Hello Chairguy,
I'm a fan of what omni-speakers do well, though I don't actually sell any omnis at this time.
Let me give a brief overview of why I think omnis sound good, as the psychoacoustic principles behind it are imho at the heart of good loudspeaker design.
The goal is to recreate, as closely as possible, the perception of being present at a live performance, and the reverberant field is critical to this. The ear finds a late-arriving, diffuse, well-energized, spectrally correct reverberant field to be rich, warm, lush, and very pleasing - enhancing timbre and giving a sense of ambience, spaciousness, and envelopment. That delicious lushness we hear in a good recital or concert hall is almost entirely the result of the reverberant field's characteristics. This is where omnis shine. In my opinion, doing a good job with at least the spectral balance of the reverberant field is an important aspect of any loudspeaker designed for anything other than nearfield listening.
Moving back into the home listening room, unfortunately not all reflections are good. In general, reflections arriving earlier than 10 milliseconds after the first-arrival sound are likely to be detrimental to imaging and timbre. This roughly corresponds to a path length difference of eleven feet. Now in a large enough room an omnidirectional speaker will not have early sidewall reflections arriving within this time window, but in a smaller room it's likely to happen. Most speakers in most rooms will have floor and ceiling bounce reflections arriving within that time interval (line source speakers being an exception).
Note that most speakers generate a reverberant field that has a significantly different spectral balance from the first-arrival sound due to radiation pattern issues. So there are a lot of speakers out there whose tonal balance goes down hill the larger the relative contribution of the reverberant field. Omnis, however, are very good to excellent in this regard (varying somewhat from one speaker to another - some omnis are more omni than others).
Line source dipole speakers (Maggies, SoundLabs, Apogees, etc.) are an especially elegant approach. Their inherent directivity in the vertical plane avoids the early floor and ceiling reflections, and if they can be placed six feet or so in front of the wall their backwave energy arrives late enough that its contribution is almost entirely beneficial (I still like to diffuse the backwave a bit). The early sidewall reflection may still be an issue, depending on radiation pattern and amount of toe-in.
Likewise, a bipole will generate a rich reverberant field. In fact, my first commercial system was a bipolar. Bipolars have some interesting characteristics and challenges, and several variatons have been tried. Although not very many bipolars have made it to the marketplace, a fair amount of innovation has gone into those that have.
Of course there are even more variations among omnidirectional systems, and most are not really true omnis. Even those with uniform 360 degree horizontal patterns (Duevel, MBL, Wolcott, Morrison) have vertical directional characteristics that can and do impact the power response and therefore the spectral balance of the reverberant energy.
I particularly like Richard Shahinian's word, "polydirectional". That word includes omnis, bipolars, dipolars, and his own carefully-chosen multi-directional configurations. Imho Richard's polydirectional speakers are very intelligently designed to work well in most people's rooms. They address the early reflection issue by being very low to the ground and bouncing a lot of energy off the upper walls and ceiling. In a small room, this increases the time interval before the arrival of the first reflections. If I wasn't pumping my spare nickels into my own line, I'd probably be a dealer for him.
The speakers that I currently build get the spectral balance of the reverberant field right and are designed to be listened to off-axis so they aren't so much "in your face", but some people (and some rooms) prefer more energy in the reverberant field relative to the direct sound. This is something polydirectionals do better than my off-axis monopoles. So I have recently revisited the polydirectional concept, building and auditioning several different configurations using identical drivers. One of the configurations I tried was clearly preferred by my own ears and also by my portable analog real-time spectrum analyzer, who doubles as my wife. So at some time in the not-too-distant future I'll probably be adding one or two polydirectional types in my lineup.
Duke
