Input from multiple readers of this page led to this quick experiment. Since it is obvious that a rectangular baffle will emphasize particular frequencies in function of the width it was time to try something different. A trapezoidal baffle as an effort to break the geometry was mentionned a couple of times, so I sacrificed half a drywall plate for this purpose. This plate was cut alongside a diagonal, voila, two triangles, get rid of the tops and I had a nice trapezoidal baffle, measuring 10 "s at the top and 40 "s at the bottom, thus "average" width 25"s. The test was performed under same circumstances of the previous test, 3 meter distance from the microphone. The height of the microphone was changed between tests however. The numbers on the legends are 41, 51 and 61 inches, this translates to a listening height of 31, 41 and 51 inches. (My 9 year old needed a ruler of 10 inches at one time ...) Results are shown below, with listening height 41"s being the "flattest" response, close to the "normal" listening height in my place, also the level where there a lot of furniture ends and naked hard walls start. The fourth line was a measurement taken yesterday from the same vantage point but with a 25" rectangular baffle, causing a peak&dip display. As can be seen out of these plots, there is a significant smoothing going on as an "expected" result of the different geometry. Just in case you were wandereing, the gain levels were "cheated" by 10 dB to keep the plots of the two different baffles separated in the overlay. Curve smoothing was in all cases set at 1/3 octave, measurement technique was the same, quasi anechoic gated SINE with large reflections after 21 milliseconds gated out.