I seem to recall you mentioning in a post over on SNA that accurately measuring Jitter below 10ps was an issue and you need special equipment to do it, with not a lot of it being about. I seem to also recall you mentioning for your converters you actually didn't measure the jitter directly but rather 'deduced' it from the jitter measurements of the clock.
I have measured worst-case direct P-P jitter over all frequencies and its in the specs for the OR5. I don't deduce from the clock or quote the clock like many other manufacturers do.
I have only recently begun using the histogram, which I am finding very useful. Even if the max deviation in the cycle-time, (which is the worst-case jitter event) is 400psec, these events can be fairly rare. The bulk of the jitter can still be in a peak about 50 psec wide. This is what I'm seeing now with this mod. Now I am convinced that not only is the spectra interesting, but the shape of the distribution of the jitter is maybe even more important.