I prepared an answer, but I see Dennis has already spoken. Dennis probably knows (and has forgotten) much more about speaker measurements and listening impressions than anything I've ever learned (mainly from him) to repeat
.
Here's my longer answer (that says less) anyway:
It's always hard to compare frequency response (FR) curves when they were done with different measurement rigs (microphones, software, rooms, and people). If you see differences, were they due to the speakers or the measurement methods? That's why the many speaker measurements published by Sound Stage (done by the NRC) and by Stereophile are so useful. You can believe that their skill, experience, and standard methods allow comparison among different speakers.
SoundStageNetwork.com | SoundStage.com | SoundStageNetwork.com | SoundStage.com
Reviews | Stereophile.com
Below 200 Hz, the NRC trace looks rather smooth, and the other trace (different speaker?) shows large peaks and valleys. When making measurements in normal rooms, the long wavelength bass reflects off of walls, floor and ceiling, and arrive at the microphone with a bit of delay compared to the sound coming directly from the speaker. Depending on the distance the reflected signals travel, these waves arrive at the microphone in or out of phase with the sound coming directly from the speaker. The direct and reflected sounds add or cancel, causing those large up and down excursions in the graph. Moving the speaker or microphone to different spots in the room will change the frequencies of the peaks and valleys.
At roughly 200 Hz and higher, as the wavelengths get shorter and shorter, this becomes much less of a problem. The NRC has a true anechoic room – little or no reflected sound. They measure the bass response (below ~200 Hz) separately from the rest of the audio frequency range response above ~200 Hz, and later splice the curves together. Stereophile does something similar. I’m don’t know what was done for the second graph you attached, but looking at the frequency response below 200 Hz, I’d guess it was measured in a room without any making corrections or splicing. So, ignore the apparent differences below ~200 Hz between those two FR graphs.