[Light] LEDs are NOT your friend. "High CRI" is a lie.

[Kastor Krieg]

Ever used LED light (strips, bulbs) and found your miniatures looking way off from actual daylight? Your eyes strained, tired? Yourself possibly strangely annoyed? That's because even the top-tier LEDs are completely crap for what we use them as.
 

What does CRI mean and why is CRI important in LED lighting?
Can’t tell the difference between the black and navy colored socks in your walk in closet? Could be that your current lighting source has a very low CRI! Not all light is made equal; some light renders color better than others. Color Rendering Index (CRI) is the measurement of how colors look under a light source when compared with sunlight. The index is measured from 0-100, with a perfect 100 indicating that colors under the light source appear the same as they would under natural sunlight.
 
This rating is also a measurement in the lighting industry to help discern naturalness, hue discrimination, vividness, preference, color naming accuracy, and color harmony.
 
- Lights with a CRI that is measured greater than 80 are considered to be more than acceptable for most applications.
- Lights with a CRI that is measured greater than 90 is generally considered “High CRI” lights.

 
SOURCE: https://www.flexfireleds.com/color-rendering-index-cri-and-led-lighting-what-is-cri/



This is a diagram of the CRI (Color Rendering Index) of a premium 94.4 CRI LED strip, which goes for $249 per 16'5" (5 meters). But, hold on! How come it has ~95% color rendering accuracy with a ~15% dip in pure yellow, ~25% dip in pure red and a huge ~35% dip in pure blue? Ah, my summer child. That's where we learn not to ever trust a CRI rating again.
 

Testing for CRI requires special machinery designed specifically for this purpose. During this test, the light spectrum of a lamp is analyzed onto eight different colors (or “R values”), termed R1 through R8. There are 15 measurements which can be seen below, but the CRI measurement only uses the first 8. The lamp receives a score from 0-100 for each color, based on how natural the color is rendered in comparison with how the color looks under a “perfect” or “reference” light source at the same color temperature as that lamp.

They just take first 8 result out of 15, barely over a half, only weird pastel mid-tones, know that greens and skintones look OK and call it a day!

And this is, like I said, a premium LED strip that boasts color accuracy. One you wouldn't normally use or see used in a mid-price bulb - those are way worse. My friend has family jewelry with natural stones which are completely different colors under LED and sunlight (as in purple vs green!). That's how skewed is the wavelength range.

Remember, if something is too good to be true - doesn't generate a lot of heat, but a crapton of lumen, takes 15W only and is cheap - that means that the engineers have severely cut the corners and are hoping that a regular user won't notice.

And they won't, usually, but for anyone who paints and treats their color composition seriously - 5000K incandescent / flueorescent light or bust. Get a good flueorescent lamp, e.g. Velleman VTLAMP6 or equivalent for cheap on Amazon/eBay (75 EUR?) and let your eyes finally have a break.

Cheers!

[Golf33]

Thank you! That could well explain some of the eye strain I get when painting (the rest is my barely-functional eyes).

[Mr Yellow]

An interesting counter point to this is; if you paint your models in artificial light and view them in artificial light (playing indoors) wouldn't the quality of the LED you use be moot? You're unlikely to compare them with sunlight to highlight any difference.

[Suspira Mortis]

Your argument about the weakness of LEDs is compelling, but you don't proceed to explain why I should prefer an incandescent/fluorescent lamp.

[NiceGuyAdi]

Krieg, I'm not sure a lot of what you're saying here is entirely helpful. I'll explain why, but preface this by saying that I've worked in video production for 15 years so I'm not coming at this cold. Of course lighting something for a camera sensor and lighting for your eyes to paint by isn't completely the same, but there's a lot of crossover.

 

First thing to say is I don't see how LED light will cause eye fatigue or a bad mood compared to incandescent or fluorescent. Mood I can't speak to, but eye strain tends to happen when you're not seeing with enough light, or something that's flickering.

 

On to colour accuracy. What your post points out is how and why CRI is a far from perfect system to measure colour accuracy. But that's not limited to the measurement of LEDs. You can have two sets of lights that have the same CRI score, but render the colours being tested differently (because the CRI is an average score), whilst the colours not being tested could be wildly different.

 

The graphic you posted show poor relative output of blue light in particular, which is consistent with the graphic below for a "warm white LED". But looking more broadly at the different technologies you can see how for instance fluorescent light can have enormous spikes at different colour wavelengths, whilst incandescent puts out much more red than blue light. These will create a different perception against daylight's fairly even production of light across the visible spectrum, which I guess we all accept is the gold standard light source.

 

 

Incidentally I've never come across a 5,000K incandescent light. They work more in the 3,200K range, so are a very warm light.

 

Feel free to correct me on any of this. I'm not the sole authority here, but I don't want people to make decisions about how they spend hours and hours painting from the wrong start point.

 

tl;dr - all current light technologies have drawbacks compared to daylight, and in my opinion LED is not inappropriate for miniature painting. CRI is a guide for how good the colour rendition of a light is, but not a perfect guide by any means.

[MegaVolt87]

Yeah I hate LED's, my magnifier lamp is LED and I use it for only 5 min max otherwise I will get a headache. I have an old school lamp my mum gave me from when she used to do building plans back in the day by hand, those warm daylight old school bulbs are the best IMO. 

[NiceGuyAdi]

 my magnifier lamp is LED and I use it for only 5 min max otherwise I will get a headache.

 

Hmmm, you just spurred me to read a bit more. LED lights, like fluorescents, both flicker. But where fluorescents drop about 35% light in their flicker, LEDs drop 100% because they completely switch off and on. It's happening hundreds or thousands of times a second, so you can't see it, but that wouldn't stop it causing a headache.

 

The old school tungsten filament bulbs didn't flicker at all. I feel like throwing all my lighting out and starting again.

 

There are apparently flicker-free LEDs out there, I'm gonna take a look at those.