Psychoacoustics ll: Temporal Processing Flashcards
Temporal patterns are
informational substrate which carry information
What are two aspects of temporal processing?
Temporal resolution (more concerned with this)
Temporal integration
What is Temporal resolution?
how OUR SYSTEM IS ABLE to follow the temporal changes, mostly the ENVELOP of the sound
What is Temporal Integration?
Temporal integration involves combining information over time to improve detection or discrimination (sensitivity)
(ex: temporal summation)
What are the two parts of a speech sound?
Temporal envelop and Fine Structure
What does this image show?
The temporal envelope of speech results from speech on and off, in addition to the amplitude changes of pronunciation of vowels, constants
The envelope is seen as an information carrier
What are the frequencies in the temporal envelope in our communication?
Ranges from a few to several hundred Hz in speech
What is the peak of the envelope in the spectrum?
Peaked at 3-4 kHz, corresponding to the normal speed of words/sec. 9 (3-4 articulations per second)
What can we see in the envelope regarding vowels?
Vowels contain Fn (n=0, 1, 2, 3 formats and fundamental frequency (F0)).
Fine structure ~ carriers; Fn ~ modulation fre.
What does the interaction among the formants produce?
Temporal fluctuation
What characteristics can mimic speech/vowels?
amplitude/frequency modulation
What do behavior studies of TR show about within/cross-channel resolution? (2)
The response to signal changes within or cross-frequency channels
For example, gap markers in the same (within) or different (cross-channel) frequency bands
Why do we use gap detection?
Gap detection is a common method to detect TR where the sound is separated by a silent gap
The sound before the gap is called __________________ and the sound after the gap is called _____________________
Pre-gap marker
Post-gap marker
Which between within and cross resolution is closer to real life?
Within-channel resolution
What is cross-channel detection?
The pre-gap marker and post-gap marker are signals in the different frequency channel
What are peripheral vs central limitations seen in behavior Studies of Temporal Resolution? (2)
Limitation from the synaptic transmission, bottom-up
Limitations due to the need of a top-down process (needed to process speech)
How are Simple Estimates of Within-Channel Acuity done for a good estimation of auditory TR? (2)
- Use a single number (index) to indicate TR
- Clicks presented in sequence~ 6ms.
Explain how temporal Resolution is done from click trains.
- Clicks presented in train
- The rate of the train is increased from low to high
- The sense of separated clicks remains up to the rate of 150/s
- Or 6 ms is the approximate of the temporal resolution with click trains.
- Similar result is seen using tone burst of 4 kHz, in which the resolution was evaluated as the minimal intervals between the tones.
What would occur if we use paired clicks instead of click trains?
- In the first pair, the first click is louder/weaker than the second one, order is reversed in the second pair
- Listeners differentiate from detecting temporal order
- Resolution is indicated by the minimal interval upon which the order can be told correctly
- TR down to 2 ms using this method!
What does this graph show related to a method to evaluate TR?
Effect of overall duration on Discrimination (the second method to evaluate TR but not popular)
- The same signal but you change the duration
- You ask the subject to tell you which signal is longer or shorter
- From this data, we can see that Weber’s fraction should be constant/followed
No difference related to bandwidth of signals
What is the Gap Detection ability?
Ability to identify a (silence) gap between two sounds or a drop in level/interruption of a sound by varied formats
What is the Gap threshold defined as?
Gap threshold Δt is defined as the minimum period of the gap that can be identified. Below that Δt, the sound is perceived as continuous
How can the Gap be like?
A silent period or one in which sound intensity is largely reduced
Gap detection can be measured in ______________________ test or ____________________test like______________________.
Behavior test, or in an objective test like evoked potential
How does the identification of the Gap threshold work in the behavior test?
By detecting the signal in response to the off-set of the pre-gap marker and onset of the post-gap marker
What can we use as Gap markers? (2)
Broadband noise and narrow band signals
Since gap markers can be broadband noise and narrow-band signals, what does it tell us about our ability to test gap detection?
Gap markers show that we can test gap detection in different frequency regions
What is the issue with using narrow-band signals as a gap marker?
Contamination of Frequency cues when using narrow band signal and sudden on/off
How can we overcome the frequency cue contaminations from narrow-band gap markers? (3)
- Masking with Notch noise
- Bandpass filter to get rid of contamination (to modify signal)
- Ramping but will make the gap not clear in duration
What does this graph show related to how we sense the gap between two sounds?
How Gap detection is based upon sensation change (degree of excitation) due to the onset and offset of signal
Pre- and post-gap markers can be different in terms of amplitude, duration and frequency.
For the gap detection test:
If different in frequency, then test______________
If both are the same in frequency, then test ____________
If different in frequency, then tests cross-channel
If both are the same in frequency, within channel test.
What is the effect of intensity on the gap marker?
Causes equal marker intensity
The impact of sound level is seen near/just above the hearing threshold and is not changed by sound level well above the threshold
What is the effect of gap marker bandwidth?
The gap is marked with equal sounds and markers are broadband signals and narrow signals resulting in a better or smaller gap detection threshold with a broader marker bandwidth.
What is the effect of bandwidth on thresholds?
Broader bandwidth causes smaller gap thresholds, our system can integrate information across broad frequency regions to improve resolution
What is the impact of hearing loss on the gap thresholds?
High-frequency HL: deteriorated gap threshold (goes down) when using broadband markers which are attributed to reduced audibility, since oftentimes SNHL occurs at high frequencies
High-frequency channels/regions have a better temporal resolution since the traveling wave occurs quicker at higher frequency region (natural advantage quickly change the signal)
What is the evidence of the impact of hearing loss on gap threshold?
High-pass filter just masks high-frequency region to create artificial hearing loss
When high-pass masking is used in normal hearing subjects, will have a similar change of deteriorated gap thresholds to SNHL.
If you compare artificial hearing loss to natural hearing loss you will not see the difference in gap thresholds to SNHL since they both become deteriorated.
High-frequency channels have a better temporal resolution
What are three settings gap detection can be seen?
(a): within the channel (same frequency band of pre and post-gap markers)
(b): Cross/between channels (different frequencies between the pre and post-gap markers)
(c): diff in onset/duration discrimination (must be “between channels”)
How does the detection of Sinusoidally Amplitude Modulated Noise work? (2)
Modulation depth can be represented in
% modulation:
average/peak to peak%;
or peak to trough /average%
percentage can be converted into dB: 20log(%), e.g., 10%~ 20log(0.1) = -20 dB (IMPORTANT)
Minimal depth of modulation is the minimal depth you identify the sound as modulated (usually smaller ar low pass)
Detection threshold with modulation frequency or Modulation transfer function (MTF)
What does this graph show?
The detection of Sinusoidally Amplitude Modulated Noise
Modulation signals
a) Time waves of modulated and unmodulated noise
b) Modulation with different MF (high or low frequency).
c) Modulation with different depth
What does this graph show related to the modulation transfer function?
In normal hearing subjects, MTF is typically a low-pass function: a larger modulation depth threshold at a larger modulation frequency (MF)
In subjects, with temporal processing deficits, a larger modulation depth threshold is seen at high MF and goes down quicker with frequency, thus they have more difficulty following high-frequency modulation or quicker change in signal
Normally, for higher Modulation frequency, the _______________________________________________________
Higher the Modulation frequency, higher the depths for us to detect sounds as modulated
How is temporal processing done in the cochlea? (2)
Phase locking or synchronization
Envelope coding
What methods are used to demonstrate envelope coding? (3)
PSTH
ISIH
PRH
Which principle describes the real neuronal envelope coding?
Volley principle
Which method is seen on this graph?
ISIH - Inter stimuli Histogram
Divided into equal-time bins
When PL is perfect at low frequencies, AN will produce 1 spike per period at a certain phase and clearer interval periodicity
When the Frequency is too high ANF fire cannot follow so it skips the period. Periods become integer numbers, less and less until eventually become only one
What does this graph show?
The Synchronization in AN firing
Increasing sound level causes better phase locking
The distribution of ANF becomes smaller and smaller in phase locking
With better phase locking, the distribution of synchronization of ANF becomes_______________
narrower
Which method is shown related to spikes of ANF?
PRH
* The x-axis is the time of the, across several cycles of the simple tone or the envelope of the complex tone.
- The complex is the combination of two tones and phase locking follows the level of combined tones
- The difference in phase resulted in different envelopes.
Give a quick summary of how auditory nerves do temporal coding. (3)
- ANs use phase locking to follow the envelope of sound/encode the Temporal envelope
- Phase relationship between Basilar membrane vibration and NT release
- Phase-locking is established by the integration of responses from many neurons.
What does this graph show?
The modulation Transfer function of SINGLE Neurons becomes bandpass by synchronization strength or by rate change
What is the take-home message of these graphs related to IC and best MF?
MTFs of IC single neurons all have different best MF (peak) meaning that they can each best detect the sound fluctuations at a certain frequency
The majority of neurons of IC have the best MF between 30-100Hz
Take home message:
Behavior MTF is _______________
Single-neuron MTF is ______________
Behavior MTF is low pass
Single-neuron MTF is bandpass
What can we see on this image related to the Place code for best modulation frequency?
Concentric (cone) distribution of neurons in IC with the same BMF for the temporal processing
- The neurons on the surface of the cone show the same BMF
- The iso-BMF contour surface is in a cone shape taping to the dorsal side (low frequency)
Place code is not only related to frequency processing but other features!
What does this graph show to temporal coding in MGB and AC?
Neuronal response can follow the envelope of the signals they show the peaks
What does this graph show related to the masking of pure tone with white noise?
Increased masking threshold with increased frequency/ increase frequency x10 causes an increase of masking threshold of 10 dB
White noise masking is more effective at
higher frequencies
What are important informations we can see on this graph related to masking increase?
Signal to noise ratio changes with the frequency
For a higher frequency signal, we need a larger SNR to hear a sound in noise
Higher masking effect.
Remember the spectrum of the noise is flat.
Definition of White noise in terms of frequency:
White noise has _______________________________
White noise has equal density across frequencies
What is important information about sound density and white noise?
Density = total sound power/bandwidth
Total power in a frequency range = density*bandwith
The energy/power effective for masking exists in the critical band:
P = density*CB
What is a critical band?
CB: The frequency spectrum that one neuron will respond to in the signal
Only the energy in a certain band around the frequency of this signal impacts the hearing of this signal.
In a broad-band masker, ____________________________ will produce masking.
In a broad-band masker, only the energy in CB will produce masking.
What can we see on the cochlea related to the width of the critical band?
The width of the Critical band changes with CF, wider CB with higher CF which is why there is a higher masking effect at higher frequency regions.
(Below 500 Hz, the Effective masking won’t change much but at higher frequencies, it gets more effective)
An increase of CF causes _________________ and an ________________________.
An increase in CF causes a wider of CB and an increase in effective masking
What are considerations we should know related the bandwidth of effective masker? (2)
- Only energy within CB is effective
- CB increases with CF in linear scale
CB increases with CF linearly with a constant ratio scale of________________________________
CB increases with CF linearly with a constant ratio scale of 20% or 1/3 octave around CF.
In clinic, the masker for pure tone should be ____________________________ to ________________________
In the clinic, the masker for pure tone should be narrowband noise of 1/3 octave to reduce the total level of masking.
What is the relationship between white noise, CF, and the effect of masker level?
A larger Signal to noise ratio is needed to be heard in noise since when white noise is used, the effective masker level increases with CF since CB gets wider.
How can we measure CB with masking?
We keep the total intensity of the masker the same
Increase the bandwidth of the masker from zero to very broad
The turning point tells CB
Within CB, the masked threshold will not change. Beyond CB, the masked threshold will decrease, because some masking energy got into other channels so that is not effective. (becomes useless)
What can we do to keep the same masking level with masker BW change to maintain masking?
Increase masking level with BW increase
Within CB: masker level should not be changed.
The energy inside CB matter
How does sensitivity change with the bandwidth change of signal?
Test hearing threshold while increasing signal bandwidth within CB
The threshold will not change
If the signal frequency range is beyond CB: ________________________________
if we don’t change the total sound level, within the CB phase-out so there won’t be enough energy to create a neuronal response
What is Co-modulation masking release?
Comodulation Masking Release (CMR) is the decrease in masked thresholds that occurs when the masker is amplitude-modulated
Separate frequency with different bands, signal and flanking band. Presenting the noise in the flanking band doesn’t influence the masking, but if presented in the signal band it will produce masking. Adding flanking bands does not change the masked threshold because they are far away from CB.
On-signal band alone
Comodulated—see release or decrease in masked threshold
Uncomodulated—no change in masked T.
What does this graph show related to co-modulation masking release?
Adding of more maskers in flanking band reduces masking, if co-modulated
What is masking overshoot?
Overshoot: Masking effect from onset to other points of signal
- Masking effect depends on the time relationship for signal in masker
- Larger masking when the signal is close to the onset of masker, moving away from onset will reduce masking effect until plateau
- Up to 10-15 dB
- Disappeared when delay (onset of masker-onset of signal)> 200 ms,
Masking overshoot is the difference between _____________________________________
the masking effect at onset of sound until plateau
What are the 3 temporal masking types?
Backward Masking after the signal (signal is at the onset of the masker)
Foward Masking before the signal
Combined Backward and Foward Masking
Masking can be done ___________________ in one ear or _____________ both ears.
Monotic S and N (Peripheral masking)
Dichotic (S in other, shows central masking because no effect of masking interaction in cochlea)
In real life, we hear sounds (monotic/dichotic/diotic) _________________________________ and shows _______________.
diotic S and N go to both ears and shows peripheral and central masking
What does this graph show?
Larger masking occurs at the onset of backward masking (central masking) and the offset of forward masking
What does this graph show?
Foward masking in dB
Linear relationship
Forward masking is ___________________ and backward masking ____________________________________
Forward masking is relatively clear and backward masking which the central masking role is unclear
What are the characteristics that show how forward masking is relatively clear? (3)
overlap in BM vibration,
neural adaptation,
Central masking (indicated by cochlear implant)
Central masking is done by presenting the masking to _______________________
the contralateral ear
What is the difference in threshold shift in central vs. peripheral masking?
Much smaller threshold shift in central masking which the maximum is only 15 dB (much smaller amplitude)
What are 2 similarities between Central and peripheral masking?
Frequency relationship
the masking effect and time-relationship between masker and signal
What does this graph show related to the effect of the masker’s frequency and presentation?
Bigger masking effect if the signal and masker are on/off together impacts the low frequencies as well.
What is informational masking?
- Interaction between masker and signal at a higher level of the auditory pathway
with a maximum masking threshold of 30 dB which is opposite to peripheral masking (in the cochlea which is also called energetic masking)
How is informational masking done? (3)
- Targeted tone (or speech) in the presence of multi-frequency masker (similarity and uncertainty impacts performance)
- Test masked threshold in 2IFC
- Frequencies of the masker randomized
- CB around the targeted signal is “protected”—to avoid energetic masking
What does this graph show related to informational masking?
Informational masking in 2IFC
- No masker in the CB of the targeted tone.
- (a) and (b): few masker components, (c) and (d): more masker components.
Informational masking by uncertainty
The effect due to randomization
Larger the randomization, the larger the effect
The difference becomes smaller with an increasing number of components in the masker
In reality, energetic and informational types of masking are presented together, we need to verify which is which.
Why do we need to know the masking and the mechanisms? (3)
To understand how masking changes our hearing.
To use masking as research tools.
Notice the gaps between what we have discussed and what we need for signal detection under masking.
How do we detect signals in noisy backgrounds? (4)
Spatial Filtering
Spectral Filtering/Frequency selectivity
Temporal Filtering
Cognitive Processing
How do we detect signals in noisy background using spatial filtering?
Detect signals by differentiating the source from noise—depending on binaural process.
How do we detect signals in noisy background using Spectral filtering/frequency selectivity?
Selectively filters out noise—but won’t work if the spectrum of noise is largely overlapped with that of signals.
How do we detect signals in noisy backgrounds using temporal filtering?
Distinguish signals based upon the time difference, e.g., signals in the trough of noise.
How do we detect signals in noisy backgrounds using cognitive processing?
Detect signals by using experiences (familiarity to the signals)—depending on top-down process. This is shown in part of “attentional filtering”.
What are the two identification settings of informational masking? (2_
The informational masking of 30 dB is typically larger than the effect of central masking (in dichotic presentation)
(1) the CB of signal (4kHz) is invaded by masker (solid symbols)-energetic masking
(2) CB is not invaded (open symbols, smaller masking)—informational masking
What are the neuro-mechanisms that contribute to signal detection in noise? (6)
- Binaural processing related to temporal processing, inhibition, efferent etc; to spatial filtering.
- Inhibition to spectral filtering and other processes.
- Mechanisms of high temporal resolution
- Low-SR ANFs and their efferent control.
- Interaction between ascending and descending pathways.
- Cochlear efferent control
What/how the neuro-mechanisms contributing to signal detection in noise are changed in subjects with HL and in old age?
Loss of temporal resolution
Which between within and cross resolution is closer to real life?
Within-channel resolution
Why do we use gap detection?
Gap detection is a common method to detect TR where the sound is separated by a silent gap