I typically road bike at night on asphalt surfaces. What temperature (Kelvin) light would give the best color rendition. Hawaii has horrible roads and i need a good light to avoid potholes. If your familiar with LEDs, i was considering either a CREE XM-L 3000k 90CRI or Nichia 219 4500K 92CRI but i'm also taking recommendations.

I have 2 lights, and they are different colors, but once you acclimate to the color it doesn't much matter when cycling. If you are doing indoor lighting, photography, makeup mirrors, or trying to create an effect it does, but I can't imagine anybody would care on the road.

I use two P7s each of which are different colors and I don't think it makes much difference. The yellow one is more pleasant but the blue-green one seems slightly brighter. What's more important in my opinion are having the two lights in different locations: one on my helmet and the other on the handlebars. It improves perspective and proportion. I run my lights about half power and find about 600 lumens total are adequate. I often think if the lower light was on a front rack or bracket it would show ground contours better. On two-lane roads with oncoming traffic I sometimes turn off my helmet light because it seems to freak people out to see two lights stacked vertically - they can't process it and the cars sometimes wander a bit. Good batteries and light placement are important, color not so much. I slightly prefer the color of the lower temp LED though.

The color temperature doesn't matter. Positioning does.

Yes it does. The human eye does not perceive certain colors as well as others.

Maybe not, but it isn't going to matter when avoiding potholes. It isn't like you're going to be shining pure red, green or blue on them.

Color temperature refers to the BALANCE of red vs. green vs. blue. It does not, in any way, denote the presence/absence of certain colors.

We aren't talking about lasers here. We're talking about bulbs. If the unit is "white" (i.e. not red, not blue, not green, not amber), it's giving off some amount of R+G+B. The 3K unit is giving off a lot more R than the 4.5K unit.

From hanging around BudgetLightForum I've found that different temperature lights do indeed have different color rendition index (cri). This does seem to make a difference in a "walking around" light but I can't say I've noticed any difference when using warm or cool lights as bike lights.

I'm a lighting engineer, and have done a lot of work on human vision on roadways. Research has shown that there is a definite sensitivity shift towards shorter wavelengths as light levels decrease. To keep this simple, assuming that the two lights you mention have the same lumen output, it is likely that you will perceive better visibility (see further into the darkness) with the higher color temperature light source. So, go with the 4500K one if you are going to base it on just this decision.

However, if they have different lumen outputs, then it becomes much harder to compare. Also, if they have different beam spread (distribution), you can't really compare them directly very easily.

Don't take my word for it however, try them, and see which one allows you to see farther into the darkness. What you want to compare is the first point where you are able to see a hazard in the road ahead as it emerges from the darkness. With higher color temperature sources (bluer light), you will typically gain 10-15% greater detection distance, given the same light output and distribution.


---Michael

Doesn't a lumen take into account the spectral sensitivity of the eye, and so is perceived intensity? In other words, a lower color temp light would have to have a higher power output to have the same lumen rating as a higher color temp light because the eye is more sensitive toward the blue. Both lights with equal lumens would have equal perceived brightness...more or less, no?

Yes it does. Of course, lumen ratings for bike lights is a crap shoot anyways with figures given often the theoretical maximums given by the emitter manufacturers when the reality is often less than half of what that is.

To answer the OP's question, I'd say look at the light emitted by all the lights, decide on what you like the best. The reality is ... it's not going to matter much, so base your decision on what looks good to you and other things like support, reliability, cost etc.

A photopic lumen does, yes. The problem is that at night you are shifting into the scotopic range, and this is why the sensitivity shift occurs. This shift is essentially done in the mesopic range, which falls between the full daylight luminances of photopic, and moonlight luminances of scotopic. Not the luminance of the moon, but the luminance of the earth while illuminated by the moon.

This is complicated, but trust me, in the luminance ranges that you are riding in at night, a blue source is going to produce 10-15% longer detection distances.

Rods vs cones and all that, but I was of the impression that typical night time driving/riding luminance levels didn't get down to that level, but I image they can, I guess, certainly in the distance where luminance has fallen off.

That's the point... hazard detection occurs on the fringes of darkness, especially with a bicycle headlamp. Those luminances are certainly in the mesopic range, although not likely to be in the scotopic range.

Typical roadway lighting does get into the mesopic range, so even running with a headlamp and some form of roadway lighting, you are still probably running in the mesopic range.