I don't have an easy way to generate a known-clean signal that low - Rigol DG4062 will do 1mv RMS but no idea how much noise may already be on it.
If you have a 24 bit sound card in a PC the (free) ARTA audio testing software will do the job:
http://www.artalabs.hr/You will loose bits as the PC volume slider is decreased, but still will probably be above the 16 bits of the high-res scope mode. If it won't go that low before cutoff with the PC slider a resistive divider would be OK, just stay above 5K or so total to keep from loading whatever the sound card is using for an output chip. I know that is a lot of set-up stuff though and probably not on the schedule any time soon.
I'll do the test when I eventually get my RTB2004's, in May now they say. I have a ultra-low THD 1KHz audio oscillator, the oscillator in a Keithley 2015, and the oscillator in a QuantAsylum QA401 audio analyzer (which is essentially a 24 bit sound card in a box).
BTW, in case folks don't remember from previous posts, this is ci11's request from back in post 450. He and I are audio guys, so the ability of that HMO1212 in high res mode to trigger and display the 20Hz 400uV waveform is pretty good stuff.
The theory here is that the RTB 2004 should be able to do at least as well, probably better, since it seems to have HAMEG DNA, lol.
Rich could easily do the test too, if he should happen to have access to a R&S UPV. It can generate 20Hz at -100dB THD+N.