Many interesting points to discuss - maybe a little off topic but the O.P. seems to have been a hit-n-run anyway.
So... Here's my SPD / phase noise story.
A few years ago I had the opportunity to assist with the creation of an RFP for a state-wide public safety system. Our group obtained the TIA documents to learn about delivered audio quality (DAQ) and coverage acceptance testing. These documents also contain the SPD comparisons of various modulation formats which I found to be very interesting.
It got me thinking about D-Star which Icom was promoting as a 6.25 kHz format. I wondered what D-Star would look like if evaluated using the TIA method. The TIA documents contain instructions on how to create the diagrams so that manufacturers can submit new formats.. I thought - I can do this! How hard can it be? Just follow the instruction, connect a radio to a spectrum analyzer, acquire the data, put the data into the TIA software - ta-da.
Then I realized that you can't test an 'unknown' and trust your results unless you first test a 'known' and achieve results identical to the published results. That is the only way can you have confidence that your instrumentation and your procedure are going to produce accurate results.
OK. Just start with one of the published modulation formats. Still easy. Not!
First, the procedure is rather involved. Once I got past that complexity to a finished diagram, I realized that my Agilent ESA-E wasn't up to the task. I (correctly) concluded that phase noise was causing the problem. OK. I need a better analyzer. So I used my R&S FSP. Better but still not good enough. So then I borrowed my employer's new CXA. 20,000 USD box. What? Still not good enough!
What to do? At that point I decided that the TIA tests must have been performed with something really good like a PSA. Wow - those are expensive. Out of my league (still are)
Then I remembered having seen some discussion on-line about the E4406A. Sure looks like a PSA. The datasheet says: "The Agilent Technologies E4406A vector signal analyzer (VSA) is a full-featured transmitter tester. Isn't that what I want to do - test transmitters? The phase noise graph in the datasheet looks pretty good. For those selling prices - what the heck - I'll buy one. Good decision! Now my results for testing published formats compared well to those published by the TIA.
Look at the attached comparison image. The signal source was the attenuated transmit output of a Motorola APX portable transceiver operating in P25 Phase one C4FM CAI mode. The exact same transceiver was used for both tests. The only difference is the analyzers.
The CXA result on the left shows much higher sideband noise than the VSA on the right. This causes a serious error in the ACPR calculation.