Hi everybody,
I don't know if this is the appropriate section for the question, but I post here.
I'm reading about digital modulation (in particular on Lee-Messerschmitt's Digital Communications) and I have this picture (RX.jpeg attached). These three structures are equivalent: so, at the end of each structure, I should have the same spectrum.
I tried to think what can happen to the spectrum of a received signal in the third case (analytic bandpass filter with baseband spectrum and I tried to design some spectra shapes ( like they do @pag.13 of this article  http://www.eumus.edu.uy/eme/cursos/dsp/material/Lyons_Quadrature.Signals.pdf ) .

In the case with a complex filter I found those spectra pictured in the attached figure RXchain.jpg, but if I multiply the Q-path signal by j and then sum to the I-path signal (like they do in the paper of Lyons) I DON'T obtain the same signal that I have from the structure with two mixers+low pass filter.

So, here my questions:
1) First of all, why the spectrum of quadrature modulated signal have the two sidebands asymmetric? How do they create such signals?
(I have found this paper  http://www.sato.gits.waseda.ac.jp/file/77a  where in Fig. 1(a) they represent some baseband spectra QPSK but I don't know if they are correct..)
2) Why the receiver chain with the complex filter doesn't give the same result as the other two structures?

Thank you if you can help me!

Nicola