This article should list some more specifics when it comes to the frequency response portion. For example, what microphone was used for the test? A microphone is very similar to a speaker and each has its own frequency response curve and almost zero of them are perfectly flat... I've only seen one ever that's supposed to be like 99.9% flat, and it's a large awkward mic used for applications like earthquake studying. Did they use a Shure SM-57 or a cheezy computer 1/16" plug mic.
Also, perhaps some room information would also be helpful? Sound from speakers hits walls, reflects and when the waves meet and interact, things additions and subtractions in dB occur. Essentially, peaks and valleys occur at many various listening positions and frequencies throughout your room. Higher Frequencies have shorter wavelengths and can attentuate before hitting a wall even, however, it's the long wavelengths of low frequencies that causes issues. So, the best way to handle this is to use bass traps (not studio foam) to control low frequencies in a room. Diffusors can be added to scatter waves in more random directions and high frequency absorbers in the first reflection areas of sidewalls and ceilings. We're not talking about studio foam either. If you want more information like this, see realtraps com for faq.
OK, so, assuming nothing like the above has been done, that COULD explain why the recorded frequency responses all have a dip at 100 hz and all have a spike at 15 khz
and all have extreme roll off in the 20 hz area. Obviously, PC desktop speakers are limited, and any non subwoofer is going to have problems beginning anywhere between 100-250 hz and all the way down. With this in mind, we can really only compare these speakers to each other, considering the environment. It's possible that while the MM-1 has the best bass response of these four below about 80 hz, it has the worst dip at 120 hz. If this room had a design that caused natural dips at 120 hz, well, then that explains why these speakers have dips there, where the Gigaworks T40 and Studiophile AV 40 would appear, still, to have better curves at those frequencies, but, it could be that this room has a -15 db dip at that frequency. If THAT were the case, then, the MM-1 would still be pretty flat there and look good, while the others would have spikes there. So, I'd be interested to know a little more about your room environment, mic choice, etc.
Also, many speaker and microphone manufacturers show an advertised frequency response curve on their websites and/or in the user manuals.
Also, this statement:
"M-Audio might not be an instantly-recognizable brand for consumers" should probably be replaced with "M-Audio might not be an instantly recognizable brand for the author of this review" because M-Audio has been making and selling speakers a long time, and primarily for DIY sound recording, as well as many other various musical products. Anyone who's looked through Musician's Friend magazines for the last few decades or played music and even thought about recording it, or possibly have been to a music store, is likely to have at least seen or heard of the name M-Audio. Trust me, there's a lot of musicians out there. While M-Audio speakers are generally more for budget studio applicatons, they more than likely have a pretty recognizable name.