Lecture 14 Flashcards
What are the localization coordinates?
Horizontal plane - parallel to the ground; behind and head (goes through your nose); degrees of azimuth
Median plane - perpendicular to the ground if upright; degrees of elevation
What are the two binaural cures for localization?
Interaural time differences, interaural level differences.
What is the phenomenon when sound waves travel from the right have to travel further to the left ear?
When is ITD the largest?
Time it takes for wave to travel extra distance, is called interaural time difference.
Largest ITD when: sound source is directly to the right or the left.
For low frequency sounds, what is the most important binaural cue?
ITD is most important for localization on binaural plane.
What happens to ITD at frequencies higher than 750Hz?
What is the waveform phase when the sound is directly to the R/L of head?
Becomes ambiguous for frequencies higher than 750Hz - period of waveform is 2x max ITD of 0.65 msec = waveform will have phase difference of 180 degrees, 1/2 cycle when sound is at L/R of head
Why are most sounds correctly localized?
Because they have both low and high frequency components
What happens when the modulation rate is lower than 700Hz for ITD cues?
The brain can use that AM to de-ambiguate the sound. Even though carrier might have high frequency.
What is the result of presenting a pure tone at the same frequency is amplitude modulated at a low frequency rate?
Because the AM right ear waveform is leading - the unambiguous ITD cue. The sound is correctly localized to the right of the head.
Brain is able to use slower AM by comparing the phase of that modulation between the two ears.
Why does phase/ITD become ambiguous as frequency increases?
If a waveform is delayed by 180 degrees, can’t tell which ear is leading. Why:
Due to time it takes for sound to travel from one ear to the other 0.6ms. Also depends on the size of the head - and sounds travel at a constant speed.
For localization in azimuth for ILD, what is ILD mainly due to and why?
What does this have to do with frequencies?
Mainly due to head’s sound shadow, because there is more attenuation at high frequencies (because wavelength is small), because sound shadow effect depends on wavelength.
Frequency components whose wavelength is long relative to size of hear, are able to bend around head - attenuated very little.
What happens to ILD differences as azimuth angle is varied?
ILD difference can be relatively large at high frequency (up to 20dB at 5-6kHz) and relatively small at lower frequency (less than 5dB at 500Hz).
What is the duplex theory of localization?
Low frequencies localized through IPD/ITD and high frequencies on basis of ILD (sound shadow not effective at low frequencies; cannot rely on ILD).
Errors between 1000-2000Hz - neither ILD/ITD can provide reliable info about localization on horizontal plane.
In the experiment for differential thresholds for localization - minimal audible threshold, what happens?
What is minimal audible threshold? (Hz)
Listeners sit in middle of array of speakers at various angles of azimuth. Pure tone first presented through one loudspeaker; then same pure tone presented to left or right of first one.
When smallest difference in azimuth is measured for different frequencies, results show that minimal audible angle is largest around 3000Hz.
Why is wideband noise accurately localized?
When you have speech, you get info from low and high frequencies and you have ability to use both cues at the same time - compared to pure tones. Localization on horizontal plane is accurate, if the two agree then you can expect to see data in a linear fashion.
How is localization on the horizontal plane accomplished?
Exploiting interaural time/phase and level cues.
When are ITD/ILD most effective? (Frequencies)
ITD: lower frequencies, about 1500Hz
ILD: higher frequencies (500-1000Hz)
Where does neural processing of ITD and ILD occur?
Likely in medial and lateral superior olive, respectively
When are ITD and ILD equal?
Sounds on a medium plane - binaural cues are useless.
What is the cone of confusion?
All locations on the surface of the cone have same ITD and ILD - because if you have sounds on the surface of concentric circles and ear entrance has circles too, each point has the same distance to the ear.
You can’t discriminate where on the cone it is because sounds will be the same.
When is the cone of confusion the longest? ITD
Directly to the right, that point is a cone in a very narrow circumference. As cone broadens, it will come close to the cone that passes directly around the head (medium plane), so ITD will almost be 0. Shorter ITD in the larger cone.
What occurs when you have head movement in the cone of confusion?
What happens when sound source is at the front/back? Above/below?
Head movement to the right will produce delay in sound arriving at the right ear. Sound source is in the front, same head movement will produce a delay at the LE, if sound source is directly behind.
If sound is directly above or below, ITD/ILD will not change
How does the shape of the pinna contribute to sound source localization on the median plane?
Spectral shape determines if sound is front or behind - high frequency energy typically more attenuated if sound source behind listener because of asymmetric shape of pinna.
Changes from filtering of pinna are dependent on location of sound source.
Resonances of cavity of pinna cause direction dependent changes in spectra of sounds entering ear canal. Happen at frequencies higher than about 4000Hz because of dimensions and cavities on pinna.
What happens to people with high frequency hearing loss for sound localization on the median plane?
Filtering properties obtained from cues to pinna to head are all about 4000Hz, so will have trouble localizing at this frequency. If you repeatedly measure on median plane, would observe abnormally large minimal audible angles for elevation for HL listeners.
In sound localization in the presence of reflected sounds, what happens to the sound? (relative to the ear, when it arrives)
Reflected sounds act as repeated copies of original sound. Reach ears with some delay relative to direct sound; arrive from different location than direct sound.
Direct sound arrives before reflected sounds, auditory system uses strategy that direct sound must be one that is most likely correct info about location - results in precedence effect.
In the precedence effect, what is the result of a delay of greater than 300ms?
Echo
In the precedence effect, what will be the perceived location if the delay is = 20ms and test signal is speech?
Delay 0 to 0.5-1.0ms
If signal from right speaker is delayed, up to 1ms to left speaker, location of sound image moves toward the right (signal played first)
In the precedence effect, what will be the perceived location if the delay is = 0.5ms and test signal is speech?
Delay 1-30ms
Sound image appears to come from the left/right speaker only. Toward the sound played first, away from the one that is delayed. FROM THE LISTENERS POV.
In the precedence effect, what will be the perceived location if the delay is = >30ms and test signal is speech?
Listener hears two sounds, first from the right speaker, second (less salient) from the left speakers.
In the precedence effect, what does the 30ms interval apply to? What does 5ms?
30ms interval applies to ongoing speech, transient sounds (clicks) interval for occurence of precedence effect is shorter - 5ms.
What is the cue that is suppressed only in precedence effect?
Location of later signal
In the precedence effect, what is the effect of level/time?
Reflected sounds usually lower in level than direct sound
Interval that generates precedence effect is longer.
With relative level in dB is higher = shorter delay (ms); relative level dB lower = longer delay (ms)
What is the precedence effect a result of?
Result of learning - when placed in unfamiliar environment, listeners are aware of reflected sounds, but after brief interval of adaptation, attention is focused on the first arrival sound = precedence effect.
If reflection pattern of spectral patterns changed, listened has to re-learn to ignore reflections again. Can be trained to ignore precedence effect hear delayed sounds clearly.
How is the perceived location of sound determined?
Determined by the location of the direct sound (first wavefront).
What time frames create precedence effect?
Delays about 1ms or less don’t generate precedence effect. Longer delays generate the perception of separate sound sources: delayed sound heard as an echo.
But direct sounds and corresponding reflected sounds that are delayed by a maximum of 5ms (clicks) or 30-40ms (complex sounds such as speech) are fused into a single sound.
What needs to happen for first and delayed sounds in the precedence effect?
Have to have similar spectra/temporal envelopes for precedence effect to work in free field conditions.
Lateralization through headphones, how does this work - what are delta-interaural phase difference thresholds?
Sounds heard within the head (lateralized) towards ear with leading phase and/or higher level.
DIPD increase above about 900Hz - size of interaural level difference someone can discriminate
Listeners require larger DIPD differences as sound is lateralized away from the centre (0 degrees)
DILD difference thresholds increase as base ILD level increases, but do not change much across frequency.
Where is ITD/ILD processing occurring in the brain for mammals?
ITD: medial superior olive
ILD: lateral superior olive
