My biggest gripe with panels would be a full frequency spectrum [no omissions] with the pin-pointability of the spatial locational cues, along with the precision of the image. Without going into detail, please allow me to explain it this way. If you have a guitarist. He is strumming the guitar. He is picking certain strings as he strums, while at the same time, he is changing notes on the neck of the guitar. The illusion or the perceived image I have with what I have is that I can place each in its own 3-D location [to a point and as always, recording dependent]. To me, with stat's and ribbons, what I get is one guitarist that may be three feet wide. Yes, this produces an image that is spatial but I'm looking for the spatial locational cues within the image itself. A 3 feet wide image of one guitarist? That is not what I am personally looking for in reproduced music.
Tom, that might have to do with set-up. Stats are dipoles, so are Genesis, Magneplanar, etc. The set-up rules are different, and you have to be careful. It is both easier and harder at the same time to set up a pair of dipole speakers. However, I've heard Martin Logan CLS's, Maggie 20's, and large Acoustats deliver the guitar image as you desire. However, this is greatly dependent on the image that was recorded on the CD.
The problem I've found is that most owners set up their speakers in relation to the listener, instead of in relation to the room. This is more important for dipoles than the usual box speakers.
If you look carefully at the pictures posted by MrAcoustat, he has a room that has an angled wall on the right, and an open space on the left. His speakers are also set up asymmetrically which is brilliant. To handle the angled right wall, he has the speaker toe-ed out slightly. This manages the rear wave of the dipole speakers, and if I am not wrong, would give him superlative pin-point and real-sized imaging and soundstage.
The other problem with most stats has to do with integration between the midrange and the bass. How do you get a 100gm woofer cone to move at the same speed (I know bass "speed" is hugely debatable) as an almost weightless ribbon or stat? Ever since Martin Logan implemented servo-bass technology in the early 2000's the integration have been very, very much better. Other speakers have relied on a large enough radiation panel (Soundlabs, CLS, etc.).
However, with any dipole, the wrap-around cancellation frequency is an factor and no matter how wide the panel is, it is difficult to get the cancellation into the bass. Some companies - Emerald Physics being a recent notable one - managed this using digital equalization. In the Genesis 1, we manage this using a 4-foot wide wing (140Hz cancellation frequency). However, this 4-foot wide wing means that most music lovers, even if the have the money and want a pair, don't have the space for a G1. I couldn't even hope to get a pair of G1's into Steve's room which houses the Alexandria and four JL Audio subwoofers.
This guy paid for the development and engineering to shoe-horn a pair into an apartment that had only 5.2m width (but 18m length). Even though I did the re-design with narrower wings - which meant that I had to change the crossovers - I was never sure that it would be a success and could be called my flagship speaker until I had them installed.
When Amir heard them in the Genesis factory (6,000 sq.ft. no walls) they were sounding small and lightweight. The image specificity was very good (a cello that was recorded to sound cello-sized sounded cello-sized) but the soundstage while it extended outside the speakers sounded small and diffuse. However, installed in the room they were designed for, they were absolutely spectacular.
After we installed them, the wife insisted that they clear the projection screen and the wings were moved another 1.5m apart. However, we still managed