Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I want to convolve two triangles. The equation satisfied by one triangle is

$$f(y) = \begin{cases} y + 1 & −1 < y < 0\\ \\ 1 − y & 0 \leq y < 1 \end{cases}.$$

So, the overall duration of a triangle is $\;-1\;$ to $\;1$.

What I have found is that the convolution of a triangle with itself is a $\operatorname{sinc}$ function but does one obtain this $\operatorname{sinc}$ from the above equation?

Secondly, what should be the limits of integral for calculating convolution?

share|improve this question
1  
It might be worthwhile asking the moderators to migrate this question to dsp.SE. With regard to your question about the limits on the integral for calculating convolutions, there is not a single integral that you have to compute, but different integrals depending on your choice for the argument of the function that is the result of the convolution. For a detailed exposition of how to go about convolving two signals, see my answer to a similar question on dsp.SE –  Dilip Sarwate Dec 22 '12 at 15:28

2 Answers 2

We have to compute $$g(x):=\int_{-\infty}^\infty f(x-t)\ f(t)\ dt=\int_{-1}^0 f(x-t)\ (1+t)\ dt + \int_0^1 f(x-t)\ (1-t)\ dt$$ for given $x\in{\mathbb R}$. Because of the nonanalytic way the function $f$ is defined this is a tedious matter. It's pretty obvious that $g(-x)=g(x)$, and that $g(x)=0$ when $x>2$.

Let an $x\in{\mathbb R}$ be fixed. Then $$f(x-t)=\cases{ 0&$(t<x-1)$\cr 1-x+t\quad &$(x-1\leq t\leq x)$\cr 1+x-t&$(x\leq t\leq x+1)$\cr 0&$(t>x+1)$\cr}\quad.$$

We distinguish the cases (i): $\ 0< x< 1$, and (ii): $\ 1< x<2$. (The cases $x\in\{0,1,2\}$ will follow by continuity of $g$.)

In case (i) we have $-1<x-1<0<x<1$ and therefore $$g(x)=\int_{x-1}^0 (1-x+t)(1+t)\ dt +\int_0^x (1-x+t)(1-t)\ dt +\int_x^1 (1+x-t)(1-t)\ dt\ ,$$ which Mathematica computes to $$g(x)={2\over3}-x^2+{x^3\over2}\ .$$ In case (ii) we have $0<x-1<1<x$. Therefore the integral $\int_{-1}^0\ldots$ vanishes, and we are left with $$g(x)=\int_{x-1}^1 (1-x+t)(1-t)\ dt={4\over3}-2x+x^2-{x^3\over6}\ .$$

share|improve this answer

Use the formula of convolution from here http://en.wikipedia.org/wiki/Convolution the limits of your integral will be from $-2$ to $2$. For bounded signals in some interval $[a,b]$, the result of the convolution will double the range from $b-a$ to $2(b-a)$.

Your range is $1-(-1)=2$, and you can expect to double it to $2*2=4$ and how is it related to defining the limits of the integral?

The resul of the convolution is not a $sinc$ function. $sinc$ function is not positive everywhere but the result of a convolution of two non-negative functions is always non-negative everywhere.

share|improve this answer
    
Okay thanks between I am also not getting it how to convolve this specific function because it has different ranges for y. –  Zara Dec 22 '12 at 14:08
    
This is just a single function. You will convolve it with itself. This means the range will be between $-1$ and $1$ in both cases. –  Seyhmus Güngören Dec 22 '12 at 14:32
    
But when I will be integrating then what will i write y+1 or y-1 inside the integral? –  Zara Dec 22 '12 at 14:42
1  
you will not write a single function in the integral. Your function says that the function is $y+1$, if $-1<y<0$ and $1-y$ if $0<y<1$. Depending on the limits you must use either of them. Suggestion: try to plot the function on a paper. –  Seyhmus Güngören Dec 22 '12 at 15:19

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.