/sci/ - Science & Math

asked another signals question earlier and I'm back again
I don't understand what convolution is. I understand a signal can be represented as a sum of scaled/shifted unit impulses.
Then what I'm trying to get is that h(n-k) is really just a representation for all the unit impulses after you shove d(n-k) through the system? And then multiplying this result by x(k) gives ANOTHER sequence, which you can sum to find the convolution sum at just ONE point if you plug in a value for n. But what you really want to do this for is ALL of the values of n. And I don't understand how this result relates back to the system itself.