Hörselorganens fysiologi Mattias Heldner heldner@kth.se Rekommendation Repetera Engstrand kapitel 10 om hörselsystemet. Betydligt mer lättillgänglig än Moore... Introduktion Hertzskalan ingen bra skala för hur tonhöjd uppfattas för röstens grundton Tryckenenheten Pascal ingen bra skala för hur stark rösten uppfattas Linjära skalor inte alltid vad man vill ha... Hur är det med Hertzskalan för resten av spektrum (formanter osv)? Hertz vs Bark Samband mellan plats på basilarmembranet och uppfattad tonhöjd Örats frekvensupplösning är tonotopiskt organiserad Ett givet frekvensintervall (i Hz) motsvaras av olika stora avstånd längs basilarmembranet
Hertz vs Bark Relation mellan frekvens (i Hz) och uppfattad tonhöjd olinjär Bark = enhet som korresponderar mot frekvensens tonotopiska representation i innerörat plats längs basilarmembranet Hertz vs Bark F2 [i] F3 [i] F4 [i] F1 [i] Hertz vs Bark Örats frekvensupplösning försämras med stigande frekvens På höga frekvenser krävs det större förändring av frekvensen för att den ska uppfattas Formanter som ligger tillräckligt nära varandra på Barkskalan integreras av örat och uppfattas som ett sammanhängande energiområde Hertz vs Bark Vanliga spektrogram och F0-kurvor plottade på Hertzskala ger inte korrekt bild av hur frekvensrelationer uppfattas Bark relevant t ex för formantmätningar Det går att göra spektrogram på Barkskala i Praat... Det går att transformera Hz till Bark http://www.ling.su.se/staff/hartmut/umrechnung.htm
Introduktion Hörselorganen gör en spektrografisk analys av auditiva stimuli Cochlea fungerar som en filterbank Hörselorganen gör även en oscillografisk analys av auditiva stimuli upp till 1500 Hz - phase locking som också bidrar till frekvensupplösningen
Organ of Corti
Inner and outer hair cells Inner haircells
Firing rate Phase locking
Psychophysical Measurement Frequency resolution
Ratio pitch (in mel) Experiments with sinusoids Subjects were asked to either divide given frequency ranges into perceptually equal intervals, or to adjust the frequency of a tone to be half as high as that of a tone given for comparison 1 mel = 1/1000 of the pitch of a 1 khz tone Critical bandwidth Critical bandwidth Bc is a measure of tonotopic resolution in audition Loudness summation experiments Different summation rules depending on whether frequency components are separated by more or less than the CB Critical bandwidth Loudness summation If the frequencies of the tones lie within the critical bandwidth, the loudness is calculated from the total intensity: S = I = I1+I2+I3+ If the bandwidth exceeds the critical bandwidth. I1+I2+I3+ < S < S1+S2+S3+ Noise with center frequency 1000Hz (930Hz-1075Hz). Bandwidth 145Hz. Critical bandwidth A measure of tonotopic resolution in audition Can be considered a measure of tonotopic position that is useful in models of hearing However, auditory frequency resolution cannot be represented on the basis of CB alone as the hearing system also performs a temporal analysis that contributes to frequency resolution for low frequencies
Equivalent rectangular bandwidth rate (ERB) Notch-noise method detection threshold for a sinusoid, centered in a spectral notch of a noise, as a function of the width of the notch Auditory frequency selectivity can be described in terms of an "equivalent rectangular bandwidth" (ERB) as a function of center frequency ERBs by Masking Tone 2000Hz. Broadband noise, 1000Hz, 250Hz, 10Hz. The critical band at 2000Hz has a band width of 280Hz 2000 Hz 1.2000Hz tone, 10 decreasing steps of 5 db 2.Broadband noise 3.Noise, bandwidth 1000 Hz 4.Noise, bandwidth 250 Hz 5.Noise, bandwidth 10 Hz Cb vs ERBs Proportional for center-frequencies above 500 Hz For lower frequencies, the ERB decreases with decreasing center-frequency, while the CB remains close to constant Temporal fine structure of the signal is not resolved in loudness summation, while it contributes substantially to frequency resolution for f < 500 Hz Psychophysical Tuning Curves (PTC)
Estimates of auditory bandwidths Frequency Selectively The ability to judge the pitch of tones is much better than bandwidth of the critical bands. 0.1% differences in frequency of pure tones. Analyzing the time pattern of neural firings higher up in the hearing system Temporal Integration Short sounds and longer sounds at the same sound pressure level are perceived differently The ear average sound energy over about 200ms Burst of broadband noise Durations: 1000, 300, 100, 30, 10, 3, 1 ms Levels: 0, -16, -20, -24, -28, -32, -36, -40 db Sound Localization Time differences Level differences Time differences as function of angle Level differences as function of angle 2 khz tone Speech
Sound Localization Vertical Plane Conductive hearing loss Deteriorated impedance conversion between the eardrum and the oval window Abnormalities at the eardrum, wax in the ear canal, injuries to the ossicles, inflammation in the middle ear. Sometimes possible to recover with surgery Sensorineural hearing loss Damage to the inner and outer hair cells Acoustic trauma, drugs, infection, congenital Usually permanent Sensorineural hearing loss
Hörselskadan Hörstyrka Ljudnivå Hörstyrka db SPL db SPL Aj!! Aj!! Hörtröskeln ökar Obehagsnivån är oförändrad Dynamiken minskar 120 120 Obehagligt! Obehagligt! väldigt starkt 100 100 väldigt starkt starkt starkt lite starkt precis lagom lite starkt 80 80 ganska svagt svagt precis lagom knappt hörbart 60 60 Hörtröskel? ganska svagt? 40 40 svagt? 50 db HL? knappt hörbart 20 20? Hörtröskel?? 0 0 normalhörande hörselskadad Hearing loss simulations