You could always try doing it the original way, that was a slower manual two step process, without any GUI:
- So you would collect power at each frequency into to a flat CSV text file over a number of minutes/hours/days, depending on your number of runs and integration time on each, using something like https://github.com/xmikos/soapy_power (there are others).
- And then you would render that collected data into an image that a human can see and actually understand: http://kmkeen.com/rtl-power/ (scroll down to " Rendering"), as long as the selected output file format is compatible any of the rendering tools should work.
At a very minimum it may help you narrow down if the problem is with soapy_power , with qspectrumanalyzer, or even an unexpected parameter being provided.