- #1
Cyrus
- 3,238
- 16
I found this interesting little problem when thinking about convolution:
[tex] \int x( \tau) \delta(t-\tau) d\tau [/tex]
Normally to solve something like this you would have to integrate by parts because of two functions in [tex]\tau[/tex]
Using the fact that:
[tex] \int u *dv = u*v - \int v*du [/tex]
Where
[tex] u=x(\tau)[/tex]
[tex] dv= \delta(t-\tau) d\tau[/tex]
Then:
[tex] du=x'(\tau) d\tau[/tex]
[tex] v= 1 [/tex]
If you plug this back in you get:
[tex]x(\tau) - x(\tau) = 0[/tex]
Total nonsense!
[tex] \int x( \tau) \delta(t-\tau) d\tau [/tex]
Normally to solve something like this you would have to integrate by parts because of two functions in [tex]\tau[/tex]
Using the fact that:
[tex] \int u *dv = u*v - \int v*du [/tex]
Where
[tex] u=x(\tau)[/tex]
[tex] dv= \delta(t-\tau) d\tau[/tex]
Then:
[tex] du=x'(\tau) d\tau[/tex]
[tex] v= 1 [/tex]
If you plug this back in you get:
[tex]x(\tau) - x(\tau) = 0[/tex]
Total nonsense!