Seneff’s Auditory Model

Report
Seneff’s Auditory Model
Miriam Cordero Ruiz
(SONY Advanced Technology Center Stuttgart)
Leuven, july 2002
Which is the best speech
recognizer?
Introduction
•Auditory System
•Seneff’s Model
•Stage I
•Stage II
•Conclusions
Human Auditory System
Human Auditory System
Human Auditory System
band
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
fc(Hz)
50
150
250
350
450
570
700
840
1000
1170
1370
1600
1850
2150
2500
2900
3400
4000
4800
5800
7000
8500
10500
13500
BW(Hz)
80
100
100
100
110
120
140
150
160
190
210
240
280
320
380
450
550
700
900
1100
1300
1800
2500
3500
Human Auditory System
Inner Hair Cells
t
t
Structure of the model
CRITICAL
BAND FILTER
BANK
STAGE I
HAIR CELL
SYNAPSE
MODEL
STAGE II
ENVELOPE
DETECTOR
Mean rate
spectrum
SYNCHRONY
DETECTOR
synchrony
spectrum
STAGE III
Stage I:
Auditory Filter Bank
40 channels (20 - 6700 Hz)
BW1channel=0,5 Barks
Design of the Auditory Filter Bank
f(Hz)
INITIAL
COMPLEX
ZEROES
ZERO OF
CASCADE
ZERO OF
CASCADE
RESONATOR
RESONATOR
RESONATOR
CHANNEL 1
CHANNEL 2
CHANNEL 40
…….
ZERO OF
CASCADE
Stage II
Model
Physiological Data
< 1kHz
STAGE I
Stages I+II
CRITICAL BAND
FILTER BANK
STAGE II
HALFWAVE
RECTIFICATION
SHORT-TERM
ADAPTATION
LOW PASS FILTER
RAPID AGC
Results
Other Peripheral Models
•Patterson-Meddis
Gammatone Filterbank
•Lyon’s Cochlear Model
Gammatone Filterbank
Adaptation Stage
Conclusions
•Based on biological data
•Front-End for Speech Processing
Speech Recognition, Speaker ID,
Localization….
•Better performance

similar documents