I wonder if Tyll (when he's around), Xnor or another member can tell me a few things about these measurements. Firstly, how is the raw frequency response data modified? By the looks to me, it might be based on Moller's diffuse field response? If not, do we have a name for and data of the modification applied?
Secondly, is the square wave response taking this into account or is it based on the raw output of the phone? I ask this because there is evidence of undershoot and ringing in the LCD-2, as well as a slope down to 0. Both of these suggest a. a boost in treble frequencies and b. insufficient bass response.
In the raw data for the frequency response of the LCD-2, it clearly shows a treble boost (which is necessary), relative to bass. However, in the compensated data it shows a clear roll-off.
Just looking for clarification, thanks.
If my suspicions are correct, is it possible to use a fourier transform to convert the compensated response to a square wave response?
I think that the compensation is based on measurements with the artificial head in a diffuse field. Somebody has written similar in this thread.
Your second question is very interesting. I too would like to know the answer.
But other than identifying some overshoot or ringing I don't see much use in the square wave measurements. I think you'd have to correct both the frequency and phase response first to make it more useful. And you need the same information for FFT matters.