LED Lights

[Rohan Dadswell]

White balance & colour meters tend to only measure the red/blue of the light spectrum not the green/magenta part (there are some meters that measure G/M separately)

So a reading of 5600K wont tell you how much green is in your light only where on the red/Blue scale it is - this is true of LED, flos & HMI (due to their nature it's less of an issue with Tungsten light)

[Brian Standing]

So, if I'm getting a reading of 4300K, instead of 5600K, these lights are probably out of whack? ...sigh...

Guess I'll have to send them back.

[Richard Andrewski]

The only real way to reliably test discharge and LED color temperatures (unfortunately) is with a spectroradiometer and integrating sphere. All those sources with a correlated color temperature (CCT) can't necessarily be accurately read with a simple color temperature meter. Most of those meters were made to work well with full spectrum lighting like tungsten. They don't understand the spectrum of anything with a correlated color temperature. When we want to accurately read the color of a bulb or LED, we have to take it to a testing facility since an integrating sphere isn't found just anywhere and they aren't portable either as you can see below...

[Matt Gottshalk]

Originally Posted by Brian Standing

Hmmm... I'll have to test the HD100's white balance against some other known light sources, just to make sure it's accurate.

If so, I'll probably have to return the LED 500's. 4300 vs. 5600 is way off the 5% margin of error quoted in FloLight's technical specs.

Good luck on that. I had to send my LED 500 back to PP because it was NOT 5600k, it was 4300k like yours.

They sent me the exact same fixture back and it was still 4300k.

Buyer beware.

[M. Paul El-Darwish]

Fabulous! I've never seen a picture of the sphere before.
I concur with your statement regarding the fidelity of readings made with a spectroradiometer versus a color temp. meter.Indeed, this reveals that the illuminations industry simply doesn't adhere to strict standards for measurement or labeling of light sources. It's a hugely frustrating issue that has had me at odds with the IEEE, and NIST at times. Surely there must be ISO compliant manufacturers?

Originally Posted by Richard Andrewski

The only real way to reliably test discharge and LED color temperatures (unfortunately) is with a spectroradiometer and integrating sphere. All those sources with a correlated color temperature (CCT) can't necessarily be accurately read with a simple color temperature meter. Most of those meters were made to work well with full spectrum lighting like tungsten. They don't understand the spectrum of anything with a correlated color temperature. When we want to accurately read the color of a bulb or LED, we have to take it to a testing facility since an integrating sphere isn't found just anywhere and they aren't portable either as you can see below...

[M. Paul El-Darwish]

Why? 4,300ºK is a fabulous place to be. It's far closer to matching the color temp of so-called 'Daylight' in the real world. The only 5,600ºK I ever see or want is in the shadows or top of a mountain scene. Otherwise, the sunlight that I've measured at the hours of the day that I actually shoot at... up till 9:00 AM and after 2:00PM is much closer to 4,300ºK. Heck, I have a special warm card that purposefully shifts my camera's manual WB towards 4,300ºK for people oriented shots.

You do realize that ambient daylight shifts enormously in CRI as well as Color Temp over the course of a day and that peak daylight color temps vary depending on geographic conditions?

Originally Posted by Brian Standing

So, if I'm getting a reading of 4300K, instead of 5600K, these lights are probably out of whack? ...sigh...

Guess I'll have to send them back.

[Richard Andrewski]

Originally Posted by M. Paul El-Darwish

Fabulous! I've never seen a picture of the sphere before.
I concur with your statement regarding the fidelity of readings made with a spectroradiometer versus a color temp. meter.Indeed, this reveals that the illuminations industry simply doesn't adhere to strict standards for measurement or labeling of light sources. It's a hugely frustrating issue that has had me at odds with the IEEE, and NIST at times. Surely there must be ISO compliant manufacturers?

Most everyone refers to "color temperature" on discharge lighting for the specs and LEDs too but its more technically called "CCT" or correlated color temperature because its a simulation. The only full spectrum light (despite a lot of hype) is real daylight or real tungsten with some rare exceptions here or there. Also, I don't know if there are small color temperature meters (usable for video/film) out there that will read CCT correctly but it would be interesting to see if some exist and it would also be good to know which (if any) camcorders that have color meters read and display CCT correctly (or not).

[Richard Andrewski]

Originally Posted by M. Paul El-Darwish

Why? 4,300ºK is a fabulous place to be. It's far closer to matching the color temp of so-called 'Daylight' in the real world. The only 5,600ºK I ever see or want is in the shadows or top of a mountain scene. Otherwise, the sunlight that I've measured at the hours of the day that I actually shoot at... up till 9:00 AM and after 2:00PM is much closer to 4,300ºK. Heck, I have a special warm card that purposefully shifts my camera's manual WB towards 4,300ºK for people oriented shots.

You do realize that ambient daylight shifts enormously in CRI as well as Color Temp over the course of a day and that peak daylight color temps vary depending on geographic conditions?

Most people believe and the consensus seems to be that daylight is between 5000K and 6500K so that's probably why they were concerned.

[Brian Standing]

Originally Posted by Richard Andrewski

Most people believe and the consensus seems to be that daylight is between 5000K and 6500K so that's probably why they were concerned.

Well, not only that, but there's the issue of trying to color-match to other artificial lights as well. If they're "daylight" balanced, they'll likely be in the 5500-5600K range. My tungsten lights with a dichroic filter come in right at 5600K, and right at 3200K without the filter. Part of the reason I want color-balanced lights, is so I can use 5600K presets in the camera and avoid the compromises of manual white balance.

Also, I can't speak for the light in Maryland, but in Wisconsin on most sunny days, my camera's white balance is showing between 5000K and 6500K. It only dips down to the 4000 range when the sun is low in the sky in the morning or at dusk.

That's interesting about the difficulties in measuring the color temperature of LED lights. Once I get my copy of DVRack working again, I'll have to light a scene with the LED 500 and look at it on the vector scope. I'll be picking up some minus-green gel tonight, so I guess further experimentation is in order.

[Bill Pryor]

When lighting with the 500LEDs, I use my daylight preset and it looks good. The end result is all I care about. If there's a green spike, I don't see it in my flesh tones, or if I do, it looks OK. Just for fun I think I'll get some 1/8-green to see if there's any noticeable difference.

[Brian Standing]

Originally Posted by Bill Pryor

When lighting with the 500LEDs, I use my daylight preset and it looks good. The end result is all I care about.

Hey, Bill. Got any stills you could share with us? Are you using all 500LEDs, or are you mixing with other lighting equipment?

Thanks in advance.

[Forrest Burger]

I just did a test shoot with the cheap LED panels that started this thread. This was shot with an EX1 with one panel overhead and another coming in from the right as fill. The fill panel was diffused.

No color correction. Not too bad for tabletop work, but these panels are very lightweight and fragile.

Forrest

[Michael Chenoweth]

Nice image, Forrest.

How fragile would you rate the panels? Are they transportable? Do you recommend reinforcement?

I just picked up a softbox kit from Coollights to test against my LED 500 - having a bulb issue (possible shipping) before I can really get down and dirty - but if $50 could get me another panel similar that throws 200+ watts or so of light, they're sure nice to tote around over larger lights, even a large softbox with huge bulb.

[Matt Gottshalk]

Originally Posted by M. Paul El-Darwish

Why? 4,300ºK is a fabulous place to be. It's far closer to matching the color temp of so-called 'Daylight' in the real world. The only 5,600ºK I ever see or want is in the shadows or top of a mountain scene. Otherwise, the sunlight that I've measured at the hours of the day that I actually shoot at... up till 9:00 AM and after 2:00PM is much closer to 4,300ºK. Heck, I have a special warm card that purposefully shifts my camera's manual WB towards 4,300ºK for people oriented shots.

You do realize that ambient daylight shifts enormously in CRI as well as Color Temp over the course of a day and that peak daylight color temps vary depending on geographic conditions?

Yes I realize that. I also realize that true daylight is usually around 5600k, not counting for twilight or early morning hours.

Otherwise HMIs would be 4300k, wouldn't they?

Also, if a light is ADVERTISED as a 5600k light, then 4300k just does not cut it.
Period.

[Forrest Burger]

Originally Posted by Michael Chenoweth

Nice image, Forrest.

How fragile would you rate the panels? Are they transportable? Do you recommend reinforcement?

I just picked up a softbox kit from Coollights to test against my LED 500 - having a bulb issue (possible shipping) before I can really get down and dirty - but if $50 could get me another panel similar that throws 200+ watts or so of light, they're sure nice to tote around over larger lights, even a large softbox with huge bulb.

Hi Michael,

I'd rate them as very fragile. They're definitely a consumer level item, but will serve nicely for this tabletop shoot. I doubt that I'd recommend them for travel. They'd require some pretty good reinforcement since they're essentially made out of plastic and a thin backing.

But, you know, for a little more than $100 bucks including shipping they're pretty useful.

Hope that helps.

Forrest