2019年1月8日 星期二

[signal] 最簡單的方法實做 lowpass filter (使用 ThinkDSP)

How to implement a lowpass filter using ThinkDSP  
井民全, Jing, mqjing@gmail.com
Back to the Main Page

Command

spectrum.low_pass(3000)

Code

from __future__ import print_function, division

%matplotlib inline

import thinkdsp
import thinkplot

import numpy as np

#import warnings
#warnings.filterwarnings('ignore')

from ipywidgets import interact, interactive, fixed
import ipywidgets as widgets
from IPython.display import display

# Load the wave
wave=thinkdsp.read_wave('92002__jcveliz__violin-origional.wav')
wave.make_audio()     // verification step

# Get the Segment
start = 1.2
duration = 0.6
segment = wave.segment(start, duration)
segment.plot()
thinkplot.config(xlabel='Time (s)')

# Wave provides make_spectrum, which computes the spectrum of the wave.
spectrum = segment.make_spectrum()
spectrum.plot()
thinkplot.config(xlabel='Frequency (Hz)')

spectrum.low_pass(3000)
spectrum.plot(high = 10000)
thinkplot.config(xlabel='Frequency (Hz)')
E.g.

Spectrum to wave

filtered = spectrum.make_wave()
filtered.normalize()     # normalize to the range -1 to 1.

filtered.apodize()




More Information

File -> Wave

wave=thinkdsp.read_wave('92002__jcveliz__violin-origional.wav')
wave.make_audio()

Wave -> Segment (wave)

start = 1.2
duration = 0.6
segment = wave.segment(start, duration)
segment.plot()
thinkplot.config(xlabel='Time (s)')

Wave to spectrum

# Wave provides make_spectrum, which computes the spectrum of the wave.
spectrum = segment.make_spectrum()
spectrum.plot()
thinkplot.config(xlabel='Frequency (Hz)')
spectrum.plot(high = 10000)

Low Pass

spectrum.low_pass(3000)
spectrum.plot(high = 10000)
thinkplot.config(xlabel='Frequency (Hz)')

Spectrum to Wave

filtered = spectrum.make_wave()
filtered.normalize()     # normalize to the range -1 to 1.

filtered.apodize()
filtered.plot()
thinkplot.config(xlabel='Time (s)')
Why Apodization?
An example of apodization is the use of the Hann window in the Fast Fourier transform analyzer to smooth the discontinuities at the beginning and end of the sampled time record.

Original
segment.normalize()
segment.apodize()
segment.plot()
thinkplot.config(xlabel='Time (s)')

Play segments

segment.make_audio()
filter.make_audio()

Reference


at