It is literally impossible to hear a subwoofer in a room without hearing the room. By the time your ears have heard enough full cycles to detect the pitch, those long wavelengths have been around the room several times. Point being, it's always subwoofer + room that we hear, never just the subwoofer.
So what we want is a sub that synergizes well with whatever the room is going to do to it. Assuming we're talking about a single sub (there are major advantages to using multiple smaller subs, but I don't think that's on the menu here), probably the best we can hope for is a sub whose native response is the approximate inverse of room gain. That way, sub + room = neither broadly rising low end, nor broadly falling low end. We'll still have major room-induced peaks and dips, but that's another problem for another day.
Neither front-firing nor down-firing have any significant advantage as far as room interaction. Box and woofer being equal, the air mass underneath the cone will change the effective woofer parameters a bit, resulting in a slightly "fatter" low-end from the downfiring iteration, but that just means the designer would start out with a different woofer and/or different enclosure in order to meet a given target.
Back to subwoofer performance targets for a minute. It's been mentioned that sealed subs are better for high quality sound, and the presumption is that that's due to their superior transient response. But the ear's time-domain resolution at low frequencies is extremely poor (which is why we can't localize the source at long wavelengths - our two ears don't detect the arrival time differences down there). It's not really the transient response itself we are hearing; rather, it's the in-room frequency response (something our ears are very good at hearing). Vented sub + room often results in a rise in the deep bass, which sounds "slow", whereas a sealed sub + room tends to have a gentle falloff in the low bass, which sounds "fast".
You see, typical broadband room gain due to boundary reinforcement is approximately 3 dB per octave. So, based on the room + sub paradigm, we'd arguably want the sub's inherent output to fall gently at 3 dB per octave as we go down the bass region. This way sub + room = roughly "flat" (aside from room-induced peak and dips). Vented cabinets are often tuned for deepest loudest bass, so the room's +3 dB per octave tends to make them sound boomy. Sealed boxes typically start rolling off higher, but it's a gentle rolloff, often between 6 dB and 12 dB per octave, so it synergizes better with the room.
As proof that it's the sub + room, and not the sub itself, haul both subs outdoors. Now the sealed sub sounds weak and anemic, but the vented sub sounds faster than it did indoors - now the vented sub is clearly the qualitative winner. These are generalities, of course - there are exceptions.
Imo vented sub technology has the greater potential for good in-room bass because a vented sub can be designed to gently roll off by approximately 3 dB per octave from 80 Hz down to 20 Hz (before room gain), whereas such performance is not really possible from an unequalized sealed box.
Sorry for the long digression, and I know that's not the question that was asked. The imo theoretically ideal 3 dB per octave native rolloff can be realized with either a front-firing or a down-firing sub. So it's not driver location itself that matters, it's the overall system design and how it synergizes with the inevitable effects of the room.
Imo, ime, ymmv, etc.
Duke
dealer/manufacturer
