Popular in Course
Popular in Psychlogy
This 26 page Class Notes was uploaded by Marco Wolf on Monday September 7, 2015. The Class Notes belongs to PSY 323 at University of Texas at Austin taught by Staff in Fall. Since its upload, it has received 42 views. For similar materials see /class/181815/psy-323-university-of-texas-at-austin in Psychlogy at University of Texas at Austin.
Reviews for PERCEPTION
Report this Material
What is Karma?
Karma is the currency of StudySoup.
Date Created: 09/07/15
Human Elephant 1 100 VIIIIII 1000 Frequency HZ Mouse 1 lllllll I 10000 l I III 100000 Frequency ranges for different organisms Auditory detection thresholds audibility curves for differe 80 Dolphin Elephant l Dolphin Mouse 3 l I Do I 60 9 I I A i I I 3 I l I m 40 Elephantquot I I 3 39 l l U I l B l 5 I l 3 20 I l E r I l l 0 7 Mouse I J 20 o l l I 1 100 1000 10000 100000 1 Frequency Hz 200000 nt species S ont Low 00 Em Medium X Nerve fibres MCL 96 9 High 0 Abs threshold 80 y E v u 8 60 O 1 0 U m 40 U E 0 20 E O 20 Tlllllll I ljilllll ill Ol lO lo 40 Frequency kHz Best thresholds for 8th nerve fibers in the cat The solid line is behavior detection data The animal s behavior approximately tracks the sensitivity of the population of high spontaneous 8th nerve neurons I I I I I Threshold 120 t Of feeling 100 80 q 80 ugt loudness 3 curves Conversational E 60 speech U 40 A0 20 Audibility u u u u u u u 7 curve 0 g1 threshold I I I l I of hearing 20 100 500 1000 5000 10000 Frequency Hz The audibility curve and equal loudness contours for humans The region in gray shows the range of frequencies and intensities for typical speech This is a particularly important range of stimuli to hear ah 50 oo ay ee lt Vowels Sound Intensity dBSPL 4O mnng dt sfth lt Nasals Stops Frlcatives a 30 lt Consonants 20 I llllllll I liillill 100 300 1000 3000 10000 Frequency Hz Here are the frequencies and intensities associated with speech sounds The consonants are less intense and higher in frequency especially the fricatives Prevalence in population 17 30 31 40 41 50 51 60 61 70 Age years Hearing loss grows with age Much of this loss is the cumulative effect of exposure to intense sound 100 80 60 40 Sound Intensity stpL 20 0 I IIIIILI I l Illlll IIIIJ 20 100 500 1000 5000 10000 Frequency Hz Typical audiogram of an individual with a 50db hearing loss Notice that this level of loss is cutting into the region needed for speech perception A cochlear implant is a device that can restore some hearing in most individuals with total hearing loss Typically about 16 pairs of electrodes are placed at various locations along the cochlea BANDF ASS NOISE BANDPASS NOISE a 39able Wid h BROADBAND Constant Width Constant Center Frequency NOISE Variable Cenier Frequency gt gt cm or a E c LLI LU Frequency Frequency gt gt gt m m a m Lu Norse w l Frequency Frequency Frequency gt gt 7 C7 3 E 5 Lu LLI Frequency demo noise types Frequency Types of noise Narrow bandwldm vng bandwidth t E E a w n z E m m 71 Hz Freuuency Frequency Large 5 g 2 g I g t None ND noise Narrow Wide Ursa WWquot demo critical band experiment E a w c 5quot E g 2 E m ellci aural respensa Minimum saun Measuring critical bands in the classical way Fletcher method lt Af gt lt Af gt Noise Auditory filter Power linear scale Frequency linear scale R D Patterson s method for measuring critical bands The solid curve represents the auditory filter centered on a particular frequency The tuning characteristic functions measured in 8th nerve neurons can be thought of as the inverse of a tuning function In other words flip the threshold tuning function in the previous slide over to get a filter function d 5 D D D I l I Relative response dB I 1 D i I 5 I I l l I I l l l 04 05 05 07 09 09 10 11 12 13 14 15 16 Frequency kHz Critical band tuning function auditory lter measured using Patterson s method 1 0 I I I I I I I I l I I I I I g 0 5 x p c39 i 2 02 U D L s 39 2 01 I I gt 5 4 3913 U 05 D lt D I I l I I l I I I I I l I I l I I 0 1 0 2 0 5 1 2 5 10 Centre Frequency kHz Auditory filter bandwidth as a function of center frequency Solid curve is based on the Patterson method Notice that the critical band is about 110th of the center frequency I BehavlounICI N ale Behavlounl BSN iirvr 39 Cochlear nerve bra lllllll Equivalent Rectangular Bandwidth kHz 0391 I I I I l I I I I I I I I I I l I I l I I 1 10 100 Characteristic frequency kHz Comparison of critical bandwidths measured in behavioral experiments with those measured on single cochlear nerve fibers D I I I I I I l I I I an 10 U 6 m 8 20 D 0 OJ 90 L 2 30 U C m 40 20 20 90 1 I I l I I l 50 m4 m5 m6 m7 as as L0 L1 L2 L3 L4 L5 L6 Frequency kHz Auditory filter as a function of sound intensity level measured with the Patterson method Note that the width increases as the noise level goes up Why should this happern Time gt The function of the outer hair cells is largely mechanical they increase the resonance magnitude of the basiliar membrane at low sound levels The inner hair cells are the real receptors They transmit sound information into the nervous system Approximately 20 8th nerve neurons connect to each inner hair cell The behavior of these neurons depends on where they connect to the inner hair cell lt Saturation r High 39 quot ii39W i Low Strength of neural response 10 20 30 40 50 60 70 80 90100 Sound intensity dBSPL lt Saturation gt LOW Strength of neural response 10 20 30 40 50 60 70 80 90100 Sound intensity dBSFL One reason is that the responses of 8th nerve neurons saturate at high intensities Responses of 8th nerve fibers as a function of sound intensity Top is a high spontaneous rate neuron bottom is low spontaneous rate neuron I50 39 O gt c 39O A 100 39 o g 0 0quoto E 80 3 o 70 E 50 63 53 O o a 30 OZ39l91 o lll O 3 LS 27 Frequency kHz Frequency response functions of 8th nerve fibers as a function of sound intensity i a m m n w m b 2a in 30 o m an so u o v 239 A V CD a D Cl 3 m E m a n m J N a m n a a m 2 a 21 391 a 395 i c g 7 1 quot I FREQUENCY a m u m b C d How the vocal tract produces three vowel sounds a Pulses from the vocal cords b con gurations of the vocal tract c Resonance spectra of the vocal tract configurations d Spectra emitted from the vocal tract Frequency kHz 05 08 10 seconds Example speech spectrogram 3000 2400 1800 1200 800 O P F di du Frequency Hz Time gt Synthetic spectrograms of the first two formants of two syllables di and du An adaptive problem faced by users at a spoken language How to maintain suf ciently high intelligibility to ensure that meanings are transmitted even under unfavorable listening conditions eg environmental noise low redundancy and possible hearing impairment A possible solution Maximize auditory distinctiveness of linguistically contrasting sounds or phonemes From acoustic to auditory representation INPUT Harmonic spectrum OUTPUT Excitation pattern Model of the auditory representation of a French vowel y ewe 1 vowel spectrum gt 5 approximate loudness spectrum Psychoacoustic basis of perceptual distance amp optimal system Optimal system of p vowels 22lDij2minimized Distance Dij related to the area enclosed by the two curves The distinctiveness of a pair of vowels is defined as the sum of the squared difference between their auditory model outputs The optimum vowel system is the one the maximizes the distinctiveness over all pairings of vowels bV dV gV 2500 l m 9 l l 1500 F2 Hz gt 1000 L39 500 200 400 600 800 F1 Hz The optimal vowel placements predicted for languages that use ve vowels This closely matches the vowels in Spanish Hawaiian Japanese and other fivevowel languages bV dV gV 2000 T The optimal vowel placements predicted for languages that use seven vowels This closely matches the vowels in Italian and other sevenvowel languages