I started a new thread here, because the original thread, Optimal color temperatures
, was getting off-topic. This is a long one, so dig in...
In the original thread, there were differing opinions on the relevance of using CRI (color rendition index) when choosing an aquarium bulb. There was also some confusion on why certain bulbs have a high CRI but don't have a color temperature that matches "daylight", i.e. 5000 - 6500K. I wanted to explore why these differing opinions might exist:
Here are two opinions expressed in the original post:
I have read that CRI for incandescent bulbs is always 100 also, and that CRI isn't of much use for aquarium lighting anyway. I tend to agree with that, and never consider CRI when I think about planted tank lighting.
I believe our goals are to provide appropriate light for the plants to grow, light that will let the non-green plant colors show up well, and light that makes the various colors in our fish look good. I don't think that requires a high CRI.
First, to be clear:
Originally Posted by Steve001
Color temp by itself should not be taken alone. A combo of color temp and color rendering index (CRI) will give you a good idea how objects will look. For the most natural rendition of color a light source with a CRI of 91 or higher coupled with a Kelvin temperature of 5500 will create a light that closely matches sunlight at noon.
CRI has nothing to do with optimal plant growth. It has everything to do with aesthetics, human vision, and how colors appear to us - not to the plants. I will also say that there is no evidence that color temperature (K) has anything to do with optimal plant growth either. The verdict is still out on whether certain wavelengths cause optimal
growth, but the general opinion is that it does not.
CRI and color temperature are largely for how we want are aquariums and plants to appear to us
. If your goal is to visually mimic the "appearance" of sunlight, then you need a bulb from 4000K - 7500K with a very high CRI. To understand CRI, you have to know a little about human vision and color perception. Secondly, not everyone has the same color perception nor the ability to distinguish between two similar colors. Color vision is a highly psychological phenomenon which makes creating a mathematical model of human vision very tricky. (See International Commission on Illumination, C.I.E. 1931 - which, still today, is largely the foundation of color space theory.)
If you want to see where you stand in your ability to perceive colors, take this color test: X-Rite Color Test
CRI (color rendition index) and color temperature (Kelvin rating) are not
related. The reason they are not related has to do with the human eye and our ability to correctly
perceive colors under a wide range of different illuminants. CRI is a metric that describes our ability to color match
under a specific illuminant. In everyday life, you do this all the time... i.e. A white car still looks "white" to you at dawn, noon, or dusk - even at night. A piece of white printer paper still looks "white" to you in an office building or in your dimly lit living room. We have the unique ability to "white balance" in real time. However, if you photographed the white car and photometrically measured its color in the resulting photograph, it would actually be blue, neutral, red, or orange depending on the illuminant and the white balance of the camera. Cameras are not very good at white balancing on their own. This is why they have a manual function for you to set it. When you use it, you are saying, "This is white. Remember it, and base all other colors relative to it." However, your brain does this feat without you even having to think about it. The phrase, 6500K Daylight or "D65", is only a standard that was created by the C.I.E. to allow observations and measurements to be taken under "known" conditions. There is also D50 and D55, for example.
The Sun and Incandescent Bulbs
Sunlight is used as the reference for CRI=100. The sun has a smooth spectrum with wavelengths present across the entire visible spectrum. As viewed from our atmosphere, the disk of the sun itself may have a specific color temperature, say 5000K; But the combination of the sky, sun, and atmospherics could produce an overall sky-light with a color temperature anywhere between 2000K - 10,000K depending on the time of day, weather, altitude, and season. In all these different situations, you will still be able to tell which object is "white". This is high CRI. It is irrelevant to the color temperature of the illuminant; And this is why an incandescent bulb with 3200K can have a CRI=100: because it is smooth spectrum and you can color match faithfully using this illuminant.
You Need a Reference
If you are in a room lit with nothing but 3200K incandescent bulbs, your eyes will quickly adapt to this light and you will not perceive other colors in the room as "orange". The same thing is true with a room lit with 10,000K bulbs. The colors in the room will not appear as "blue"... even though these lights are orange and blue respectively.
If, however, you put an aquarium lit with 3200K bulbs into a room lit with 10,000K bulbs, the aquarium will appear very orange to you. This is because your reference, the room, is 10,000K. Your brain has already decided that 10,000K is "white". The reverse is true, as well, and you will end up with a blue "washed out" looking aquarium.
The most natural visual effect would be to match your aquarium bulb's color temperature to that of your room lighting. But, just because the color temperatures are matched, does not mean that you will be able to judge separations in color accurately. This is where CRI comes into play. If you want to accurately perceive different colors under that lighting, then choose a high CRI light.
Fluorescent Bulbs and CRI
Unfortunately, things are more complicated for fluorescent bulbs. The reason has to do with a human vision phenomenon called "metamerism". Metamerism comes into play when you start creating colors with RGB, or when you perceive colors with RGB sensors, (i.e. the human eye). The C.I.E. (1931) determined that any
color can be described in terms of just three primary colors: i.e, red, blue, and green light. However, when an observer arrives at two different RGB definitions for the same color, this is called metamerism. Another example of metamerism is when the colors of an inkjet print look different under different lighting - this largely due to the inks which are metamers.
Fluorescence is notorious for producing metamerism. Do "tri-phosphor" fluorescent bulbs sound familiar? These are the higher quality bulbs that create light using multiple phosphors which produce red, green and blue wavelengths. When you look at the spectrum plot of these bulbs, these are the wavelength "spikes" you see. Their spectrum is not smooth like the sun or incandescent bulbs. Ultimately, it's the relative proportions of red, green, and blue in the fluorescent bulb that will define it's "correlated" color temperature. However, it's CRI rating will be dependent on how much metamerism it produces and how faithfully you can color match using this illuminant.
Have you ever noticed how your car looks under sodium vapor lighting (orange street lamps)? The color really changes: red can look gray, white looks orange, blue looks black; You cannot accurately judge color using this illuminant. Sodium vapor lights have a very low CRI=25. Standard cool-white fluorescent bulbs have a CRI=62. Higher quality fluorescent bulbs have a CRI=85.
(Keep in mind that metal halide lights also have low and high CRI versions. If you don't specifically buy a high CRI bulb, you most probably have a low CRI bulb. This is why simply obtaining a metal halide fixture doesn't magically make your tank look "amazing". HID light fixtures for marine/coral tanks are notorious for this. When it comes to metal halide, I always buy my bulb first, and then build the lighting fixture components around it.)
What About LEDs?
For now, see fluorescence above.. the same thing applies to LED because its "white" light is produced by using a fragmented spectrum with peaks at R, G, and B. However, this may not be the case for long.... See this article for info: LED Technology
There are two sides to the lighting of a planted aquarium: Light that is good for plant growth, and light that makes the tank look visually pleasing. But the term visually pleasing is subjective and unique to each person. This is the reason for the differing opinions. Some people simply will not be sensitive to accurate color, while others will. I am a person that is sensitive to color and its reproduction, so I pay attention to CRI when lighting my aquariums. "Sunlight" is not always my goal, but accurate, pleasing color is.
Fortunately, most lighting that is visually pleasing to us is also good for plant growth. You do not have to suffer through a "pink" grow bulb, nor a cheap "green" fluorescent bulb, nor a "washed out" metal halide - even though all these will grow plants just fine.