Not really a new feature, but expanding a current one: Spectroid allows to plot the
max level of the signal reached at every frequency.
It would be great to have the possibility to choose the mean (either from the recording start or adaptive to the last recorded N seconds). On continuous signals where a particular section is not relevant (e.g. background noise analysis, noise tests, voice recording,...) this would provide a time averaging profile that would smoothen the signal profile providing a more realistic view of the sound being analysed.
Regards