Obtaining Sine Wave From Square Wave

Hello,

I have a square wave but this wave is not an ideal square wave. What I mean is, for example from the first sample of positive peak to the last sample of the positive peak, I have 100 datas. In the next positive peak, I have 91 datas. I want to plot a sine wave which changes adaptively by the number of samples in the positive and negative peaks. This sine wave's positive period should end when the square waves positive peak drops to the zero. Or the best continuous wave that fits with this square wave. Thanks for your advices.

[Walter Roberson]

You can certainly splice together cycles, but I am wondering if that is good enough, or if you need more continuity? And I wonder if you need a pure sine for each cycle? For example if you built a sine function with increasing frequency over time then it might be continuous but it would be an increasingly compressed wave (frequency modulation)

[tinkyminky93]

No i do not need a pure sine wave, enough is enough. It can be like linear interpolation.

[Image Analyst]

"I have a square wave" <== but you forgot to attach it. Please attach it as a .mat or .txt file with the paperclip icon.

Answer 1

[Chunru]

% generate a square wave with no fixed period
f = 0.125; fs=1;
s = 0.3+sin(2*pi*f*(0:40));  % a sine wave with offset
x =2* (s>0) -1;                 % square wave
stem(x);
y = lowpass(x, f*1.5, fs);
hold on
plot(y)

[tinkyminky93]

Why you add offset sır? And is it possible to add zeros to sine wave? When the period ends, i have 5 sample of zeros

[Walter Roberson]

The offset biases there to be more positive values than negative values.

However the end result has fixed frequency, f: the 0:40 are acting as time.

[tinkyminky93]

Simply low pass filtering my square wave gives me a sine wave. Seems like I do not need a solution like that.