And it has 4000K variant. Generally it’s very similar to the NTG50, they both most likely come from the same factory. I wish someone tested the XHP50.3 4000K R70 (bin J4 5D) from Convoy because I think Simon may’ve got shipped the 80CRI variant.
Very true. This also clarifies a common misconception that CRI measures how much colors pop or how nice colors look. CRI is simply a measure of similarity to a reference spectrum that is purely derived from physics and is under no obligation to please the human eye.
With LEDs, though, CRI is often a useful proxy of pleasantness, since current low CRI LEDs universally suffer from spectral deficiencies (e.g., blue spike, cyan dip, red deficiency) that make colors look bad. I have seen spectra that are low CRI but look better than 100CRI, but I have not seen one produced by an LED.
Also true. I think a very useful sanity check is to go outside at night with full moon high in the sky and no light pollution, and see how much you like seeing things illuminated by moonlight. That’s around 4000K, and I almost always find it too cool. My sweet spot is around 3500K, but sadly I can’t find any ultra high CRI LEDs (like SFT40) in that range.
2 Thanks
I even googled that because I couldn’t believe that’s true - moonlight appears to be so cool white. You learn something newa every day… I don’t think it applies to flashlights though because once turned on they break your night/twillight vision.
1 Thank
It really does! I’ve assumed moonlight is 5000K+ because it’s just reflected sunlight (of course, different wavelengths have different albedo), and that it appears cool to the eye. But indeed it is around 4000K, and I was able to confirm by putting a 219C 4000K on moonlight, holding it next to the moon, and staring into it.
I find that below low-medium settings, my flashlights–when taken as a walking light–still engages mesopic vision. But for higher outputs or upclose use, photopic vision returns.
On a tangential note, it would be interesting if someone engineered an LED that emits a spectrum that, under mesopic or scotopic vision, appears to the human eye the same way a 100CRI source would under normal (photopic) vision.
Interesting. To reverse the Purkinje effect and thus ‘enhance’ the dynamic range of the photopic vision.
But I wonder if there is an evolutionary reason to why nature appears the way it does under mesopic vision (of-course there is…). In my use-case it is precisely this balance to be aware of your whole surroundings without any bright spots that reduces visions sensitivity. Does it make sense? Not to illuminate everything like sunshine (yeah, to calculate the energy needed for that…) but make the whole surrounding environment seem like twilight. I don’t think a proper ~10x45, hazy spot TIR could accomplish that, that’s what I’m trying to accomplish. A dual or quad 45-25-10 -10 certainly does.
1 Thank
Precisely what I meant, but much better stated!
I think one hypothesis I’ve heard is that historically, low intensity sources tend to also be low CCT (moonlight, campfire), so humans adapted to prefer low CCT sources in low intensity conditions.
I’m not sure if I understood, but I’ll make a guess: you want to illuminate things in a way that is visually uniform, but not necessarily truly uniform: for example, a true uniform illuminator makes far objects in front of you very dark, and close objects under your feet very bright.
For this, I think some sort of frosted TIR would do the trick. I haven’t had any experience with the striped TIRs, but suspect that if they are sufficiently throwy, then some diffusion film might give it enough spill to cover close objects. I personally really like the intensity gradient of diffused narrow-angle TIRs, which are not the same as wide-angle TIRs. The latter gives a sharp cutoff, while the former has a continuous gradient from very throwy to very floody, like your 10-10-25-45 setup.
Precisely. To get a perceptually uniform illumination, we need a continuous gradient, like we use in acoustics with line-arrays. Near the source getting the same intensity (power per area) as far. No hotspots.
To illustrate an ‘ideal’ lens+LED combo:
- Carclo 10049 driven by a small, ‘trowy’ emitter (horizontal plot):
- Same Carclo driven by a large, floody emitter (vertical plot):
Getting a nice balance with a single emitter is ‘fun’. But seems like a large, floody emitter makes the horizontal pattern only a little bit worse - by introducing a slight falloff with angle.
Notice broad, constant horizontal ‘coverage’ with a nice, wide plateau and a nice constant falloff w. angle in the vertical. Vertical could be still too steep for use-case, jostling around causes too sudden changes in light output if the hotspot is not precisely aimed at the distance. A 20-30deg vertical pattern could be better suited.
Sadly Carclo (also LEDIL and other major TIR lens manufacturers) makes only one type of polar-pattern elliptical lens (45 by 10/15 deg, depending on LED). Everything else is a custom manufacture. Luckily Aliexpress is full of ‘interesting’ TIR optics, mostly in 20mm-23mm size.
I see! For the sort of uniform plateau beam profile with a smooth fall-off, I find that beaded TIR optics do a great job.
I current have a SFT40 running a 20mm “20 deg” optic; I assume that the 20deg is radial angle, since I measured that the beam cone has a diametrical angle around 40 degrees. It produces a nice large circle (a bit hexagonal due to the beading being a hexagonal lattice) of uniform illumination, though the falloff is a bit fast.
I would try Convoy’s 5050 TIRs in 20mm and 21mm, and try some beaded options with at least 20 degree angle. One of them might work great for you.
1 Thank
… moonlight. That’s around 4000K …
I didn’t know this. To my eyes it looks cool, closer to 5000k, but I believe the 4000K statement. At night, I definitely don’t want to go over maybe 3500K for a light.
I was just earlier outside, illuminating my yard with two different lights. DW4K 5000K FFL351A, and Maeerxu XT3 3000K XPL-HI. Even though those XPL-HI gen2s have pretty bad angular tint shift (a floodier lens helps), it looked far better, and had better perceived color reproduction, even though statistically they’re worse CRI.
I really like those 5000K FFL351A during the day, they’re very neutral and high CRI. I can’t think of a more neutral DUV neutral-to-cool tint emitter that I have. they make 5000K domed 519A look quite green. However, at night, they do not make me feel good like the warm CCTs. the FFL feels unwelcoming, cold, and clinical.
1 Thank
I find this equally surprising! At night (or just low light conditions in general), the human perception of color (and thus preference of CCT) changes.
To your brain, perhaps the XPL-HI has higher CRI!
It is known that the human eye’s luminosity function (weight/brightness given to different wavelengths) changes under low light conditions. During daytime, the signal your brain receives is not the literal spectrum of the emitting source, but the product of the spectrum and your daytime luminosity function.
At night, your luminosity function changes, so the same high CRI source no longer sends the same signal to your brain. It is quite likely that the low CRI, low CCT source simulates the “high CRI signal” better than the higher CCT, high CRI source, under the low-light luminosity function.
1 Thank
This is very interesting. First time I’ve heard anything like it, really.
I’m almost always up very late, due to my scheduling for the last decade, even though I would prefer the opposite nowadays. I use my lights every single day without fail. I guess I never really noticed that until last night, but usually if I have two contrasting CCT lights in use, the cooler one is most times a thrower, these were both bloody lights.
I was looking around the yard to see if I dropped something. the 5000k did seem to wash out the colors of the leaves and grass more, like a cheap 70cri light that “normal” people are used to, even though that was not the case technically.
I wonder if I could find somewhere to research this more.
1 Thank
You might find these two articles interesting:
-
Kruithof curve - Wikipedia
“The Kruithof curve describes a region of illuminance levels and color temperatures that are often viewed as comfortable or pleasing to an observer.” -
Purkinje effect - Wikipedia
“As illuminance decreases, human sensitivity to blue light increases.”
“In consequence, reds will appear darker relative to other colors as light levels decrease.”
1 Thank
Neal have some in stock, I will be ordering them soon from him 
Wonder who is this Neal fella? Maybe he could send also my way a couple.
I assume Neal from nealsgadgets.com altough I don’t see them on his website.
1 Thank
That is actually a pretty significant 25% gain vs. typical G4 R9050. Sadly it still uses the dreadful/mind-bogglingly idiotic Cree-s EasyWhite chroma binning.
BTW Convoy has a J4 R70 to compare with - that is in a perfect 5D chroma bin. Digikey and Mouser seems to keep all the binning codes a secret, only order codes are available (E4 and E5). Quite tolerant to slightly greenish tints, but a 5B or 5C bin is probably too much.
@BlueSwordM, I don’t see the 4000k high CRI variant of the N5-165HPE++ sale on Taobao on the ‘Nightwatch DIY’ listing.
Could you send me a link where they are for sale?