I don't know if this will be of any interest to anyone... but I'd like to share the results of my little science project involving a common nitrate test kit. I'm new to the forum and hope I'm not repeating something that's been overdone. Apologies in advance... Also, I appreciate any comments and ideas!
Everyone knows that it's sometimes really hard to use a test kit and match your resulting sample to one of the colors on the printed color chart that comes with the test kit. This problem stems from the fact that the color chart is reflected light from a printed medium, and the sample vile is transmitting light through water - two completely different mediums! Even under optimal viewing conditions and viewing angles, it can be hard to discern subtle color differences - "Ummm, <squints> Do I have 3ppm nitrate or 6ppm nitrate?"
And I question, "Can my test kit even resolve small differences of nitrates?"
This really isn't just about nitrates, it's about any water parameter that has a test kit which requires visual color matching to get the result. But nitrate is the example here. Since I recently started a dosing regime (yay!), I really
wanted to observe how my tanks consume nitrates and I needed to know if an "average" test kit would allow me to do that. I figured if I could remove the "human" visual complexities, then I might be able to get better performance out of my test kit and discern smaller intervals in the nitrate concentrations.
I have the perception that most people serious about testing their water parameters would, first and foremost, test their test kit against a "known" concentration - to see if it's even working correctly. But I decided to take that a step further and try to improve the visual matching aspect involved in these types of test kits.
My idea was to carefully create several known concentrations of nitrate solutions, test them, and then photograph these together with the printed color chart that comes with the test kit. IF
the whole scenario was lit and photographed with care, it would be possible to create a better match using a digital image and see how the test kit was really performing.
Here is the result of my experiment... the photograph of the tested samples together with the color chart. (For setup details, see separate section below.) The small squares of color in the test tubes are simply areas cut-n-pasted directly from the corresponding color chart swatches on the right side of the image. Please let me know if image does not appear below...
My first reaction was, "Wow. - Those colors are pretty freak'n close!"
My second reaction was to notice how different the color swatches appear when they are separated away from the color chart. For example, the yellow color for 0ppm looks much more vivid (saturated) on the color chart than it does when placed as a little square above the water sample - but it's a direct copy from that swatch. (Check it yourself! Copy the image and use a paint program to move the little squares to their corresponding color on the chart.) Needless to say, I was pretty darn satisfied with the relative accuracy of this particular test kit (Aquarium Pharmaceuticals). <Who knew!> But the implications of this little experiment are even more exciting...
Expanding on the Results:
Once I verified that the test kit (two years old!!
) was working correctly, I knew that I could completely remove the color chart from the visual matching equation. For example, if I always perform a 0ppm and a 5ppm "known" test along with my tank samples, then all I need to do is compare the tank samples to the "known" samples and easily see if I'm over or under my 5ppm target. In other words, match to the colors of other test tubes, not the color chart. Same medium, easier matching.
Alternatively, if I use the EXACT photo setup (lights, exposure, position, lens, etc.), then I could photograph my tank samples and literally cut and paste the resulting image colors into my reference image to see where they match up!
Another extension would be to create known concentrations of nitrate from 0ppm-10ppm in 1ppm increments, then photograph that set of samples and use it as a reference image against other photographed tank samples. Your test kit can probably resolve those small differences, it's just that it's difficult for us
to discern the color difference. The digital reference image can really help. (But you gotta be a total nerd to do it. Like me.)
Notes On Photo Setup:
Simple part: I used a block of wood and drilled holes big enough to accommodate the test tubes. I used tape and thin piece of balsa wood to secure color chart.
The lighting part: The scene is lit from both the front and back with "soft boxes" (very EVEN light). Next, if you have it, use the histogram feature of your digital camera to make sure you are not clipping (over exposing). Technically it's best to use lights that are close to 6500K in color temperature because this is how the manufacturers design and verify the color chart to avoid things like metamerism
-- that's when the same printed colors look different under different kinds of light sources. I was careful to match the intensity of the back light to the apparent reflected values on the front of color chart. Next, position the front light to avoid any reflections in the glass and on the color chart. Do not use a front mounted camera flash... shoot long exposures if you have to. The key is even light from all sides with a sufficient back light intensity and a non-overpowering front light.
The camera part: Nikon D200, 19-28mm lens. Exposure: f29, 1/250sec. Two strobes with mounted softboxes. Two white foamcore bounce cards.
The hard part: Of course, it's always more complicated and technical than I just outlined above. But even so, this is not meant to be a technical method of color matching... the actual process of color matching is severely more complicated! It's just an idea to create a consistent viewing environment and help discern small color differences.
It makes for a great Saturday project and really gets you familiar with your favorite analog test kit.