Chromatize - Color Blindness Correction and More

Two days late for the Assistive Technologies round deadline. I thought it was today (not that I would have won anything). Oh well.

I stopped work on this project because I thought my color filter wheels and shutter glasses had been stolen. The other day I found them again: I had only misplaced them. Therefore I will be resuming work on this project.

In the meantime I mostly built a demo/test box to demonstrate the concept and test out different color filters—I bought a Lee filters sample pack at a photography store to try different filters for possible future incorporation into the glasses. I had intended to demonstrate it at Calgary Mini Maker Faire, but I only started building the box the night before, and didn't get it done in time. Next year™.

I also bought a second spectrophotometer of the same kind, which works fine. I'll probably put some kind of controller on it at some point for automated spectrum capture, and use the first one for parts. (Aside: The other week Surplus Sales had an LKB Biochrom Ultrospec II, a quite nice UV/VIS spectrophotometer which would also have been easier to automate, which I was considering buying, but by the time I'd looked it up online and returned to buy it, someone else had bought it. I can't go there very often anymore, because their hours are similar to my school hours.)

Anyway, now I need to do the following:

• (preferably) Characterize the filters in the existing filter wheels
• Finish the box
• Try out different Lee filters
• Get an Arduino to control the filter wheels and shutter glasses

After two months away from it (lots of schoolwork, and then a bit of time to relax after the semester ended), I'm picking up this project again.

One of the things that was on my mind when I last worked on it was my desire to obtain a spectrophotometer to characterize the color filters in the filter wheels I have, as well as any filters I may use or make in the future. Just last week, while browsing at the provincial government surplus store, I noticed that they had six Turner 350 spectrophotometers for $25 each! One of them, which I bought, was a slightly newer version with a black faceplate and power switch on the front panel. Before I bought it, I first plugged it in and powered it on (which resulted in the lamp lighting up and the meter going to full scale, as expected) and then tried to test it by sticking a strip of colored plastic from a drink bottle in the sample holder (which had no result, not as expected).

I brought it back to Protospace on a Tuesday night (when we have our weekly open house) and guests April and Rachael very eagerly helped me disassemble it (being somewhat careful to protect the PMT from light), and then they and Protospace member Nick helped me attempt to diagnose it. Unfortunately, we didn't get very far. The main thing we noticed was that the needle would vibrate, and sometimes wander erratically down the scale, when we touched various parts inside, but it was never consistent. Probing the 1 Vpp output with an oscilloscope resulted in similarly inconsistent signals that sometimes looked like multiple frequencies superimposed.

Further investigation of the connections to the PMT revealed that each of the PMT's pins had a stub of another wire on it, suggesting it had been cut out of another piece of equipment (possibly another Turner 350) and installed in this one. Perhaps this individual machine has a problem that causes it to damage PMTs, or perhaps the PMT was damaged by light during the replacement or during our disassembly. (Edit: No. PMTs can only be damaged by excessive light when powered up.) Next I will attempt to reverse engineer the whole circuit and figure out which points would be useful to probe.

In other news, I intend to perform proof-of-concept testing soon with filters alternating under the user's direct control, with both colorsighted and colorblind subjects. With the mockup I did a couple of months ago, I did notice that things matching the color of the currently active filters looked brighter to me, which was the expected/desired effect. That counts as a successful prototype, right? I also need to get the color filter wheels turning under a microcontroller's control. Protospace member Derek mentioned something that does that in the context of discussing another member's project and I thought it might be useful for this, so I'll be looking into that as well.

This morning I woke up to the news that my project had been written up on the Hackaday blog. Thanks, @Benchoff!

Thanks also to everyone who has liked my project so far, and, I guess, thinks it's worth a dollar!

Unfortunately, I'm a bit sick this week (as well as recovering from biting my cheek really hard, twice, in the same place), so work will be slowed/delayed for a while. However, I can still type fine, so I'll be posting another project log entry or two soon.

Some of the blog comments also alluded to other implementations and approaches to correcting color blindness, which I'm planning to cover in the next few days.

A couple of the blog post comments mentioned that the DLP filter wheels can shatter while spinning. I don't know what the risk of this in my application is (given that they're probably not being heated and cooled as much), but I think this is another reason to make custom filter wheels out of a safer material such as a plastic. I don't think the very narrow pass band of dichroic filters is necessary for this application (with the current approach, anyway). For the prototypes I'll put some kind of shield around them, I guess.

The other day I was at Protospace for our weekly mini open house. Protospace happens to be where I have my collection of color filter wheels (on my member shelf), and I intended to make a mockup that night. However, our magical wish-granting free shelf made it even better.

That night, several pairs of active shutter 3D glasses had appeared on the free shelf, and proved quite popular. I had been thinking that my device would need shutters of some kind to eliminate motion blur as the filter wheels spin (analogous to the shutter in a movie projector eliminating motion blur as the projector switches frames), and these LCD shutters will probably do quite nicely. They work the same way as an auto-darkening welding helmet, but have a separate shutter for each eye, which (for 3D viewing) alternate in sync with the image being displayed on a screen to produce the 3D effect.

I, however, will be closing and opening the shutters together rather than alternately, because the filter wheels will spin in sync with each other. My plan is to remove the shutters from the glasses and mount them and the filter wheel assemblies together onto a head harness of some kind. This mockup (color filter wheel assemblies taped to shutter glasses) was pretty heavy:

At some point I might design my own color filter wheels for this project (especially with a larger radius to increase the field of view), but I want to experiment with these ones more first (both because I already have them, and for an exciting reason to be explained in an upcoming project log entry). As well, I found out who supplied them to the free shelf the other month, and he said he can get me more if I want them. I already have a bunch, but if this proves popular among local colorblind people, maybe I'll use those to make them some goggles to try out before getting custom ones made.

For initial testing, I have to figure out a good way to control the filter wheels to go to one filter on command and stay there until the next command. They use brushless motors, which I can probably set up to microstep if necessary (?). This, however, is going to require me to investigatively disassemble one of the assemblies again, further than I did last time, because I was unable to see and trace out the entirety of the PCB then. The internal circuitry is pretty simple, though. I think I'll be able to control it with a simple microcontroller or something, and because the motor is so low-powered-looking, I probably can just use some transistors to drive it instead of using a brushless ESC (which would be more difficult to control for static positioning).

I'm in a software development program at a local polytechnic institute currently, and the first semester, being common among all IT majors, included a course intended to help orient students to the whole field and the other majors offered in case they want to switch. This included a field trip one day, a tour of the telecom department. I learned about interesting technology, the hardware was fascinating, and there was a demo of splicing optical fibers at the end.

But what's relevant to this project is that one of my classmates asked if he would be able to do cabling even though he's colorblind, and was told he wouldn't be able to, because it requires color vision. Later we had a brief conversation about EnChroma, which we'd recently seen in the news, but neither of us really knew anything about it. Later, I looked it up and found that it works by providing a greater separation between red and green wavelengths entering the eye, and therefore doesn't work for all types of color blindness (though red–green is the most common).

Months later, I found some color filter wheels from DLP projectors on the Protospace free shelf. These are used to enable the (single-chip, white-lamp) DLP projector to project a color image, while the DLP chip itself is a grayscale device, by filtering the light from the lamp to one color at a time, and cycling through the colors faster than the human eye can detect. I also knew that while most digital cameras on the ground use a Bayer filter array or other color filter array to take color photos, many of those on space probes use a color filter wheel, because this enables greater resolution and sensitivity (and sometimes is used to take a photo in only one band, full-color photos not always being necessary for science and data return being carefully budgeted). Such mechanical color filter wheel arrangements were also used in early color television technologies, but those were soon surpassed by the purely electronic system that was used until the 2000s.

I pretty much immediately had the idea to put those color filter wheels in front of a colorblind person's eyes to enable them to see in color the same way a space probe's camera does. However, I didn't make any significant progress, after investigatively disassembling the color filter wheel assembly, for a while. This year's Hackaday Prize gave me motivation to resume work on this project.

In the intervening period, I quizzed a bunch of people at Protospace and at school about the intended purpose of my project, giving clues such as that it's to solve a common problem, how space probes take color photos, the existence of a commercial product that solves this problem for only some sufferers, that anaglyph glasses could be used as a 'poor man's version', etc. I found it interesting that my friends who generally work on more technical things had a harder time guessing what it was for, and those who generally work on less technical things guessed the purpose more quickly, on average.

Just yesterday I showed it to one of my friends in the networking major, and she guessed what it was for as soon as she saw the photo, and said one of her classmates (not the same one as before—he's in software development with me) needs it because he's colorblind and is always asking what colors the wires are when he's doing cabling.