The toroid in LPA is normally shaped. The enclosure is rectangular because it is easier and cheaper to manufacture that way. A rounded enclosure would not make sense.

It would seem to make more sense for it to be square than rectangular - a rectangle's no easier to make than a square, plus it takes up more space and wouldn't cover a cylindrical toroid any better than a square. Obviously I could be totally off base, but it just doesn't look right for a cover on a toroid.

This was really bugging me, so I dug into Photoshop and did some experimenting. If we assume that the transformer is mounted upright (not lying flat), similar to the way that Outlaw mounts the dual toroidal transformers in the 7500 and 7700, then the width of the enclosure would be equal to or very similar to the height of the enclosure. Cropping all of the images (front, back, and top) to just the exact width of the unit (no more than a few pixels of white space to the sides) and resizing to 1000 pixels wide gives us a common scale for all three images. With this done, the width of the transformer enclosure is 410 pixels (just over 40% of the width). That's six pixels less than the height of the cropped and resized front panel image and two pixels less than the height of the cropped and resized rear panel image. Based on that, it would appear that there is a toroid - the cover and the depth of the chassis just make it hard to tell. I'd guess that the toroid is about 6" in diameter and probably 3" or so tall. I e-mailed Emotiva earlier this morning to get their input. They should be able to confirm this (three earlier e-mails relating to the LMC-1 have gone unanswered over the last month, but hopefully this will be a simpler thing to reply to).

Speaking of bugging someone, are these units you are comparing exactly identical size? Physically, not the picture.

If not, scaling the images in this way does not give you a common scale.

It may be a mute point for what you are trying to prove but I have seen many mistakes (including my own) made in this way.

The three images should be of the same unit, but even if they're not (even if they're off three different samples of the LPA-1) there should not be enough variation in dimensions to throw things off - although my guess of actual toroid dimensions inevitably have some margin for error.

Cool. I couldn't tell from the thread.
My sister is an architect/interior designer and she uses this technique quite often to do some virtual rearrangements and planning with her customers. It is a pretty neat trick because it doesn't matter how far the camera was, etc. (or even if you use real images for that matter). The only problem is you have to resize the images proportionally to the size of the object.
I designed an entire elementary school computer lab in this manner - that is except for that one station that I resized improperly...
Didn't mean to hijack, please don't mind me.

It sounds like both you and your sister are familiar with the adventures and woes of scale factors. Even in AutoCAD, I am forever answering questions about scaling - which is easy in comparison to images that have no dimensional data embedded in them.

I called and Emotiva confirmed that this is the case. I understand that it is cheaper to bend the metal into a rectangle than to shape it into a rounded enclosure, just as bugbitten has written.

Yep - the part that tripped me up was the orientation of the toroid.

Sorry to report that Emotiva is not so great. First and foremost, it has a crappy remote control...quirky default remote control behavior
No active IEEE 1394/Firewire interface
No HDMI/DVI Video Switching
No subwoofer distance/delay settings
* Subwoofer calibration test tone inaccurate
* No optional bass management / delay compensation on EXT Analog Inputs