We Keep Coding

Separating two interleaved arrays into continuous sequences

I'm not sure how to best explain my question in words, so I will provide a code example below. But to at least give it a try. I am solving an eigenvalue problem as a function of some external parameter, which results in two eigenvalues. Those two eigenvalues cross, as the function of the external parameter. Eigenvalue sorting then leads to the wrong classification of which eigenvalue belongs to what 'branch' of the problem. I'd like to disentangle that.
Okay, as for the example. We start from the desired result and then mess it up according to what happens in the diagonalization routine.
xs = np.arange(-np.pi, np.pi, 0.001)
fun_a = np.cos(xs) + 0.5*np.sin(xs)
fun_b = np.cos(xs) - 0.5*xs*np.sin(xs) + 0.2
plt.figure()
plt.plot(xs, fun_a)
plt.plot(xs, fun_b)
This results in two smooth branches that cross each other:
Now, what happens is that instead the eigenvalues are sorted, which I can mimmic as follows:
fun_c = np.zeros_like(fun_a)
fun_c[fun_a>fun_b] = fun_a[fun_a>fun_b]
fun_c[fun_a<=fun_b] = fun_b[fun_a<=fun_b]
fun_d = np.zeros_like(fun_a)
fun_d[fun_a>fun_b] = fun_b[fun_a>fun_b]
fun_d[fun_a<=fun_b] = fun_a[fun_a<=fun_b]
plt.figure()
plt.plot(xs, fun_c)
plt.plot(xs, fun_d)
The result is these two branches, which are no longer smooth, continuous branches
What my question boils down to is how we can go from the second case (fun_c, fun_d) to the first case (fun_a, fun_b). I suppose one could use the information in the derivative, to some extent. Using np.diff() reveals a sharp discontinuity at the point where things go wrong. But I don't immediately see how to nicely use that.
edit: I'm starting to think the best assignment would be such that each subsequent point is chosen to minimize the change in the slope. But I'm not yet sure how to do that..

I'm sure there are many adjustments to be made to improve performance, but here's my guess, based on second derivative thresholding:
# first derivative
deriv = np.diff(fun_c)
# second derivative
deriv2 = np.diff(deriv)
# adjust threshold to detect discontinuities
switch_points = deriv2 > 0.0002
# indices of points of intersection
touch_index = np.where(switch_points == True)[0]
# adjustment to prevent false positives
# (sometimes contigous samples like 127,128 are both detected)
duplicate_index = np.where(np.diff(touch_index) == 1)[0]
touch_index = np.delete(touch_index, duplicate_index)
# begin and ending points of sections to swap
begins = touch_index[::2]
ends = touch_index[1::2]
from itertools import zip_longest
# swap values in selected sections
tmp = fun_c.copy() # need a tmp array to swap
for begin, end in zip_longest(begins, ends, fillvalue=len(fun_c)):
tmp[begin:end] = fun_d[begin:end]
# complete swapping, correcting fun_d
# on the indices we changed before
swapped = tmp != fun_c
fun_d[swapped] = fun_c[swapped]
fun_c = tmp
plt.plot(xs, fun_c)
plt.plot(xs, fun_d)
I found it quite dependant on the sampling_rate (might fail if it is too low).
Edit:
If it wasn't for the
touch_index = np.delete(touch_index, duplicate_index)
you could entirely skip all this:
begins = touch_index[::2]
ends = touch_index[1::2]
from itertools import zip_longest
# swap values in selected sections
tmp = fun_c.copy() # need a tmp array to swap
for begin, end in zip_longest(begins, ends, fillvalue=len(fun_c)):
tmp[begin:end] = fun_d[begin:end]
with just
np.logical_xor.accumulate(touch_index) | touch_index

Change the melody of human speech using FFT and polynomial interpolation

I'm trying to do the following:
Extract the melody of me asking a question (word "Hey?" recorded to
wav) so I get a melody pattern that I can apply to any other
recorded/synthesized speech (basically how F0 changes in time).
Use polynomial interpolation (Lagrange?) so I get a function that describes the melody (approximately of course).
Apply the function to another recorded voice sample. (eg. word "Hey." so it's transformed to a question "Hey?", or transform the end of a sentence to sound like a question [eg. "Is it ok." => "Is it ok?"]). Voila, that's it.
What I have done? Where am I?
Firstly, I have dived into the math that stands behind the fft and signal processing (basics). I want to do it programatically so I decided to use python.
I performed the fft on the entire "Hey?" voice sample and got data in frequency domain (please don't mind y-axis units, I haven't normalized them)
So far so good. Then I decided to divide my signal into chunks so I get more clear frequency information - peaks and so on - this is a blind shot, me trying to grasp the idea of manipulating the frequency and analyzing the audio data. It gets me nowhere however, not in a direction I want, at least.
Now, if I took those peaks, got an interpolated function from them, and applied the function on another voice sample (a part of a voice sample, that is also ffted of course) and performed inversed fft I wouldn't get what I wanted, right?
I would only change the magnitude so it wouldn't affect the melody itself (I think so).
Then I used spec and pyin methods from librosa to extract the real F0-in-time - the melody of asking question "Hey?". And as we would expect, we can clearly see an increase in frequency value:
And a non-question statement looks like this - let's say it's moreless constant.
The same applies to a longer speech sample:
Now, I assume that I have blocks to build my algorithm/process but I still don't know how to assemble them beacause there are some blanks in my understanding of what's going on under the hood.
I consider that I need to find a way to map the F0-in-time curve from the spectrogram to the "pure" FFT data, get an interpolated function from it and then apply the function on another voice sample.
Is there any elegant (inelegant would be ok too) way to do this? I need to be pointed in a right direction beceause I can feel I'm close but I'm basically stuck.
The code that works behind the above charts is taken just from the librosa docs and other stackoverflow questions, it's just a draft/POC so please don't comment on style, if you could :)
fft in chunks:
import numpy as np
import matplotlib.pyplot as plt
from scipy.io import wavfile
import os
file = os.path.join("dir", "hej_n_nat.wav")
fs, signal = wavfile.read(file)
CHUNK = 1024
afft = np.abs(np.fft.fft(signal[0:CHUNK]))
freqs = np.linspace(0, fs, CHUNK)[0:int(fs / 2)]
spectrogram_chunk = freqs / np.amax(freqs * 1.0)
# Plot spectral analysis
plt.plot(freqs[0:250], afft[0:250])
plt.show()
spectrogram:
import librosa.display
import numpy as np
import matplotlib.pyplot as plt
import os
file = os.path.join("/path/to/dir", "hej_n_nat.wav")
y, sr = librosa.load(file, sr=44100)
f0, voiced_flag, voiced_probs = librosa.pyin(y, fmin=librosa.note_to_hz('C2'), fmax=librosa.note_to_hz('C7'))
times = librosa.times_like(f0)
D = librosa.amplitude_to_db(np.abs(librosa.stft(y)), ref=np.max)
fig, ax = plt.subplots()
img = librosa.display.specshow(D, x_axis='time', y_axis='log', ax=ax)
ax.set(title='pYIN fundamental frequency estimation')
fig.colorbar(img, ax=ax, format="%+2.f dB")
ax.plot(times, f0, label='f0', color='cyan', linewidth=2)
ax.legend(loc='upper right')
plt.show()
Hints, questions and comments much appreciated.

The problem was that I didn't know how to modify the fundamental frequency (F0). By modifying it I mean modify F0 and its harmonics, as well.
The spectrograms in question show frequencies at certain points in time with power (dB) of certain frequency point.
Since I know which time bin holds which frequency from the melody (green line below) ...
....I need to compute a function that represents that green line so I can apply it to other speech samples.
So I need to use some interpolation method which takes as parameters the sample F0 function points.
One need to remember that degree of the polynomial should equal to the number of points. The example doesn't have that unfortunately, but the effect is somehow ok as for the prototype.
def _get_bin_nr(val, bins):
the_bin_no = np.nan
for b in range(0, bins.size - 1):
if bins[b] <= val < bins[b + 1]:
the_bin_no = b
elif val > bins[bins.size - 1]:
the_bin_no = bins.size - 1
return the_bin_no
def calculate_pattern_poly_coeff(file_name):
y_source, sr_source = librosa.load(os.path.join(ROOT_DIR, file_name), sr=sr)
f0_source, voiced_flag, voiced_probs = librosa.pyin(y_source, fmin=librosa.note_to_hz('C2'),
fmax=librosa.note_to_hz('C7'), pad_mode='constant',
center=True, frame_length=4096, hop_length=512, sr=sr_source)
all_freq_bins = librosa.core.fft_frequencies(sr=sr, n_fft=n_fft)
f0_freq_bins = list(filter(lambda x: np.isfinite(x), map(lambda val: _get_bin_nr(val, all_freq_bins), f0_source)))
return np.polynomial.polynomial.polyfit(np.arange(0, len(f0_freq_bins), 1), f0_freq_bins, 3)
def calculate_pattern_poly_func(coefficients):
return np.poly1d(coefficients)
Method calculate_pattern_poly_coeff calculates polynomial coefficients.
Using pythons poly1d lib I can compute function which can modify the speech. How to do that?
I just need to move up or down all values vertically at certain point in time.
for instance I want to move all frequencies at time bin 0,75 seconds up 3 times -> it means that frequency will be increased and the melody at that point will sound higher.
Code:
def transform(sentence_audio_sample, mode=None, show_spectrograms=False, frames_from_end_to_transform=12):
# cutting out silence
y_trimmed, idx = librosa.effects.trim(sentence_audio_sample, top_db=60, frame_length=256, hop_length=64)
stft_original = librosa.stft(y_trimmed, hop_length=hop_length, pad_mode='constant', center=True)
stft_original_roll = stft_original.copy()
rolled = stft_original_roll.copy()
source_frames_count = np.shape(stft_original_roll)[1]
sentence_ending_first_frame = source_frames_count - frames_from_end_to_transform
sentence_len = np.shape(stft_original_roll)[1]
for i in range(sentence_ending_first_frame + 1, sentence_len):
if mode == 'question':
by = int(_question_pattern(i) / 500)
elif mode == 'exclamation':
by = int(_exclamation_pattern(i) / 500)
else:
by = 0
rolled = _roll_column(rolled, i, by)
transformed_data = librosa.istft(rolled, hop_length=hop_length, center=True)
def _roll_column(two_d_array, column, shift):
two_d_array[:, column] = np.roll(two_d_array[:, column], shift)
return two_d_array
In this case I am simply rolling up or down frequencies referencing certain time bin.
This needs to be polished as it doesn't take into consideration an actual state of the transformed sample. It just rolls it up/down according to the factor calculated using the polynomial function computer earlier.
You can check full code of my project at github, "audio" package contains pattern calculator and audio transform algorithm described above.
Feel free to ask if something's unclear :)

How can I cut a piece away from a plot and set the point I need to zero?

In my work I have the task to read in a CSV file and do calculations with it. The CSV file consists of 9 different columns and about 150 lines with different values acquired from sensors. First the horizontal acceleration was determined, from which the distance was derived by double integration. This represents the lower plot of the two plots in the picture. The upper plot represents the so-called force data. The orange graph shows the plot over the 9th column of the CSV file and the blue graph shows the plot over the 7th column of the CSV file.
As you can see I have drawn two vertical lines in the lower plot in the picture. These lines represent the x-value, which in the upper plot is the global minimum of the orange function and the intersection with the blue function. Now I want to do the following, but I need some help: While I want the intersection point between the first vertical line and the graph to be (0,0), i.e. the function has to be moved down. How do I achieve this? Furthermore, the piece of the function before this first intersection point (shown in purple) should be omitted, so that the function really only starts at this point. How can I do this?
In the following picture I try to demonstrate how I would like to do that:
If you need my code, here you can see it:
import numpy as np
import matplotlib.pyplot as plt
import math as m
import loaddataa as ld
import scipy.integrate as inte
from scipy.signal import find_peaks
import pandas as pd
import os
# Loading of the values
print(os.path.realpath(__file__))
a,b = os.path.split(os.path.realpath(__file__))
print(os.chdir(a))
print(os.chdir('..'))
print(os.chdir('..'))
path=os.getcwd()
path=path+"\\Data\\1 Fabienne\\Test1\\left foot\\50cm"
print(path)
dataListStride = ld.loadData(path)
indexStrideData = 0
strideData = dataListStride[indexStrideData]
#%%Calculation of the horizontal acceleration
def horizontal(yAngle, yAcceleration, xAcceleration):
a = ((m.cos(m.radians(yAngle)))*yAcceleration)-((m.sin(m.radians(yAngle)))*xAcceleration)
return a
resultsHorizontal = list()
for i in range (len(strideData)):
strideData_yAngle = strideData.to_numpy()[i, 2]
strideData_xAcceleration = strideData.to_numpy()[i, 4]
strideData_yAcceleration = strideData.to_numpy()[i, 5]
resultsHorizontal.append(horizontal(strideData_yAngle, strideData_yAcceleration, strideData_xAcceleration))
resultsHorizontal.insert(0, 0)
#plt.plot(x_values, resultsHorizontal)
#%%
#x-axis "convert" into time: 100 Hertz makes 0.01 seconds
scale_factor = 0.01
x_values = np.arange(len(resultsHorizontal)) * scale_factor
#Calculation of the global high and low points
heel_one=pd.Series(strideData.iloc[:,7])
plt.scatter(heel_one.idxmax()*scale_factor,heel_one.max(), color='red')
plt.scatter(heel_one.idxmin()*scale_factor,heel_one.min(), color='blue')
heel_two=pd.Series(strideData.iloc[:,9])
plt.scatter(heel_two.idxmax()*scale_factor,heel_two.max(), color='orange')
plt.scatter(heel_two.idxmin()*scale_factor,heel_two.min(), color='green')#!
#Plot of force data
plt.plot(x_values[:-1],strideData.iloc[:,7]) #force heel
plt.plot(x_values[:-1],strideData.iloc[:,9]) #force toe
# while - loop to calculate the point of intersection with the blue function
i = heel_one.idxmax()
while strideData.iloc[i,7] > strideData.iloc[i,9]:
i = i-1
# Length calculation between global minimum orange function and intersection with blue function
laenge=(i-heel_two.idxmin())*scale_factor
print(laenge)
#%% Integration of horizontal acceleration
velocity = inte.cumtrapz(resultsHorizontal,x_values)
plt.plot(x_values[:-1], velocity)
#%% Integration of the velocity
s = inte.cumtrapz(velocity, x_values[:-1])
plt.plot(x_values[:-2],s)
I hope it's clear what I want to do. Thanks for helping me!

I didn't dig all the way through your code, but the following tricks may be useful.
Say you have x and y values:
x = np.linspace(0,3,100)
y = x**2
Now, you only want the values corresponding to, say, .5 < x < 1.5. First, create a boolean mask for the arrays as follows:
mask = np.logical_and(.5 < x, x < 1.5)
(If this seems magical, then run x < 1.5 in your interpreter and observe the results).
Then use this mask to select your desired x and y values:
x_masked = x[mask]
y_masked = y[mask]
Then, you can translate all these values so that the first x,y pair is at the origin:
x_translated = x_masked - x_masked[0]
y_translated = y_masked - y_masked[0]
Is this the type of thing you were looking for?

Using find_peaks in loop to find peak heights from data

This problem is about using scipy.signal.find_peaks for extracting mean peak height from data files efficiently. I am a beginner with Python (3.7), so I am not sure if I have written my code in the most optimal way, with regard to speed and code quality.
I have a set of measurement files containing one million data points (30MB) each. The graph of this data is a signal with peaks at regular intervals, and with noise. Also, the baseline and the amplitude of the signal vary across parts of the signal. I attached an image of an example. The signal can be much less clean.
My goal is to calculate the mean height of the peaks for each file. In order to do this, first I use find_peaks to locate all the peaks. Then, I loop over each peak location and detect the peak in a small interval around the peak, to make sure I get the local height of the peak.
I then put all these heights in numpy arrays and calculate their mean and standard deviation afterwards.
Here is a barebone version of my code, it is a bit long but I think that might also be because I am doing something wrong.
import numpy as np
from scipy.signal import find_peaks
# Allocate empty lists for values
mean_heights = []
std_heights = []
mean_baselines = []
std_baselines = []
temperatures = []
# Loop over several files, read them in and process data
for file in file_list:
temperatures.append(file)
# Load in data from a file of 30 MB
t_dat, x_dat = np.loadtxt(file, delimiter='\t', unpack=True)
# Find all peaks in this file
peaks, peak_properties = find_peaks(x_dat, prominence=prom, width=0)
# Calculate window size, make sure it is even
if round(len(t_dat)/len(peaks)) % 2 == 0:
n_points = len(t_dat) // len(peaks)
else:
n_points = len(t_dat) // len(peaks) + 1
t_slice = t_dat[-1] / len(t_dat)
# Allocate np arrays for storing heights
baseline_list = np.zeros(len(peaks) - 2)
height_list = np.zeros(len(peaks) - 2)
# Loop over all found peaks, and re-detect the peak in a window around the peak to be able
# to detect its local height without triggering to a baseline far away
for i in range(len(peaks) - 2):
# Making a window around a peak_properties
sub_t = t_dat[peaks[i+1] - n_points // 2: peaks[i+1] + n_points // 2]
sub_x = x_dat[peaks[i+1] - n_points // 2: peaks[i+1] + n_points // 2]
# Detect the peaks (2 version, specific to the application I have)
h_min = max(sub_x) - np.mean(sub_x)
_, baseline_props = find_peaks(
sub_x, prominence=h_min, distance=n_points - 1, width=0)
_, height_props = find_peaks(np.append(
min(sub_x) - 1, sub_x), prominence=h_min, distance=n_points - 1, width=0)
# Add the heights to the np arrays storing the heights
baseline_list[i] = baseline_props["prominences"]
height_list[i] = height_props["prominences"]
# Fill lists with values, taking the stdev and mean of the np arrays with the heights
mean_heights.append(np.mean(height_list))
std_heights.append(np.std(height_list))
mean_baselines.append(np.mean(baseline_list))
std_baselines.append(np.std(baseline_list))
It takes ~30 s to execute. Is this normal or too slow? If so, can it be optimised?

In the meantime I have improved the speed by getting rid of various inefficiencies, that I found by using the Python profiler. I will list the optimisations here ordered by significance for the speed:
Using pandas pd.read_csv() for I/O instead of np.loadtxt() cut off about 90% of the runtime. As also mentioned here, this saves a lot of time. This means changing this:
t_dat, x_dat = np.loadtxt(file, delimiter='\t', unpack=True)
to this:
data = pd.read_csv(file, delimiter = "\t", names=["t_dat", "x_dat"])
t_dat = data.values[:,0]
x_dat = data.values[:,1]
Removing redundant len() calls. I noticed that len() was called many times, and then noticed that this happened unnecessary. Changing this:
if round(len(t_dat) / len(peaks)) % 2 == 0:
n_points = int(len(t_dat) / len(peaks))
else:
n_points = int(len(t_dat) / len(peaks) + 1)
to this:
n_points = round(len(t_dat) / len(peaks))
if n_points % 2 != 0:
n_points += 1
proved to be also a significant improvement.
Lastly, a disproportionally component of the computational time (about 20%) was used by the built in Python functions min(), max() and sum(). Since I was using numpy arrays already, switching to the numpy equivalents for these functions, resulted in a 84% improvement on this part. This means for example changing max(sub_x) to sub_x.max().
These are all unrelated optimisations that I still think might be useful for Python beginner like myself, and they do help a lot.

How to pick points under the curve?

What I'm trying to do is make a gaussian function graph. then pick random numbers anywhere in a space say y=[0,1] (because its normalized) & x=[0,200]. Then, I want it to ignore all values above the curve and only keep the values underneath it.
import numpy
import random
import math
import matplotlib.pyplot as plt
import matplotlib.mlab as mlab
from math import sqrt
from numpy import zeros
from numpy import numarray
variance = input("Input variance of the star:")
mean = input("Input mean of the star:")
x=numpy.linspace(0,200,1000)
sigma = sqrt(variance)
z = max(mlab.normpdf(x,mean,sigma))
foo = (mlab.normpdf(x,mean,sigma))/z
plt.plot(x,foo)
zing = random.random()
random = random.uniform(0,200)
import random
def method2(size):
ret = set()
while len(ret) < size:
ret.add((random.random(), random.uniform(0,200)))
return ret
size = input("Input number of simulations:")
foos = set(foo)
xx = set(x)
method = method2(size)
def undercurve(xx,foos,method):
Upper = numpy.where(foos<(method))
Lower = numpy.where(foos[Upper]>(method[Upper]))
return (xx[Upper])[Lower],(foos[Upper])[Lower]
When I try to print undercurve, I get an error:
TypeError: 'set' object has no attribute '__getitem__'
and I have no idea how to fix it.
As you can all see, I'm quite new at python and programming in general, but any help is appreciated and if there are any questions I'll do my best to answer them.

The immediate cause of the error you're seeing is presumably this line (which should be identified by the full traceback -- it's generally quite helpful to post that):
Lower = numpy.where(foos[Upper]>(method[Upper]))
because the confusingly-named variable method is actually a set, as returned by your function method2. Actually, on second thought, foos is also a set, so it's probably failing on that first. Sets don't support indexing with something like the_set[index]; that's what the complaint about __getitem__ means.
I'm not entirely sure what all the parts of your code are intended to do; variable names like "foos" don't really help like that. So here's how I might do what you're trying to do:
# generate sample points
num_pts = 500
sample_xs = np.random.uniform(0, 200, size=num_pts)
sample_ys = np.random.uniform(0, 1, size=num_pts)
# define distribution
mean = 50
sigma = 10
# figure out "normalized" pdf vals at sample points
max_pdf = mlab.normpdf(mean, mean, sigma)
sample_pdf_vals = mlab.normpdf(sample_xs, mean, sigma) / max_pdf
# which ones are under the curve?
under_curve = sample_ys < sample_pdf_vals
# get pdf vals to plot
x = np.linspace(0, 200, 1000)
pdf_vals = mlab.normpdf(x, mean, sigma) / max_pdf
# plot the samples and the curve
colors = np.array(['cyan' if b else 'red' for b in under_curve])
scatter(sample_xs, sample_ys, c=colors)
plot(x, pdf_vals)
Of course, you should also realize that if you only want the points under the curve, this is equivalent to (but much less efficient than) just sampling from the normal distribution and then randomly selecting a y for each sample uniformly from 0 to the pdf value there:
sample_xs = np.random.normal(mean, sigma, size=num_pts)
max_pdf = mlab.normpdf(mean, mean, sigma)
sample_pdf_vals = mlab.normpdf(sample_xs, mean, sigma) / max_pdf
sample_ys = np.array([np.random.uniform(0, pdf_val) for pdf_val in sample_pdf_vals])

It's hard to read your code.. Anyway, you can't access a set using [], that is, foos[Upper], method[Upper], etc are all illegal. I don't see why you convert foo, x into set. In addition, for a point produced by method2, say (x0, y0), it is very likely that x0 is not present in x.
I'm not familiar with numpy, but this is what I'll do for the purpose you specified:
def undercurve(size):
result = []
for i in xrange(size):
x = random()
y = random()
if y < scipy.stats.norm(0, 200).pdf(x): # here's the 'undercurve'
result.append((x, y))
return results