Extract Numerator and Denominator of a vector (ratio)

Hello Everyone,

I appreciate your helps in advance as I am a beginner in MATLAB!

I have a matrix and I should measure a ratio of a simple variable (it is a signal) and I need separate numerators and denominators (separated vectors) for next steps. My code is below. Thank you!

clc;
clear all;
close all;
%.....................................................................
data=xlsread('1');
amp=data(:,2);%us/f
time=data(:,1);%ms
t=time(1:1:end);% distans approximately 1 ms
am=amp(1:1:end);
l=length(t);
for i=1:l
    if am(i)==-999.25
        am(i)=am(i-1);
    end
end
for j=2:l
    r(j)=(am(j)-am(j-1))/(am(j)+am(j-1));
end

Answer 1

[Scott MacKenzie]

It is not entirely clear what you are trying to do. Your 1st for-loop does nothing, since all the values are positive. However, if I understand correctly, your second for-loop reveals how you want to spit the signal into terms for the numerator (us) and denominator (f) terms. If that's correct, then, instead of a single vector r, you likely want separate vectors, us and f. Given this, here is what I put together:

data = readmatrix('https://www.mathworks.com/matlabcentral/answers/uploaded_files/651850/01.xlsx');

amp = data(:,2); % signal (us/f)
time = data(:,1); % time (1 ms intervals, not needed)

% this is equivalent to your second for-loop (except separating the terms)
us = amp(2:end) - amp(1:end-1);
f = amp(2:end) + amp(1:end-1);

tiledlayout(2,1);

nexttile;
plot(amp);
xlabel('Time (ms)');
ylabel('Signal');

nexttile;
p1 = plot(f);
hold on;
p2 = plot(us);
xlabel('Time (ms)');
ylabel('Signal (separated)');
legend({'f' 'us'}, 'location', 'east');

[Maria Amr]

Thank you sir for your responding but there is an error:

Error using readmatrix (line 157)

Invalid filename :

https://www.mathworks.com/matlabcentral/answers/uploaded_files/651850/01.xlsx

Would you please attached the excel file? Appreciated!

[Scott MacKenzie]

It's your file! I have no idea why you are getting that error. You can try replacing this lines with

data = readmatrix('01.xlsx');

Again, this is your file. Assuming the file is your workspace (or MATLAB's path), the data should be read without a problem.

[Maria Amr]

@Scott MacKenzie

Dear Scott,

I have rune the code but the figures are different. Would you please look at this?

[Maria Amr]

@Scott MacKenzie, Excuse me! I assumed a new file was attached. I run it but the figures are different.

[Scott MacKenzie]

Hmm, not sure what the issue is. In my solutions, I usually omit...

clear; clc; clf; % fresh start
close all;

Please insert these at the top and try again.

Also, make sure 01.xlsx is the same file you originally posted. It should have 2 columns of data and 940 rows.

[Maria Amr]

@Scott MacKenzie

Dear Scott, I did but the error is remaining. I truly apprecite your time. The procedure is correct. Thank you!

[Scott MacKenzie]

By "error" you meant the figures are different, right?

I'm running R2021a. You?

Also, after running the script, type whos in the command window:

>> whos

Name Size Bytes Class

amp 940x1 7520 double

data 940x2 15040 double

f 939x1 7512 double

p1 1x1 8 matlab.graphics.chart.primitive.Line

p2 1x1 8 matlab.graphics.chart.primitive.Line

time 940x1 7520 double

us 939x1 7512 double

Is this what you see?

[Scott MacKenzie]

I just thought of something else. The plot function in MATLAB has different versions, depending on the toolboxes installed. Try typing which plot in the command window:

>> which plot

built-in (C:\Program Files\MATLAB\R2021a\toolbox\matlab\graph2d\plot)

This is the standard 2D version of plot that is built-in to MATLAB. Is this what you see?

[Maria Amr]

@Scott MacKenzie,

Thank you for your explanation. I am using MATLAB 2020b and I am going to update it to 2021a.

I hope it be solved!

[Maria Amr]

@Scott MacKenzie,

Dear Scott,

I have removed MATLAB 2020b and Installed 2021a , but it seems the results is not compatible. In my code the second loop measures r and its plot is attached. Each coefficients is counted as the numeratore of the above layer and the denominatore of the below layer (It is a layerd medium) and changes between -1 to 1. I need to separate them and get the coefficients vectors (between -0.1 to 1). I appreciate your point.

[Maria Amr]

@Scott MacKenzie,

f and us values are so large.

[Scott MacKenzie]

@Maria Amr In my code I just added a line to compute r as follows:

us = amp(2:end) - amp(1:end-1);
f = amp(2:end) + amp(1:end-1);
r = us ./ f;

This calculation for r is equivalent to what you do in your second for-loop. I also added code to plot r using the stem function and I get the same plot as you, so that's good.

I'm still guessing what your overall goal is here. I think what you are trying to do is take the original signal (the vector amp in your code) and convert each value into a numerator term (us) and a denominator term (f). Is that correct? I suspect amp was actually calculated or measured that way in the first place and that you are trying to reverse engineer amp into the original data, us and f. These are what you call the upper layer and the lower layer. Is that correct?

So, if the formulas you provided for us and f are correct, then the plot for r should be the same as the plot for the original signal (amp), because that's what r is -- a recreation of the original signal from the original numerator and denominator terms. But, the two plots are different. That tells me that the formulas you provided for the numerator (us) and the denominator (f) are wrong. This raises the question, why are you calculating us and f like this:

us = amp(2:end) - amp(1:end-1);
f = amp(2:end) + amp(1:end-1);

This is my starting point -- trying to help with a MATLAB coding problem. But, it seems the fomulas might be wrong in the first place, and I'm not sure I can help with that.

[Maria Amr]

@Scott MacKenzie,

Yes sir. You are totally right. This is aeismic data (a layered earth) and is not original data. Now, I understood how I should use the output of the us and f. Your explanation is so useful! Appreciated!