Correcting tint with LEE Filters

Correcting tint with LEE Filters

From a local e-store I obtained (for free!) a test sample book of LEE Filters.
The purpose was to find a filter to insert into Xiaomi USB lamps, which I found had too green a tint for my taste.
The filters are a thin polyester foil delivered in sheets of 122x53cm (4x1.7 feet). They are not expensive, around 10$ for a full sheet and they can be had in half sheet size also.


http://leefilters.com/lighting/colour-list.html
(Perhaps find a dealer to see a more simple presentation)

I will present the test of a few (anti-green) filters in one picture:

When you remove green it is gonna cost you in light output. The percentages in the picture are what should be left when using a filter. So it is a trade-off between how much green you wish to remove and how much light you want to preserve. The PLANCK LOCUS is where you want to go and the letter 'd' is the shortest distance to the locus from each tint.
The result (measured with Xiaomi lamps and a Color Analyzer which gives the xc,yc of a tint):


And the winners are:
The filter '003 Lavender Tint' keeps the same color temperature and removes excess green very fine. The nominal loss is 24% light.
The various 'minus green' filters (salmon color) add some warmth to the tint, shifting the color temperature. I would use '279 eighth minus green' on a light that I find a little too cool and with a greenish tint (many stock lights, in fact). The nominal loss is here 14% light.
The directions theese filters move the original tint will vary with other 'original' tints. A 6000K light will be changed somehow different so it is good to have a sample book and experiment to find the best filter.

I have not yet bought the sheets (there are so many I may want) to cut the filters from, so I have no pictures yet, but with the test sheets from the sample book it looked very pleasant.
You can also have a 'eighth plus green' filter (loss 12%) which I think could fix the lilac tint from some zooming lights (SK68 etc).

edit: on the picture, labels xc and yc should be interchanged.

No questions or comments?

1. Is it too technical?

2. Is the subject tedious?

3. Am I so good at explaining that everyone has got it?

That is a nice type of testing :-) , even though I wonder how many people accept 14+ % light loss to improve tint.

I'm not familiar with a 'Color Analyser' (I have an app for my phone, but I don't trust it because I doubt it takes into account which camera and CCD chip is in the phone), is it an instrument used in photography?

Don't be afraid that it is too technical, this is not the type of forum where you are blamed when you dive into the fine details of flashlights

Thanks for asking, djozz. Now I'm relieved.


The Color Analyzer is a discarded prototype of Philips PM5639 Color Analyzer that I repaired. Apart from the measuring head I was the sole developer of this apparatus. It is(was) used to calibrate colour monitors in TV factory lines.

Yes I have visited Eindhoven several times in the past.

As I don't have access to the calibrating equipment any more I found out a matrix to multiply the readings of colour coordinates xc and yc to obtain sane results for a multitude of known lamps. Somewhat like we have a factor for our integrating spheres. Mine is just a 3-dim factor.

Yes, 14% seems much but it is'nt visually. But I know most people want quantity over quality when it comes to flashlights.

Cool instrument! (I'm pretty jealous of that you are able to measure the xc and yc values actually, I want that too. )

Using instruments from a distant past sounds familiar, just yesterday I used the monochromator from an ancient (but quality!) Zeiss spectrofotometer to measure the output spectra of two UV-leds :-)

Yes, with some more resources one could have:
http://www.amazon.com/Ikan-MK350-Spectrometer-UPRtek-Black/dp/B00DOPW0RY

Although I worked 10 years in an opticals laboratory, I newer fully understood to operate the monochromator, perhaps because it was always the same guy that was trusted to operate it.

I have used Lee filters to alter the output color temp of camera flashes over the years. My concern would possibly be the light lost and the heat generated as the light strikes the filter in close proximity to the LED, with a high duty cycle.

Like I said, watching with interest.

Glad to have the pointer to the Lee filters.
I’ve mentioned similar ones from Rosco here and there, mostly for making “low-blue” or “no-blue” filters.
There are several other comparably large companies — if you don’t find them in your country look up theatrical or photographic filter suppliers.

Most of them have little “sample book” collections cheap or often free —- each sample big enough to cut two or three flashlight filters. Ask around. They go obsolete every year or so and theatrical/photo supply places often give them away.

I’ve been using Lee filters for a couple of lights one even on a Nichia 219A HCRI light and it looks really good. The visual difference in light loss is negligible in real world use for me but the nice warm tint it gives is great for night time use when everything outside is extra blue from moonlight.

I had heard about them from a fellow member on cpf. Here is the thread from cpf. I recommend checking it out as there are very nice before and after beamshots with different filters and how they change the tint and color rendition of the beams.

Definitely recommend the filters for anyone interested in a quick cheap way to change tints on there cool white lights.

There is now a special range of Lee LED filters with a longer liftetime to correct color casts (five “Minus Green” and four “Amber” tints):

Cree Zircon Filters

Regards, Juergen

Thanks for the info, good stuff.

I’ll admit when I first read it I immediately thought the trade-off wasn’t worth it. But if you consider 14% is exactly two CREE bins, that helps put it in perspective. I’ll drop to an S6 if it meant ideal light quality. I currently favor my old 5B1 T4 85CRI XML2 (from before they had bond wire issues) over most any other tint in my collection… Yes… that means the Nichia’s too. I’ve tried them all. A, B, R9050. I think they are slightly green and a bit too cool for my taste (4000K is my sweet-spot)… They are pretty darn nice dedomed though, but just not bright enough and somewhat hard to focus.

Seems like it could benefit from LEE filter 279 'eighth minus green'. You could try getting a free sample book (here sheets are 90x35 mm) from a reseller to try it out.

I’ll move my questions here to keep from flooding clemences high cri nichia thread with off topic chatter:)

So have you tried the 1/8 minus green with a nichia 219c that has a yellowish tint? What is the result?

If 1/8minus green with 003 would give a tint close to nichia 219b 9080 then I would think the output loss worth it. At least in my case with home lighting.

they lose a lot of light, is the main negative…

since they absorb light, the filters may overheat at high light levels, it’s just something to think about

wle

Ya… the heat loss could be calculated I suppose. But that means Ild have to turn on my brain and I usually don’t start it up unless I’m trying to show off or something. :disguised_face:

The honest answer is that I have not (yet) used filters on Nichia LED’s because the offset from the BBL has been acceptable as is (offset lower than 0.003 - that is slight green/yellow. Green for CCT’s higher than 5000K and yellow for CCT’s 4000K and lower).
Today I made a test with some lights with Nichia 219C 4003 from clemence and Nichia 219C 4005 from Wfans and from KD. To make it easier to judge, I converted color coordinates (xc,yc) to CCT (in Kelvin) and offset from the BBL in units of 0.001, like this:
Color point example: 4250/3 (CCT / tint offset).
As 3 is positive the tint is to the green side. If negative to the magenta side.
This means that the color is 4250K and lies in the (shortest) distance from the BBL of 0.003.
0.001 is considered the smallest offset a human can see by comparing. 0.005 is easy to see and is approx. the amount that one layer of filter 279 changes the tint offset. A side effect of 279 is that it also draws the color some 100-200K towards warmer.

I do not have the Nichia 219B 9080 to compare with (yet), but here are results for 219C 4000:

(sorry, I have trouble with the Table editor, hope it is readable anyway).
table(table#posts).
|before | with filter | LED | vendor|

3959/1 3823/-3 4003 clemence
3961/3 3823/-2 4003 clemence
4033/1 3863/-4 4003 clemence
4322/-3 4223/-8 4005 Wfans
4066/-3 3894/-8 4005 KD

The LED’s from clemence has a rather yellow hotspot that looks much better with the filter. The spill becomes somewhat magenta, like from a 219A 4500. Interesting is that the LED’s 4005 from Wfans and KD does not need a filter.
But that could just be one lucky batch.

Here is how one layer of filter 279 worked for some other lights I put them in.
table(table#posts).
|before|with filter|LED|filter|
|4531/10|4381/4|xml T6|279|
|5673/15|6006/12|xml T6|279|
|5506/16|5686/12|xml T6|279|
|4007/8|3894/5|xml2 T5|279|
|4791/5|4584/1|xpl hi 3a|279|
|4552/10|4406/8|xpg R4|279|

Using other filters, 206 orange and 278 plus 1/8 green:
table(table#posts).
|before | with filter | LED | filters|

10088/5 6498/8 chineese 206
9850/-13 6614/-7 xpg2 206+278
11950/-13 6235/-4 chineese 206+278

No need to toss all the bad LED’s when a filter can do the job.

I’m sorry for not providing pictures, but it needs a better camera than mine and I doubt that the small differences involved can be seen on a picture anyway.
Regarding the concern about the filters being exposed to high temperatures, consider that they were developed for stage floodlights. What I found negative though, is do not rub them with water or spirits, the color will wash off.
Have fun and good filter hunting.

Interesting to read that, sixty545, and not just what the filter does, also the variation that is in the Nichia leds. Thanks for testing and posting! :slight_smile:

The change in color temp is very interesting. High temp gets higher and low temp gets lower?

The leds I have are 219c from clemence and yes they are yellow to me. I am very intrigued now. I don’t mind a bit of cct shift or the loss in output. The 219b has a lower output anyway. I was going to sell them off and wait for some 219b but I just might give this 279 filter a try and stick with the yellowish 229c.

Just a note to readers: when I say yellowish, I mean compared to other earlier nichias. When comparing to a cree led the 219c actually seems less yellow. I don’t want to give the wrong impression of the 219c.

Thanks for the helpful info!

Another thing about filters is, they are usually designed for ‘white’ light as input, or at least broad spectrum.

But LEDs don’t make white that way, they make BLUE, a very limited wavelength, then that goes through something that absorbs the blue and emits a YELLOW, again a narrow band of ‘colors’.

So if you are trying to make it come out green or red, it will be very inefficient.