But the rates are the same regardless of memory setting of the scope - in other words, they are linked to the displayed samples (like the Agilent)?
As I have told, there is (1): 1 captured waveform is horizontally equal to display area (14 div). There is not invisible part of waveform outside of display horizontal area. What is captured is also visible (less blind time) Also as in published table, example with 1ns/div and 2GSa/s capture length is 28pts.
(1)this is now with this FW what is not at all ready. (it is fully possible there is coming (or not coming - who knows) feature where user can also select if he want minimize "blind" time or select long capture also with fast horizontal times. But, in real, it do not give so much, becouse 1ms/div there is full 2GSa/s and if want long capture but detailed look inside some small part of capture of course zoom can use. Example 1ms/div and details with 1ns/div if need.
If look example oscilloscope where is 10k minimum memory and 10 horizontal div display. If there is 1GSa/s sampling and 2ns/div speed it means that displayed visible area have 20 sample points and invisible there is 99980 points (capture is, in this case, 5000 div but visible is only 10 div. Alone this give visible/invisible time ratio 1:499. and then of course time between every single capture and not visible/visible ratio is - as we know - terrible. But if is is acceptable or not, it depends what user is doing with scope.
(why capture something what can not see..
![Wink ;)](https://www.eevblog.com/forum/Smileys/default/wink.gif)
)