chinese 500 led light-color correct-how?

R

robfilms

Veteran
chinese 500 led light-color correct-how?

so i recently bought a chinese sourced 500 led day light unit.

i got the adorama brand flashpoint (then on sale for $159)

http://www.adorama.com/FPVL500.html

there are similar lights on amazon (fancier is another brand) and ebay.

the light seems nice enough for the price.

i'm not sure i would use the dimmers.

in my testing i have set it to full blast.

at 10 feet on my interview subject (without diffusion), i was 4/5.6 at 1/48 and asa 500 (my fs100 is rated at asa 500.)

i ended up with the 500 led as a key light at approx 4.5 feet shooting thru diffusion which was about 2 feet from the subject and that gave me a useable 3.2.

but i am unsure of the color.

i've read of magenta spikes (david w. jones below suggests it is a green spike that is at issue) but without the proper tools and the knowledge to use those tools, i cannot "see" the spike.

in fact, the image looks fine to my eyes.

i manually white balanced, and the interview subject's skin tone looked lovely.

so, those who have also used these chinese 500 led fixtures:

do you gel to lessen the magenta spike?

do you have a standard color correct in post?

do you simply perform an accurate manual white balance and move on?

any and all thoughts are appreciated.

be well.

rob
smalltalk productions/nyc
 
Last edited:
David W. Jones said:
It has been my experience that these light generally have a green spike.

But if you are happy and your clients are happy?
Click to expand...

david-

thanks for sharing your experience.

how do you handle the green spike?

will a manual white balance "get rid" of the spike?

thanks in advance.

be well.

rob
smalltalk productions/nyc
 
Personally, if you are really worried, then there are, as mentioned all sorts of meters and analysers you could use - BUT - what do the pictures look like. People who see green or pinky tinges to what should be white usually refer to these as spikes, but the reality is that the spectrum from all kinds of lighting has weird humps and bumps, and they're only an issue, when they are! White balance, and tweaks in post can sort out most 'problems'. The only real issue is when you mix light sources - and then the usual correction filters can be used. I tend to just use a monitor I trust, and just look at the white and judge for myself. The white from my LEDs is one colour, my tungsten lights another, and of course the two different class lamps in my discharge fixtures are yet again both different.

I've got sheets of all sorts of colour correction - greens, oranges, blues, full and half and I never use them automatically - I use them when it's needed. My view is that real life is not white, and light bounces off coloured surfaces everywhere, so this quest for absolute white just produces something a bit sterile.

Stills photographers, especially sports ones, have been coping with green grass for years, and anything shot in the street has been sodium spiked for quite a while now.

Lots of my work has RGB LED light - and just recently I discovered that there appears to be a red component in my deep blue colour - no idea where it comes from, but it's there, plus a sizeable UV output. It makes the blue look very vivid.
 
david-

Chroma Meter CL-200A...

...quite the tool!

unfortunately, such expertise is not available to me.

paulears-

i LOVE the idea of mixing lights, each providing a different pallet.

my question the is, how?

be well.

rob
smalltalk productions/nyc
 
I am a fan of mixing lights also. I have found that working across a color continuum, from say cool to warm or warm to cool helps to add and enhance depth in a frame. I have run into some nasty green spikes with using CFLs as inexpensive daylight sources. I use minus green gel to knock down the green spike. I use very minimum strength gel, usually just a 1/8 minus green, and that works just fine. To see the color of your lights, and not having access to a relatively expensive color meter, I suggest setting up a white source, lighting it with your lights, then looking at the image in your NLE video scopes. The RGB cascade view in the scopes and the vector scope view will give you a lot of information about the color of your lights. I would also shoot your white reference under lights of known color, tungsten at 3200K and a reliable color daylight source at 5600K and compare them to the white reference that you shot with the LED lights that you are questioning on the scopes. If you do not have an actual light that is a reliable daylight color temp, set up your white reference so that it is lit by reflected skylight (North sky) on a clear day at about noon. The actual color of that light is probably going to be higher than 5600 K, but is it will give you a reference to compare your LED lights to. I am really just getting into using the color scopes in my NLE and color correction app (Apple Color) and there is a wealth of information available from testing your lights this way.
 
David G. Smith said:
I am a fan of mixing lights also. I have found that working across a color continuum, from say cool to warm or warm to cool helps to add and enhance depth in a frame. I have run into some nasty green spikes with using CFLs as inexpensive daylight sources. I use minus green gel to knock down the green spike. I use very minimum strength gel, usually just a 1/8 minus green, and that works just fine. To see the color of your lights, and not having access to a relatively expensive color meter, I suggest setting up a white source, lighting it with your lights, then looking at the image in your NLE video scopes. The RGB cascade view in the scopes and the vector scope view will give you a lot of information about the color of your lights. I would also shoot your white reference under lights of known color, tungsten at 3200K and a reliable color daylight source at 5600K and compare them to the white reference that you shot with the LED lights that you are questioning on the scopes. If you do not have an actual light that is a reliable daylight color temp, set up your white reference so that it is lit by reflected skylight (North sky) on a clear day at about noon. The actual color of that light is probably going to be higher than 5600 K, but is it will give you a reference to compare your LED lights to. I am really just getting into using the color scopes in my NLE and color correction app (Apple Color) and there is a wealth of information available from testing your lights this way.
Click to expand...

david-

thank you for the thoughtful suggestions.

you are kind.

i owe you a coffee.

be well.

rob
smalltalk productions/nyc
 
Is that meter improved over the old Minolta Color Meter? ( looks similar) If not i would be careful because the Minoltas were wildly inaccurate re green and magenta usually telling you to over correct flourescents and LED's by about double the correction needed.

Best test i know is just to get a reliable monitor and set your camera to daylight and tungstun presets , then use 1/8 Green or magenta and fractional CTO and CTB to match. warming to cooling. Comp[are to tungstun lights until they match or tungstun with full daylight on them.
usually you need a 1/8 to 1/4 magenta to match daylight properly
 
For what we currently do, we're using RGBW LED, so each fixture has 4 different colours - the white is there because because the LED spectrum is so spiky - as in lots of little peaks and the occasional big one, full Red, Green and Blue create a sort of dirty white, which while bright, is not very flattering. However, the white channel uses LEDs very similar to the ones found in video kit. The nice thing, using these types of fixture for video, is that for effect, you can add plenty of red, green or blue. I've not tried it yet, but the blue is VERY blue. Some stuff I have shot in theatres with the blue LED lighting looks amazingly like the moonlight we assume moonlight is (but of course isn't). The scope for some heavy effects laden video pops into my head. Colouring tungsten with something like Congo Blue is very similar, but more than 90% of the white light is lost - LEDs don't do this. You can buy LED PAR cans very cheaply now, with RGB, RGBW and even RGBAW - with an amber channel too. If you only want video white - steer clear, but if you do pop videos or weird stuff, well worth a look.
 
Lenilenapi said:
Is that meter improved over the old Minolta Color Meter? ( looks similar) If not i would be careful because the Minoltas were wildly inaccurate re green and magenta usually telling you to over correct flourescents and LED's by about double the correction needed.
Click to expand...

It was designed to work with LED sources.

i own 3 different calibrated color temp meters, and have a buddy with a very high end color measurement system, and you know what, my experience has been the reports of inaccurate metering of Fluorescent and LED are FUBAR. In our testing all the metering systems were very very close to the known baseline of his color measurement system.
 
I've done longer posts regarding this, but you can test all of your lights and dial in what sort of gels each might need.

Get a gray card and shoot it with a DSLR set to RAW and shoot with your LED, flo, whatever. Open the raw file at 5200k and eyedropper the grey. If the RGB levels are out of balance (pure grey will have all three numbers equal - like R125, G125 and B125). If the green number is more than 3 or 5 digits higher, you may see that spike. (You can also eyedropper the gray and adjust the color palette to the closest pure RGB - if you've got 166, 148, 169, make a square of 166/166/166 on a separate layer - this will give you a very effective visual reference - I mean, you may be really surprised to see your gray against mathematically correct gray).

You can use Photoshop's photo filter layers to play with correcting... 12% cooling filter (in my experience) comes close to 1/8 CTB for instance, 25% magenta approximates 1/4 minus green. Gel your light and retest. For example, I have Biax flos that need violet - they're kind of yellow. I have 6500k stuff that needs CTO. Mark the gel combos on the lights or on the globe packaging.

Camera Raw may not reproduce color the same as your camera is setup to, but you can get a daylight baseline for all of your lights this way.

I thought all my daylight stuff looked OK, but I recently did a shoot with a 400w mogul growlight, an Alzo 150 in a china ball, and a dual biax flo. Dialing in the gels (1/2 CTO on the mogul, 1/8 minus green here, minus green plus CTB there) was pretty startling to me... white seamless background product-demo type shoot, and the BG fell off to a great, even grey. I had control over tone & warmth in camera settings, but having everything shift the same via settings was really great.

Test test test!
 
I agree test test test . I've found that different cameras also seem to respond to the green spikes differently .

David, Don't know what color meters you were using but I can assure that my tests with a Minolta were not FUBAR but generally had me putting twice as much correction as actually desired.
 
I applaud you for doing your own testing. Most of the reports I have seen we're second hand repeats of internet banter. Our Chroma Meter and Color Meter II were within 100K of each other. The Color Meter III purchased off eBay was way off and needed to be sent off to be calibrated, which seemed like it took forever, very long wait!
 
Just been tinkering around with a dozen Nichia 219 neutral whites lately. 4200K neutral tint and 92 CRI. No green spike in my eyes.
Just recently made a 90 CRI light with 2 CREE MTG-2's in the warm binned version, no green spike there.
Finished a camera-top light using 4 CREE XM-L2's in U2 bin 5200K, green spike ? nah !!
If there is a green spike, then perhaps my aging eyes do tend to filter it out most effectively.
 
I'm not sure which model my friends meter was but I'm sure its pretty old. However for me the issue wasn't to compare it to another meter but to see if the recommended correction worked when compared to a tungstun source at 3200 or with CTB on it. In fact it was always substantial overkill. Useful information but it needed to be taken with a grain of salt. The bottom line is what looks good on a monitor with your camera. Generally I found about half the recommended correction was good. That was both for LED's and fluorescents.
 
Norman99 said:
Just been tinkering around with a dozen Nichia 219 neutral whites lately. 4200K neutral tint and 92 CRI. No green spike in my eyes.
Just recently made a 90 CRI light with 2 CREE MTG-2's in the warm binned version, no green spike there.
Finished a camera-top light using 4 CREE XM-L2's in U2 bin 5200K, green spike ? nah !!
If there is a green spike, then perhaps my aging eyes do tend to filter it out most effectively.
Click to expand...

norman99-

thanks for the sharing but for us uninitiated...

...i need you to speak slower and use language i can understand. :-)

any and all additional info is appreciated.

be well.

rob
smalltalk productions/nyc
 
You must log in or register to reply here.
Back
Top