I really think chasing the deep bass response like this is a waste of time. Wavelengths are huge. 30Hz is about 10 metres. It is impossible to remove the effect of the room. The room becomes part of the speaker at these wavelengths.
The different measured shapes of the response make this somewhat clear. A ported speaker is a 4th order system. No ifs, no buts, it is 4th order. There is some wiggle room with the parameters of the system, and the QB3 (quasi-butterworth 3rd order) looks a lot like a 3rd, and has been a popular alignment. But no matter what, the bass response curve will always eventually show a 4th order rolloff. If it doesn't in measurement, it isn't the speaker that is the problem. Designers can choose the Q of the system, usually from a limited set. Many will go for maximally flat, Butterworth aka Q = 0.7, lots go for something with a bit of dialled in extension, so Chebychev, Q = 0.8, or even more, with Q's up to say 1.2 with a clear artificially accentuated bass hump. A few will go for low Q's with a lean bass buy chasing some ideals about improved group delay, say Q down to 0.5. All of this stuff has been known for many decades, and not much has changed.
Designers don't have a lot of room to play. Physics is a harsh mistress. Unless the designer messes up (and some do) the bass response really is predictable from basic physics. Given the box and port dimensions, and the speaker's mechanical and electrical parameters (collectively usually terms the TS orTheile-Small parameters) you can predict from first principles the free field bass response. Designers can mess up in only a few ways. (And these can be predicted from the basic physics as well, just with a bit more effort than just the simple 4th order equations.) But any competently designed speaker will have an iron clad deep bass response that varies very little from the very simple 4th order response.
The F208 and 228Be use almost the same bass driver. The SB 23NBAC45 in the 208 and the SB23CACS45 in the 228Be. (Harmon could commission custom parameters, SB are more than happy to oblige, but I somewhat doubt there would be much point here.)
These are the parameters for both drivers. Differences highlighted. Where different, the 228Be driver is first.
- Nominal Impedance 4 Ω
- DC resistance, Re 3.3 Ω
- Voice coil inductance, Le 0.27 mH
- Effective piston area, Sd 216 cm2
- Voice coil diameter 45.5 mm
- Voice coil height 19 mm
- Air gap height 6 mm
- Linear coil travel (p-p) 13 mm
- Magnetic flux density 0.86 T
- Magnet weight 0.8 kg
- Net weight 2.7 kg
- Free air resonance, Fs 23 Hz
- Sensitivity (2.83V/1m) 90 dB
- Mechanical Q-factor, Qms 6.0 vs 5.4
- Electrical Q-factor, Qes 0.35
- Total Q-factor, Qts 0.33 vs 0.35
- Moving mass incl. air, Mms 33 g vs 32 g
- Force factor, Bl 6.7 Tm
- Equivalent volume, Vas 95 liters
- Compliance, Cms 1.43 mm/N
- Mechanical loss, Rms 0.8 kg/s vs 0.9kg/s
- Rated power handling* 60 W
Basically the very slight difference is down to the slight difference in effective cone mass - 32 vs 33grams. The other differences are a direct consequence as they include effective mass in their definition. (The surround termination is likey a different design to account for the different cone properties, this may be reflected in the mechanical loss.) The final result that matters is a tiny change in Qts. Critically, Vas and Fs are unchanged. Note that things such as linear coil travel and the like are identical. The speakers use essentially the same motor and have identical power handling and ability to move air.
Where one might expect a difference is in a barely measurable difference in bass alignment due to the different Qts. This can be trivially tuned out with either changes to the crossover resistance and maybe be a small change to the port length.
Differences between the speaker's bass response will be extraordinarily difficult to find. Any difference in measured results is almost certainly an artefact of the way the measurements were done. And these measurements are hard, simply because, as above, the room couples to the speaker at large wavelengths. Even jamming the microphone against the cone and sticking it in the port can't fully avoid these issues. If a room mode coincides with the location of the driver, it can affect the measurements.
This is especially important when looking for such defining parameters as the 3dB point of the speaker. There is little chance it can be reliably identified from an in-room measurement. Indeed the entire shape of the roll off is difficult to reliably estimate.
Overall, the true deep bass character of any speaker is governed by a few simple iron clad parameters. Choosing a higher Q than 0.7 can lead to a more emphasised bass, and there is little doubt that many bookshelf speakers do this (all the way up the the infamous LS3/5a). Some of this is taste. A zillion years ago it was noted that there was a bit of a cultural divide, with American speakers (and we assume consumer taste) going for higher Q values, and British speakers going (for the most part) for the more accurate but leaner Q values.
What one also notices is that many people ascribe qualities to the bass that are not part of the true deep bass response. Bass slam and impact come from much higher frequencies. A goodly dose of harmonic distortion in the bass might be part of this, and lead to perceptions of better bass. That might be more measurable.
ETA.
Eventually bass comes from moving air. And for deep bass, lots of air. No amount of messing about can change this simple rule. You can make any speaker go as deep as you like. But at the cost of maximum output. In the limit, in-ear monitor earphones have fabulous bass. Deep loud bass means you must shift lots of air. Even a small extension in bass roll-off frequency can mean a quite noticeable drop in maximum output.