We are on the way to the Hackaday Superconference, and will be bringing the EmotiGlass prototypes.  Please feel free to find us during the conference if you'd like to see one or try it out.

The front LCD displays use modules from the Electronic Assembly DOG series.  This line is fairly unique in that the LCD displays and backlights are sold as separate modules, and can then be assembled in the combination suiting the application.  For this project, the modular design allows us to easily use only the transparent display portion (with no backlight or reflector) without having to extract it from a complete module.  The series we are using, the EA DOGM128x-6, offers the display module in 5 "technology" variants.  After eliminating the reflective and colored variants, two options remain.

More details can be found in the datasheet here:
https://www.lcd-module.com/eng/pdf/grafik/dogm128e.pdf

The 'E' variant (DOGM128E-6) is a STN positive transmissive display.  Because it is positive, when it is open (pixels off/power off) the display is completely clear.  For this reason, we had selected this display to use initially. These displays do have a yellowish tint to them, but it is not noticeable after a few minutes.  Because there is a small clear space between the pixels even when they are on (darkened), contrast is limited, and these displays don't block the light quite as well as we would like.  

The 'S' variant (DOGM128S-6) is a FSTN negative transmissive display.  When off or unpowered, it is dark.

After building the initial prototypes with the 'E' variant, we decided it would be worthwhile to test the 'S' variant.  Initially, we attached an extra display to one of the prototypes and tested it in the clear and dark states.  We found that the 'S' variant performs better in the dark state.  We had some concern that the dark area between the pixels would be problematic in the clear state, but it's actually less noticeable than the tint from the 'E' variant.  Because the 'S' variant works better overall, we changed the displays in one of the prototypes to use them.

This change also require the firmware to be changed to flip everything as the display is now negative.  To allow only one firmware version to be maintained, we selected an unused microcontroller pin (A2) and tied it to ground on the prototype that has the 'S' displays.  The firmware will be updated to apply the internal pullup to that pin and then check its state, at which point it will know which display variant is installed and can flip everything (or not) accordingly.