That "hump" might look awful to you, but it is nothing compared to what happens in the 8kHz-20kHz in the on-axis curve from the spinorama.
Anyway, the on axis curve is only a (minor) part of the whole picture here, and the EQed curve was the listening window average.
Here are my gated measurements (correct settings in red, you can ignore the blue curve), 0° and 20°
http://www.audioheritage.org/vbullet...l=1#post393721
You can clearly see that the bump around 8-9kHz results in a flat curve at ~30° off axis (which is probably a good approximation of the listening window curve...)