ITYM the opposite of what you said. There is an efficient constant Q fast transform that gives something vaguely useful for audio and studio band equaliser type displays. But it has worse resolution with increasing frequency in qualitatively the same was as the human ear.
There are techniques to get very high resolution frequency measurements from limited time sampled data provided that the signal to noise ratio is sufficiently high. Heuristically time between zero crossings on a simple waveform will give you a feel for why this is possible. To do it for an unknown signal requires some complex mathermatics and it isn't really amenable to realtime displays taking typically a couple of orders of magnitude more processing time. But if you are stuck with the measurements you can actually make and need the answer it can be useful.
Regards, Martin Brown