Tutorial 2.5 - support

Generic ECE - University of Houston

Han Q Le (c) -copyrighted

Problem 3 - generate musical sound

For the Mathematica-Python translator beginner. This is based on the Mathematica algorithm given in Tutorial 2.5, section 3. See other notebook for alternatives (slightly more complex, advanced)

In [1]:
import numpy as np
import pandas as pd
import re
import matplotlib.pyplot as plt
import sounddevice as sd # see conda install sounddevice instruction google
In [2]:
notes = ["D5", "D5", "E5", "D5", "G5", "F#5"
   , "D5", "D5", "E5", "D5", "A5", "G5"];
notedur = 1/np.array([8, 16, 4, 4, 4, 2
    , 8, 16, 4, 4, 4, 2]) 
tau = 0.4;  beatdur = 2. ; srate = 11025;
tdur = notedur*beatdur;  # actual time duration of each note 
dt=1/srate
tdurarr=[np.arange(0,td,dt) for td in tdur]
In [3]:
datimp = pd.read_html("https://pages.mtu.edu/~suits/notefreqs.html");
freqdata=np.array(datimp[1])
freqdata=[list(u) for u in freqdata]
def getfreq(note,flist):
    for fr in flist:
        if re.search(note,str.replace(fr[0],' ','')):
            f=fr[1];
            break
        else :
            f=0 ;
    return f
In [4]:
sndout=[]
for i in range(len(notes)):
    freq=getfreq(notes[i],freqdata)
    sndout=sndout+[np.exp(-t/tau)*np.sin(2*np.pi*freq*t) for t in tdurarr[i]]

sd.play(sndout,srate)

note that the code in cell above isn't exactly parallel with the Mathematica code, but quite similar in concept. An alternative code similar to Mathematica code - and can be used in more advanced version to play different instrumental sounds is below. Here, we get the frequency array, so that a play function can be defined, given any input frequency, duration, overtone characteristics, sampling rate etc. Then a simple sound list just play the play function. See Tutorial 2.5, section 3.2

In [5]:
freq=[getfreq(n,freqdata) for n in notes]
sndarr=[[np.exp(-t/tau)*np.sin(2*np.pi*freq[i]*t) for t in tdurarr[i]] 
        for i in range(len(freq))]
sndout=[]
for s in sndarr:
    sndout=sndout+s
sd.play(sndout,srate)
In [ ]: