On Wed, Apr 9, 2025 at 8:54 AM Elmore Family <wa4...@comcast.net> wrote:
> I understand that the block is essentially an FIR high pass filter with > the passband starting at 0. Thus it would be removing the negative > frequency of the imaginary part of the signal as well as shifting the > signal by 90 degrees. > > How does it operate to remove the negative frequency of the real part of > the signal as well since this part of the signal is not shifted? Neither the real part nor the imaginary part have, or don’t have, negative frequencies by themselves. The distinction between positive frequencies and negative frequencies only exists in a complex-valued signal. “Take this real-valued signal and create an imaginary part with a phase shift of 90 degrees” is mathematically the exact same operation as “Take this real-valued signal, add an imaginary part that is all zeroes, then filter out all of the negative frequencies”. The presence of the imaginary part allows the signal to have an asymmetric spectrum, and specifically the 90 degree phase shift makes the negative half of the spectrum zero. > I also observed, by using the Frequency Sink that the Hilbert output > performs a 70 dB reduction of the negative signal using 125 taps. However, > when I connect a Complex to Float block to its output and observe the > frequencies, the negative frequency reduction is no longer present. I guess > I also don’t completely understand the functionality of that block well > either. When you discard the imaginary part, you get a real-valued signal, which no longer contains any distinction between positive and negative frequencies. You would get the same result (except for phase) by discarding the real part.
