Lecture 22 Frequency Response Hung-yi Lee Outline (Chapter 11) Frequency Response Amplitude Ratio Phase Shift Ch 11.1 Filter Lowpass Filter Highpass Filter Bandpss Filter Notch Filter Ch 11.2 Bode Plot Ch 11.4 Draw frequency response Review • Network Function X Y Hs X Network Function Y Hs X Hs Output depends on input and complex frequency s. Frequency Response In this lecture, we only focus on s=jω. That is, only consider H(jω) Y Hs X | Y || Hs || X | Y Hs X Only focus on the change of frequency H(jω) is the frequency response. Frequency Response • H(jω) is a complex number, so frequency response is usually represented by two curves |Y| | H j | |X| a Magnitude Ratio H j Y X Phase Shift Example 11.1 R 8 L 0.2H vin (t ) 10 cos 20t 10 cos 300t vout (t ) ? Network Function: Vout R H (s) Vin sL R Frequency Response: R 8 H ( j ) j L R j 0.2 8 0 40 H ( j ) j 40 tan 1 a ( ) H ( j ) 40 40 1600 2 ( ) H ( j ) tan 1 40 Example 11.1 R 8 L 0.2H vin (t ) 10 cos 20t 10 cos 300t vout (t ) ? a ( ) 40 1600 ( ) tan 1 2 40 if 20 a (20) 0.894 (20) 26.6 if 300 a (300) 0.132 (300) 82.4 Based on Superposition: vout 8.94 cos(20t 26.6 ) 1.32 cos(300t 82.4 ) Example 11.1 Input vin (t ) 10 cos 20t 10 cos 300t Output vout 8.94 cos(20t 26.6 ) 1.32 cos(300t 82.4 ) Example 11.2 Find Network Function 1 sC Vout VC Vx 1 sC 1 Vin Vin 1 2 R sC 2 1 R sC sC Vin 1 2 R sC 1 R sC V 1 in 2 R sC H (s) Vout Vin 1 1 s sC 1 RC 1 1 2 s 2 R RC sC R Assume C=1/aR a=1/RC 1 sa H (s) 2 sa 1 K , z a, p a 2 Example 11.2 1 sa 1 H (s) a 2 sa RC 1 K , z a, p a 2 Set s=jω 1 j a H ( j ) 2 j a a j 2a j a a 2 2 tan 1 a a tan 2 2 2 1 1 sC a ( ) H ( j ) a2 2 2 a2 2 1 2 The amplitude ratio is constant. ( ) H ( j ) tan 1 a tan 1 a 2 tan 1 a Example 11.2 1 j a H ( j ) 2 j a Distance of jω to a Magnitude Ratio 1 | j a | 1 | H ( j ) || | 2 | j a | 2 Distance of jω to -a Phase Shift 2 tan 1 a Example 11.2 • Phase shift network: Waveform distortion 1 a ( ) 2 2 tan a vin (t ) 10 cos10t 10 cos 20t 1 1 a 10 RC vout 5 cos(10t 90 ) 5 cos(20t 127 ) Another Example for Phase Shift Network 1 sC vout vin 1 sC V2 1 sC 1 R sC Vin V1 R R 1 sC Vin Vout V2 V1 1 R sC R 1 R 1 sC sC 1 R sC R 1 sC Vin sCR 1 Vin sCR 1 Vin Another Example for Phase Shift Network vin vout Vout sCR 1 Vin sCR 1 sCR 1 Hs Vin sCR 1 Vout jCR 1 H j jCR 1 | jCR 1 | a 1 | jCR 1 | jCR 1 jCR 1 tan 1 CR tan 1 CR 2 tan 1 CR Computing Frequency Response by Poles and Zeros Frequency Response s z1 s z 2 Hs K s p1 s p2 K: gain factor p1, p2, …: poles z1, z2, …: zeros p1 z1 z2 j z1 j z 2 H j K j p1 j p2 j p2 Frequency Response - Magnitude Ratio j z1 j z 2 H j K j p1 j p2 Distance of jω to z1 a ( ) | H j | K Distance of jω to z2 j z1 j z 2 j p1 j p2 Distance of jω to p1 Distance of jω to p2 Frequency Response - Magnitude Ratio Magnitude ratio at ω p1 d4 j =| Gain factor| X d2 all distance to zeros all distance to poles z1 z2 Gain factor is K d1d 2 a ( ) K d 3d 4 d1 p2 d3 Frequency Response - Phase Shift j z1 j z 2 H j K j p1 j p2 ( ) K 0 or 180 1 2 3 4 3 j 1 z1 z2 j p1 j p2 ( ) K p1 2 j z1 j z 2 p2 4 Example 11.3 H ( j ) 20 j 25 2 j 20 500 H (s) 20s 25 s 2 20 s 500 y tan 1 x 2025 j 500 2 j 20 a ( ) 20 625 2 500 2 2 400 2 20 1 1 2 tan tan 500 2 25 500 ( ) tan 1 tan 1 20 180 2 500 25 500 2 Example 11.3 20s 25 H (s) 2 s 20 s 500 K 20 z1 25 p1 , p2 10 j 20 Example 11.3 20s 25 H (s) 2 s 20 s 500 K 20 z1 25 p1 , p2 10 j 20 Homework • 11.2, 11.4 Application: Hearing Reference: http://hyperphysics.phy-astr.gsu.edu/hbase/sound/hearcon.html dB Range of Hearing Range of hearing: 20Hz - 20KHz Range of Audio: 300Hz – 3.4KHz Frequency Equal Loudness Contour Blue Line: Same Energy dB Red Curve: Equal loudness phon (unit): 60 phon means as loud as the 60dB 1kHz sound. Hearing v.s. Age Application: 柯南 68 集 (黑暗中消失的麒麟之角) Phase and Sound Pass a Phase shift network Structure of Ear Cochlea (耳蝸) Pass 20kHz Each neuron only passes a specific frequency. Pass 20Hz to brain Pass high frequency Pass low frequency Cochlea (耳蝸) vin (t ) 10 cos 20t 10 cos 300t Neuron for ω=300 vout (t ) 10 cos 300t Neuron for ω=100 vout (t ) 0 ω=300, Amplitude=10 ω=20, Amplitude=10 to brain Neuron for ω=20 vout (t ) 10 cos 20t Machine Hearing ( ) A set of filters (filter bank) Ear of machine Feature Send to computer How? Each triangle is the amplitude ratio of a filter Energy of the signal of the output of the filter How to write a program a. Observe data, and write down for speech recognition? some rules b. Simulate human brain, and let the computer learn by itself. Speech Signal Signal Processing (Filter Bank) Features Acoustic Model Machine Learning ㄨㄣˊ ㄧˋ … 文藝? 聞腋? Text Language Model Deep Neural Network (Brain) Speech Recognition Thank you! Answer • 11.2: 1.075 cos 2t 3.9 1.346 cos cos 10t 21.8 0.482 cos 50t 71.9 • 11.4: 0.186 cos 2t 71 0.485 cos cos 10t 39.1 0.837 cos 50t 22.6 Acknowledgement • 感謝 陳俞兆(b02) • 上課時指出投影片的錯誤 • 感謝 陳均彥(b02) • 上課時指出投影片的錯誤 • 感謝 林楷恩(b02) • 指出投影片方程式的錯誤