Hi Mark. I enjoyed watching the video and explanation of your design/project. Especially as I've been tackling many of the same issues myself with my Arduino televisor.
Things I found interesting/unusual/worthy of comment... in no particular order...
a) you mentioned frame rate for the Baird - 18fps . I never realised this was the case until now - I had always assumed the "club standard" 12.5fps was modelled on Baird's system but it appears not to be so. I have to find out why we're using 12.5fps. Odd.
b) It was interesting to me you are using a single hole for frame sync. It wasn't clear how you achieved that (that is, getting the first scanline in the right position). I assume your Pi software is (re)starting a frame when it sees the synch hole, rather than running independently and hoping that the disk is up to speed at some point. That is, do you detect a synch pulse, then start "sending" the image?
c) I thought the 512x512 GIF was both overkill and a significant limitation. Probably fit for purpose, but it would be nice to have period-clips (consider, for example, Don MacLean's restorations of actual 1930s TV recordings). I converted these to my own format - "NBTV8" - for display on my Arduino televisor. I did contact Don for permission, which he kindly gave, but also he seemed to be very positive to allowing these restorations/digitisations to be shown in museum situations. I think these would be very useful/relevant for your project.
d) The 4-bit resolution surprised me. I understand the human eye is sensitive to 6-bit. What was the reason/limitation/purpose of limiting the display to just 4-bits? I know it would be difficult to change now - you'd have to increase your resistor ladder and add more trimming pots and opamps etc. But yeah.... is 4-bit really representative of the quality that was achievable "back in the day"?
e) Rather than one system to provide the same image/data to two televisors, and the conversion issues of your 512x512 this brings... I had the advantage of designing for just one output system and consequently could pre-convert MPG data into WAV files stored on a SD card (and, as it happens, selectable on a touch-panel LCD for playback). I did have some ambition to get this into a museum context somehow.
f) A note on the sound in your videos - I am hard of hearing, and found the significant variation in volume quite hard to deal with. At some points you were hard to hear, and sometimes way too loud.
g) As to the "electrical noise" of the baird motor connected to the disc and thus interfering with the infra-red sensor - this was not clear to me what exactly the problem was. Wouldn't the sensor be isolated from the motor and thus not electrically affected? Are we talking about some sort of EM interference and induced signal? I'm a bit of a novice with electronics so just trying to understand exactly the issue requiring the shielded box.
h) I was surprised at the seeming flimsiness of the baird disk. Is this an original disk?
i) finally, since you synch with an IR led, I assume that you have added a synch hole to the disks on both televisors. Did this mean modification of the originals, or were there some sort of holes already present.
So that's it. Thanks for posting the video link!
PS: as an aside - I'm working on a project (CSIRO in Australia) which is/was using Edisons. I guess we have to find an alternative now