There are no flaws in the simulation. In my first post on this thread, I said "For real signals time, this gives the conjugate symmetry of the spectrum. This symmetry doesn't exist for complex signals." If you look at real physical signals, they will automatically have conjugate symmetry in the spectrum. Period.
You have made the following statements in your posts and sheets:
"The purpose of the following is to create 1000 data point 1/f noise in both the time and frequency domains with no conjugate symmetry about the midpoint in the frequency domain."
"I suspect this would be another way of getting around the symmetry problem."
The first is simply not mathematically possible for a real-valued noise signal. The second is perhaps the root of the issue. As long as you see the symmetry as a "problem," rather than a characteristic that is inextricably linked to the DFT of a real signal, I don't think you will get out of the loop that you're in. The spectra of time sampled signals is simply not the same as those of continuous time signals. What you can see on a spectrum/network analyzer is not necessarily what you will get with a DFT (emphasize Discrete) simulation.
The OFDM simulation takes the spectra of complex-valued time signals, hence the lack of conjugate symmetry. However, complex signals require both positive and negative frequencies. The symmetry issue is lying low with white noise, but is more apparent with 1/f noise. If you truly want 1/f noise, then presumably you want to have the spectral density at -10Hz to be the same as the spectral density at +10Hz (strange 1/f noise otherwise). This is already the conjugate symmetry - mirror symmetry when reproduced near the sampling freq - that you labeled a "problem."
All the hoops of "...converted to the frequency domain, the first half of the complex data points extracted..." etc. are unnecessary. I think the core issue is your expectation that the upper half of the DFT should look like a "real" continuous spectrum that you would see on an analyzer. Address why you think the symmetry is a problem, and you will be able to move on to the simulation you are trying to run.
Lou
