narrow and wide baffles and imaging

@twoleftears  there are clear explanation  how width of baffle
make sound or sound stage different , The problem coming from 
different diffraction lost from low frequency and mid freg. 
the call baffle step , you can read more on google
http://diyaudioprojects.com/Technical/Baffle-Step-Correction-Circuit-Calculator/ https://www.trueaudio.com/st_diff1.htm and much more.
Good design 3 way Loudspeakers usually dont have this issue. 
but one way crossoverless and some 2 way get this problem
result -pure soundstage, lack midbass, voices and piano sound  thin   

Also, Voxative  made one nice Loudspeakers with huge baffle
try to avoid this issue. The Good Horn design  also is good for
this reason and make  good  soundstage

The width of the baffle really only affects the baffle step. If the crossover designer knows what they are doing, there is very little difference. Having chamfered edges or a curved face (like the LS50’s) have a bigger impact than the width of the baffle (assuming equal efforts of dealing with the baffle step).

What can people tell me about Troels Gravesen?

I've heard from a speaker manufacturer that the main reason we're seeing so many ultra-narrow baffles is WAF -- narrow speakers can be shoehorned between the furniture.

Otherwise, I think it's a tradeoff. As Alon Wolf said, a larger baffle reduces the baffle step effect, something that can never be completely corrected. On the other hand, a large baffle causes diffraction issues that are known to interfere with imaging and can also block room reflections that add to the sense of space. In any speaker, these issues can be minimized by for example curving the edges of the baffle and using felt.

Interestingly, though, the most holographic imaging I've ever heard was from the IRS V, at Lyric. Which of course had a huge baffle, curved to minimize diffraction effects. As I recall, the room was also quite dead, which tends to give you pinpoint imaging at the expense of spatiality (width and depth) and a sense of ambiance. And it was being used with tube electronics (CJ) that were known for enhancing depth (out of phase second harmonic distortion).

Otherwise agree with those who say to get the electronics out from in behind the speaker -- the rack will cause a reflection and create a virtual speaker that confuses imaging.

One of the most important factors in imaging seems to be the distance from the speaker to the walls -- you want ideally to have no early reflections greater than -20 dB for the first 25 ms or so, because the ear uses first reflections to judge the size of a space and early reflections from the room mask the recorded ambiance of the original venue. In all but very large rooms, some acoustical treatment (absorption or diffusion) at the early reflection points is needed to do this.

You also have to consider your distance to the rear wall. I've found that if you're too close to the rear wall, rear wall reflections interfere with the imaging. Since in my own undersized listening room I have to sit a few feet in front of a wall, I put some Owens Corning 703 on it and the imaging problems disappeared.

Another imaging rule that I've found useful is that good mono makes good stereo. If you listen to a mono recording or pink noise it should be pinpoint centered between the speakers. That will give you palpable, pinpoint stereo imaging. Asymmetries in the listening room and setup can cause the sound to wander with frequency. I recently put a sheet of plexiglass in front of a window behind my right speaker to flush it out with the wall, and I was astounded by the improvement in imaging. I didn't think a couple of inches would make such a difference, but it did.

It's amazing how much of a difference small changes to the room can make. For example, I have a computer monitor off to the side, and when I angle it so that it faces the rear wall and doesn't reflect the sound from the speakers to my ear the imaging improves. A Umik-1 measurement mic and the impulse response display in Room EQ Wizard are great tools for catching early reflections.