I am not aware of a Cadence app note, so in case I will be glad to take a look :)
However, there is quite some treatment on this and related topic in Kundert's designers-guide.org site.
This link is a good start : http://www.designers-guide.org/Analysis/PLLnoise.pdf
If I may say, if you want to characterize the *phase* noise, then PSS and Pnoise are what you need. Just feed your divider with a pure signal (whatever it needs, sin or square), and the Pnoise will give you exactly the output power spectrum of the divided waveform.
The 'strobed PNoise' is I think what you refer to, and was an indirect method based on sampling the noise of your 'edge producing circuit' only at the zero crossing times, and manipulating the PSD of this process (basically, jitter) into phase noise. Quite thick subject, I may add :)
Hope this helps,