Plug-ins for 5xxx and 7xxx Tektronix scopes. If you can't do better, you just might as well abandon hope of measuring noise. The scope used by Jim Williams (RIP) was capable and 1960's vintage!
Analog Devices LT1113 dual JFET opamps, a pair (i.e. 4 in parallel) can do better than this with very high input impedance. With respect to the Siglent it is a very simple matter to build a 40dB low-noise preamplifier ahead of the input.
I would wager those photos were made with a Tek 547 with a 1A7 plugin https://w140.com/tekwiki/wiki/1A7 , 10 uV/division. The 7a22 https://w140.com/tekwiki/wiki/7A22 also goes to 10 uV/div. They are really to help illustrate the low frequency noise component. Midband noise is better measured using spectrum analysis, preferably with a noise corrected level indication. Noise and steady state signals have different properties that make amplitude measurements misleading if not corrected and a suitable averaging time provided. This is a more complete explanation https://www.edn.com/dsos-and-noise/
