Marco: It is not abnormal for the frequency response of a system ( especially speakers ) to change as volume is varied. As such, you could simply be "highlighting" a flaw in the recording and / or demonstrating system non-linearities. My guess is that you might be experiencing a little bit of both.
As a side note, you REALLY have to watch how speakers are rated and measured, especially in reviews. By measuring a speaker at a lower volume, it is possible to produce a flatter i.e. "better looking" response curve.
I just stumbled across a review a while back where the baseline frequency response reading for a $4K+ speaker was taken at an average of about 77 dB's. They tested the speaker in this manner even though it is WAY more sensitive than that with one watt fed into it. Why would anyone do such a thing? Well...
If someone wasn't paying attention or know what to look for, the results look much better than they really are. At this level, the speaker measured 18 Hz to 20 KHz +/- 3.5 dB's. To try and persuade the reader that this speaker was "really flat", they even went so far as to say that the speaker was +/-2.5 dB's from 35 Hz up to 10 KHz. This same article went on even further to say that "across the midrange, response tightened up even further, to be flat +/-1.5 dB's". If one didn't know better or really study the graphs that they displayed, this would look to be a pretty well designed speaker.
In another section of the same review, they drove the speaker well above this 77 dB baseline and charted the individual output of each driver. When doing an over-lay and splicing all of the composite curves together onto one graph, they revealed that this speaker had a huge plateau that starts at 40 Hz, hits a peak of appr +7 dB's at 110 Hz and then levels back out at appr 200 Hz. Needless to say, the linearity of this speaker went out the window, regardless of how smooth they tried to tell us it measured at ridiculously low listening levels. When the speakers' linearity went out the window at higher spl's, it also took any credibility that the reviewer had with it. Sean
>
PS... Anyone wanna guess what make / model speaker this was and what magazine reviewed it?
As a side note, you REALLY have to watch how speakers are rated and measured, especially in reviews. By measuring a speaker at a lower volume, it is possible to produce a flatter i.e. "better looking" response curve.
I just stumbled across a review a while back where the baseline frequency response reading for a $4K+ speaker was taken at an average of about 77 dB's. They tested the speaker in this manner even though it is WAY more sensitive than that with one watt fed into it. Why would anyone do such a thing? Well...
If someone wasn't paying attention or know what to look for, the results look much better than they really are. At this level, the speaker measured 18 Hz to 20 KHz +/- 3.5 dB's. To try and persuade the reader that this speaker was "really flat", they even went so far as to say that the speaker was +/-2.5 dB's from 35 Hz up to 10 KHz. This same article went on even further to say that "across the midrange, response tightened up even further, to be flat +/-1.5 dB's". If one didn't know better or really study the graphs that they displayed, this would look to be a pretty well designed speaker.
In another section of the same review, they drove the speaker well above this 77 dB baseline and charted the individual output of each driver. When doing an over-lay and splicing all of the composite curves together onto one graph, they revealed that this speaker had a huge plateau that starts at 40 Hz, hits a peak of appr +7 dB's at 110 Hz and then levels back out at appr 200 Hz. Needless to say, the linearity of this speaker went out the window, regardless of how smooth they tried to tell us it measured at ridiculously low listening levels. When the speakers' linearity went out the window at higher spl's, it also took any credibility that the reviewer had with it. Sean
>
PS... Anyone wanna guess what make / model speaker this was and what magazine reviewed it?