My photoshop skills are bad but below is an overlay of graphs posted by Craig. These are measurements that he took of his 340 SE in his room in response to me taking issue with the passive level-matching switcher which was used. The first graph is with the speakers behind the screen with no attenuation (switcher being used but no padding). The second is with 2dB attenuation and the third is with 4dB attenuation.
I normalized the graphs to keep the relative sensitivity levels the same between 1 kHz and 2 kHz. It is fairly easy to see the effects of the inline resistance (attenuation).
Only 2dB of attenuation caused a 4dB drop in the 6 kHz + range relative to the 1 kHz – 2 kHz range, and a 3dB drop starting from about 2.5 kHz. Additionally, the graph also clearly shows the beginning of the dramatic midrange dropout starting at about 700Hz and lower (as indicated on the graphs I posted). I must say once again that this would constitute a dramatic change in the character of any loudspeaker.
A "neutral" loudspeaker will now have its mids and highs recessed, sounding very laid back, a term that was indeed used to describe many of the higher efficiency speakers. A bright or overly forward sounding speaker will now sound more neutral.
It is also important to note that Craig's frequency response graphs clearly show the nasty effect of any attenuation and these effects are identical to the measurements I originally posted. At no fault of Craig's, his graphs do not show the change in filter slopes and Q (which is what my electrical response graph reveals) which also have a dramatic effect on loudspeaker performance, even though these effects may or may not show up in a simple frequency response graph. For example, even though an on-axis graph between two different speakers may look exactly the same, they will not sound the same because the crossover points are different and the filter slopes and Q are different. It is the shape and the "blending" of these crossover filters that precisely control phase and off-axis response....
I must additionally stress that even 1 dB of attenuation will result in similar performance degradations.
In my opinion, it is a bit silly to go back and forth with this -- the proof is right there. Regardless of which speaker is being claimed as the more sensitive -- even as little as 1dB of attenuation for any of the loudspeakers changes the performance of that particular loudspeaker. It is simply not a valid way to compare 2 different loudspeakers. That issue is not subject for debate.
