You basically want to see something resembling the the original input square wave. Having recognizable straight sides rather than skewed looking sides (compared to the the square waveform) indicates good transient response which means the headphone sounds "quick" (can react to changes of sound frequency and amplitude faster). It's a bit academic though, I'm not sure if anybody has ever correlated specific sound quality attributes to measurements like this....but sometimes you can see a headphone or speaker really mess up the shape, which might help you back up listening impressions that a headphone just "sounds wrong" or "incoherent"
But what's going on with the Denon D7000? The bass is not as flat as the D5000 and there seems to be less bass extension. This seems to contradict what everybody is saying about the D7000 having better bass response and extension when compared to the D5000.
Am I reading the graph correctly? The trailing corner for the D7000 seems to be much lower than the D5000. I've compared it to many other headphones (DT880, K701, SR80, and others) and the trailing corner is lower than all of those as well.
A square wave is made up of all frequencies. The input to the headphones will not be a perfect square wave because, despite Tyll's magical abilities, all function generators have a finite bandwidth. However, it is reasonable to assume that the highest and lowest frequencies reproduce by the source extends far beyond that of the headphones under test (which really only operate over a very small bandwidth). So the reproduction of the square wave will be indicative of the linearity and extension of the headphone's frequency response. Still, I am not sure how reliable we would choose the linearity since headphones are not designed to have a flat frequency response since they are near field sources. How you measure the headphones and how you normalize the measurements would affect the frequency response.
Could somebody explain the square wave graphs?
