I understand that the Siglent wins miles from the Rigol if the low-end model can be "upgraded" to the high-end model
And if it stay 2.1GHz (SSA3021X) it still outperform Rigol clearly if we compare DSA815. SSA3021X and SSA3032X are just same in all things exept frequency.
In all other things they are same, not like low-end and high-end if think how we normally classify things to "low-end", "high-end" and "state of art" categories. All these spectrums, whole DSA800 serie and SSA3000X series are in "low-end" category if we talk stand alone table top spectrum analyzers. Bit, there is "low-end" category SSA3000X is in top corner, far over DSA815.
Just only one example:
Due to low near carrier phase noise down to < -90dBc/Hz @100Hz distance from carrier (-91dBc/Hz in this image including RBW normalization to 1Hz and then +2.5dB correction for 1Hz noise BW) you really can use RBW10Hz
You can not do there just anything with DSA815. Same if you try put cow to fly.
This is standard SSA3021X out from factory box. Trace A, signal from HP8644B free running with its own internal reference and SSA free running with its own reference. Trace B signal from SSA own 10MHz reference out. (this is from very different work but it can use also here as example.)
In this image trace C is also on but not visible but it is there if move screen to more down, C is there for detect displayed average base noise level (in this case -127dBm/Hz)