Discussion Forum

Hello,

I do not hope for this question to be answered before the midterm exam, but I still would be glad if someone could help me understand where is the error of my reasoning (or confirm that there is something strange with the provided solution).

In the provided correction, it is stated that non-zero coefficients must verify

$(n \geq 1) \land (m < \log_2(\frac{n}{2^{-i}+1})) \land (m > -i -\log_2(n+1))$

My problem is that I find a counter-example by choosing i=1,n=0, m=2. (Violates first part of the boolean expression and one can argue that the second expression in the composition is ill-defined in this case).

Integrating the product of the wavelets yields $\frac{1}{2}$. And if I plot it (wavelet in blue and f_1(t) in red), it does not look like the coefficient is zero: 

Thank you in advance !