gain, not MPO
* VC affects amount of input that reaches the compressor
30
clinical uses for input compression (ACGi)
* Mild-to-moderate SNHL
* Larger dynamic ranges
* Large fitting application
31
clincial uses for output compression (ACGo)
*Severe-to-profound SNHL
*Very small dynamic ranges
*Helps prevent damage to
residual hearing
*Often selected for children
32
MPO compression control
aka output limiting control
* typically used with output compression
* affects KP and MPO
* does not affect gain
Clinical uses:
* limits MPO
* protect residual hearing
33
threshold-Kneepoint "TK" control
* used with input compression
- --specifically with wide dynamic range compression
* affects KP and gain for soft and moderate inputs
- --,60dB SPL
- --noes not affect MPO
* clinic uses:
- --imitates function of OHC
34
output limiting compression (OLC)
*associated with output compression & MPO control
*high KP and high CR
---high KP of >60 dB SPL
---high CR is >4:1 to 10:1 (textbook definition)
-----clinically, >8:1 to 10:1
*Provides a strong degree of compression over a
narrow range of inputs = Super compression
*OLC with MPO control provides less distortion than
peak clipping
*Limits the output of HA to protect residual hearing &
avoid loudness discomfort with loud incoming sounds.
35
wide dynamic range compression (WDRC)
* associated with input compression and TK control
* low KP and low CR
- --low KP < 60dB SPL
- --low CR is >1:1 to 4:1 (textbook definition)
- ----clinically <2:1
* provides a weak degree of compression over a wide range of input intensities
* "low-level" compression: active with low-moderate level inputs
* "medium-level" or "high-level" compression: active with higher level moderate inputs
36
expansion
* opposite of compression
* reduces the gain for very soft input sounds
* why do we need this?
- --reduces internal mic and amplifier noise
- --reduces low-level environmental noise
* examples of CR for expansion:
- ----1:5
- ----1:2
37
dynamic aspects of compression (2)
attack time
release time
*dynamic aspects of compression determine how fast or slow compression will begin and end
38
Attack Time (AT)
* the time taken for output to stabilize to within 3dB of its final level after the input to the HA has increased above the KP
* basically time taken for the compressor to react to an increase in signal level and reduce gain
39
release time (RT)
* the time taken for the output signal to increase to within 4dB of its final value following a decrease in input level below the KP
* basically the length of time taken for the HA to come out of compression and restore gain
40
what happens when the dynamic aspects of compression are too slow?
* loud sounds may be uncomfortably loud
* compression acts too slowly
* a sudden and loud transient will not be compressed fast enough
41
what happens when the dynamic aspects of compression are too fast?
* gain fluctuates rapidly
* may cause a "pumping" sensation
* may lead to distortion of a signal
42
general time lengths for attack and release time
AT<10ms RT>20ms
| *AT and RTs are set to achieve the best compromise between too slow and too fast
43
effect of AT/RT on speech signal
AT and RT have a major effect on how compression affect the levels of different syllables of speech
44
Rationale for use of compression (5 purposes)
1) avoiding discomfort, distortion, and damage
2) reducing inter-syllabic and inter-phonemic intensity differences
3) reducing difference in long-term level
4 )increasing sound comfort
5) normalizing loudness
45
how compression is used to avoid discomfort, distortion, and damage
*pt's UCL provides upper limit to the HA OSPL90
* output limiting compression:
- --limits discomfort associated with excessively intense signals
- --limits damage to residual hearing caused by excessively intense signals
- --created less distortion than peak clipping
46
numbers for compression to control maximum output
1. Avoiding Discomfort, Distortion, &
Damage
Compression to Control Maximum Output
Output-controlled Compression: AGCo
High CR > 8:1
Short AT < 15 ms
Short or Adaptive RT: 20-100 ms, or adaptive
CT: Set low enough to avoid discomfort
47
adaptive compression
* fixed, short AT
* adaptive RT
- --changes based on duration of incoming sound
- --short or transient input=RT is short
- --longer input= RT is longer
* Results in:
- --protection against brief, excessively loud sounds; no reduction in gain for sounds that follow
- --reduction of "pumping" sensation
48
why we need compression to reduce inter-syllabic and inter-phonemic intensity differences and what type of compression is used?
intense sounds (vowels) can be 30dB more intense than the weakest sounds of speech (unvoiced consonants)
* potential problem for pts with reduced dynamic range
* solution: fast-acting compressor
- ---syllabic compression
49
numbers for compression to reduce inter-syllabic and inter-phonenic intensity differences
Compression to reduce inter-syllabic level differences
*Input-controlled compression: AGCi
Low CR >1.5:1
< 3:1
Short AT: 1 to 10 ms
Short RT: 10 to 50 ms
CT < 50 dB SPL
50
how we use compression to reduce difference in long-term level
* decrease long-term dynamic range without changing intensity difference b/t syllables
* longer AT and RT
* automatic volume control
51
numbers for compression to reduce differences in long-term level
Compression to decrease long-term level differences
*Input-controlled compression: AGCi
Low CR > 1.5:1, but < 4:1
Longer AT > 100 ms
Longer RT > 400 ms
CT < 50 dB SPL
52
average detection for change in level in the environment and decreasing short-term level differences
* adaptive AT and RT
* looks at the average of incoming signal over a given length of time
* when average SPL exceeds KP, compression begins (gain is reduces)
* method of achieving adaptive AT and RT:
- --multiple detectors used to control a compressor
- --dual front end compressor
* results in:
- --protection against brief, excessively loud sounds without rapid fluctuations in gain with high intensity speech
- --reduction of "pumping" sensation
53
adaptive dynamic range optimization (ADRO) for avoiding discomfort and reducing differences in long-term level
applies multiple time constants and several rules in sequential order so that gain is:
1) reduced to avoid MPO exceeding the user's LDL (UCL)
2) reduced to avoid the upper levels of speech exceeding the user's UCL
3) increased to avoid the lower levels of speech being inaudible
4) never allowed to exceed a certain value to help avoid feedback and amplification of background noise
54
how we use compression to increase sound comfort
* OSPL90 set low enough can prevent loudness discomfort, but some pts may not tolerate the signal being close to their discomfort level
* solution: for higher input levels, compression that is more gradual than compression limiting
* comfort controlled compression
55
numbers for compression to increase sound comfort
Compression to increase comfort
Input-controlled compression= AGCi
Low CR > 1.5:1, but < 4:1
Slow AT > 100 ms
Slow RT > 2 sec
CT ~65 dB SPL
56
how we use compression to normalize loudness
principle of loudness normalization: for any input level and frequency, give HA gain needed for the HI pt to report loudness to be the same as a normal hearing pt would report
57
numbers for compression to normalize loudness
Compression to normalize loudness
Input-controlled compression= AGCi
Low CR > 1.5:1, but < 4:1
Slow AT > 100 ms
Slow RT > 2 sec
CT ≤ 50 dB SPL
58
multichannel options of HAs that each compression scheme/rationale can be applied to
* multichannel compression limiting
* multichannel inter-syllabic intensity reduction
* multichannel automatic volume control
* multichannel comfort control
59
the advantage of a compression system or rationale depends on:
* alternative to which it is compared
* goal of amplification (if intelligibility or quality of speech is desired)
* the signal level (ie speech, music, environment)
* presence and type of noise
* SNR
* frequency response shaping used
* HL characteristics of PT
60
default HA fitting should have
* fast-acting compression limiting
* WDRC with fast, slow, or preferably adaptive time constants
* compression ratio appropriate to the degree of HL at each frequency
61
purpose of directional mics
improve SNR
62
what are the three common directional mic configurations
* conventional
- --one mic with two mic ports
* Dual-mic
- --two omni mics
* multi-mic
- --three or more omni mics
63
what is a d-mic
one omni mic with one mic port in addition to one conventional DIR mic with two mic ports for a total of 3 mic ports
64
when do DIR mics turn on (2)
* manual: pt presses a button or switch to engage DIR mics
| * automatic: HA samples environment and determines if engaging DIR mics is appropriate
65
how do we know DIR mics are effective? (3)
methods used to quantify the effectiveness of DIR mics:
* directivity index
* polar plots
* front-to-back ratio
66
directivity index for directional mics
* is a frequency dependent numerical index that describes a device's directivity
* the ratio of sound pressure of the mic from the front to the sound pressure of the mic in a diffuse sound field of the same level
67
fixed arrow for dir mics
polar plot never changes
68
adaptive arrow for dir mics
polar pattern changes based on noise source
69
fitting factors affecting directivity of mics (4)
* orientation of mic ports
* compression
* venting
* directivity degradation
70
how does the orientation of mic ports affect directivity
* need to be as horizontal as possible
| * horizontal location created optimal external delay
71
how does compression affect directivity
* amplitude compression
| * the more compression, the less sensitive dir mics will be
72
how does venting affect directivity
* the larger the vent, the more low freqs are allowed to escape
* this can reduce directivity
73
how does directivity degradation affect directivity
* dual mic systems:
- --must have identical mics
- --mismatch of amplitude and phase can occur
- --the may eliminate directivity or reverse polar patterns
74
consequences of cochlear dead regions
* poorer speech reception thresholds on noise
* distortion
- --can occur at low, mid, and high frequencies and a variety of audiogram configs but most commonly with sloping high freq hearing loss
75
how do we help pts with cochlear dead regions?
move inaudible frequencies to an audible freq region
76
frequency lowering
sometimes called frequency shifting
* frequency information is moved to some other frequency region
* accomplished using frequency transposition or frequency compression
77
frequency transposition
*frequency info is shifted by a constant number of Hz
78
linear frequency transposition
sometimes referred to as linear frequency compression
* output frequency is a constant fraction of input frequency
- --frequencies are shifted by a constant number
79
frequency compression
non-linear frequency compression
| *compresses and shifts frequencies above 1.5kHz
80
fitting frequency transposition/compression
* lower every frequency where hearing threshold exceeds 80dB HL
* may need to implement an aural rehab program during this initial adjustment
* trial for 2 weeks then fine-tune or adjust at next appointment
* always leave signal components below 1500Hz at original frequencies
81
purpose of adaptive noise reduction
improve speech intelligibility by removing noise
82
what other ways do we have to improve SNR?
* DIR mics
| * accessories
83
how do adaptive noise reduction algorithms attempt to remove noise?
*by providing less amplification to noise than to speech and/or UNFINISHED ON HER SLIDE
84
how does the HA detect speech and noise? (3)
* modulation of amplitude
- --HA detects the envelope power
- --speech is likely present when the envelope power in the speech frequencies is greater than the envelope power at other frequencies
* co-modulation
- --when the fluctuations are present in more than one HA channel
* fine structure of speech
- --discrete speech bursts across frequency channels
85
in order to begin the process of adaptive noise reduction, 3 things have to occur:
* estimation of spectral power of noise
* estimation of spectral power of noise + speech
* spectral power of noise is compared to speech+noise to estimate the spectral power of speech alone
86
ways to accomplish adaptive noise reduction
* weiner filter
- --decreases gain as the SNR deteriorates
* spectral subtraction
- --spectral power (amplitude) of noise spectrum is subtracted from the spectral power or speech and noise
87
noise reduction dynamics of adaptive noise reduction (2)
* onset time: the time from when the noise begins to the time when the gain has been reduced to 3 dB of final value
- --can take from 2-30 seconds
* offset time: when the noise stops to the time when gain has been restored to within 3dB of the value in quiet conditions
- --can take from 5ms to several seconds
88
noise reduction dynamics in filters
* wiener filter: slower approach; can vary from ms to seconds
- --intended to react to changes in listening environment
* spectral subtraction: fast acting; 4-8ms
- --reduces noise in frequency regions where noise dominated but can also reduce noise between pauses or words and syllables
89
benefits of adaptive noise reduction (4)
* improve overall SNR
- --however, this does not result in an increase in speech intelligibility
* noise reduction systems are preferred fro:
- --comfort
- --ease of listening
- --quality or overall preference of signal
90
feedback
some of the output of HAs gets back to the input of the HA (fed back to the path)
91
feedback may occur if:
gain from mic to the EC is greater than the attenuation from the EC back to the mic at a specific frequency
92
ways to reduce feedback (3)
* gain-frequency response control
* phase control
* feedback path cancellation
93
gain-frequency response control for feedback
* decrease gain at all freqs
| * or only decrease gain where feedback is occurring
94
narrowband notch filters for feedback?
?
95
phase control for feedback
* phase of the signal is manipulated to reduce the likelihood of feedback
* any freq where gain is enough to cause feedback, the phase response around the loop cause the feedback to be negative instead of positive
96
feedback path cancellation
*a second feedback path is intentionally created
*the internal feedback path has the perfect gain & phase response to cancel
the external feedback path
97
De-reverberation and echo reduction
*reverberation can drastically reduce speech intelligibility of HI and normal hearing pts
*reverberation that does not overlap with speech signals can be recognized by a characteristic drop-off in intensity with time
*once HA recognizes this type of reverberation, i can then reduce the
intensity rapidly
*this then gives the perception of reduced reverberation and increased
speech quality
98
advanced HAs can classify environments as: (4)
* speech-in-quiet
* speech-in-noise
* specific noise types (ie wind)
* music
99
HA classification of environment is based on: (4)
* overall intensity level
* spectral shape
* rate of amplitude modulation
* co-modulation across channels
100
spectral characteristics of wind noise
* f;fluctuates
* high intensity
* low frequency
101
two ways to detect wind noise
* correlation coefficient method
| * power ratio method
102
what happens with HA once wind noise has been detected?
essentially, a low-cut is given to the frequency response to help reduce the intense low frequency noise created by the wind
103
data logging
HAs can now store info about usage:
* listening situations most encountered
* how the pt adjusted HA (volume, programs)
* if any advanced features (noise reduction) were activated by HA)
* daily usage and pattern or usage
