Mike Harrison of mikeselectricstuff gave a talk at the 2016 Hackaday Belgrade conference on how insanely complex these things were.

https://www.youtube.com/watch?v=3-BvMcqEc98

He mentions the Eidophor's only competitor, the GE Talaria, which is also interesting:

https://en.wikipedia.org/wiki/Talaria_projector

I am quite impressed by the image quality and the sharpness for something that was used in the 60s. Of course, digital screens are better now but whoever invented the Eidophor at that time did extremely well.

I remember "live" projections of a TV video signal used in BBC programmes in late seventies, where you see on the stage a huge screen and seemed that it's not a film projection, and at that time I just couldn't imagine how they managed to do that, knowing only how the CRTs worked -- and big projections simply "didn't fit" that model. I considered that a pure magic.

And now searching the internet, they really did have this technology (see the comments):

https://hackaday.com/2016/03/15/retrotechtacular-eidophor-an...

"When I worked at the BBC West London studios in the 70’s we regularly used Eidophors for back projection"

Fascinating. I always wondered what was used in the 1960s to project the huge screens in NASA Mission Control.

Since it uses a "slowly rotating disk" for the projection, what was the latency of the technology?

The rotating disk is only there to replenish and renew the oil surface on which the electron beam is drawing the image. The speed of disk rotation has no effect on the latency.

It does have an effect on persistence of the image, but my understanding is that the disk was slow enough that movement was negligible compared to how fast the oil deformations faded.

Eidophor has practically zero latency, same as a CRT - whatever the electron beam draws is immediately visible.

From http://www.earlytelevision.org/yanczer_eidophor.html they could run at 120 Hz.

I wonder if you could do digital holography with this?

not sure about holography, i think one can imagine how an electron beam modulated surface (or may be volume) can say change laser direction thus providing kind of non-mechanical scanning in a lidar for example. Though it sounds like a last century approach. These days i'd bet on something like nanotube channeling the beam and say flexing in response to the field.

I wonder what the advantages / disadvantages were compared to CRT projectors that came out in the 1950s.

From the article: the Eidophor was 80 times brighter than CRT projectors of the time.

Presumably because it was a modulator of an independent light source, rather than the source itself.

Doug Engelbart's famous 1968 demo used an Eidophor.