Light Sources and "True" Color Matching

Problems matching an existing color because of a difference between the lighting on site and in the shop? Here's an extensive discussion of the theory and practice of lighting and color-matching. June 9, 2007

Question
Fluorescent lighting in finish area bit us in the butt. Just got back from the jobsite. What we made and what we were supposed to match was not even close. We brought a piece of the furniture back to shop to match colors. It looked dead on in shop. Way off on site. I have short ceilings in finish area. What kind of fixtures can I get that are low profile that will have "true color" bulbs in them? I need to rip out existing and put in new immediately.

Forum Responses
(Finishing Forum)
From contributor O:
It should be so simple! The best "true daylight, color balanced" fluorescent tubes, which are readily available, are at best a compromise. GE lighting consultants state that "if a color sample matches under 'daylight' fluorescent (lacks red) and under incandescent bulbs (lacks blue), then it should match under all conditions. They further state that colors should be matched under the lighting conditions of the final installation. Our shop has large windows N and S, and we light with a mix of tubes. Won't even attempt critical matches under artificial lighting (at night). Test samples and target are often carried from unfiltered sunlight to areas with diffused lighting, with and without artificial lighting. Fabric, carpet and paper sample targets (which are interpretations at best) can be frustrating. Key is samples, as always, signed off on by final decision maker (bill payer). We've made sets of samples for client/designer to take to job-site and have them select one.



From contributor M:
Try checking the samples in natural daylight.


From the original questioner:
"They further state that colors should be matched under the lighting conditions of the final installation."

Shop is basically a cave with little to no natural light. I want lights in the area we are doing samples. Don't want to walk outside. This will be more realistic to onsite conditions.




From contributor T:
Ever notice that a car parked under a halogen street light is a different color than the same car parked in sunlight? Why? Because the spectrum of light emitted by a halogen light is very different than the spectrum of sunlight and your pigments will only reflect light from the spectrum that illuminates them. All different light sources emit different spectrums and will make a given sample look different - halogens are the worst. That's why they say match colors under the lighting conditions of the installation. Even light reflected from drapes, countertops, carpets, walls, etc. will have an affect. In other words, there's no way to get to where you want to go.

Sounds like you got bit by halogens. You may want to install some switchable halogen lights in the shop so if your client has halogen lighting (and only if they have halogens), you can come much closer to a match in the shop. Still no substitute for checking it in the actual installation lighting to be sure it's right.



From the original questioner:
11 years and first time ever having to strip finish. Arg. Almost positive it has halogen focal spotlights in the room directly above.

Pretty rare we have to do an exact match to something existing. It is benches for a table. I took leaf back to shop to match. Out of room, leaf has a green basecoat with red/amber stain. In the room, the green disappears on table, but not benches.

Shop is located 1-3 hours one way from 99.9% of all projects. That coupled with the fact of bad/no natural light is why I want something in shop. Perhaps a mix of fluorescent, halogen and incandescent together to get the full range? Is there such a thing as a full spectrum bulb?



From contributor J:
Due respect, but contributor T's got some bad information. You need to be looking for lights with a high color resolution index. This is a measurement of a light source's ability to let you distinguish between colors. Streetlights have lousy CRI ratings, but they are not halogen; they are HID (high intensity discharge). Halogen is a totally different animal.

Sunlight is the industry standard in defining CRI, and is considered to have a CRI of 100. Standard incandescent and halogen bulbs also have CRIs of 100, not because they make things look the same as they would in sunlight (they don't), but because they are just as good at letting you distinguish between different colors. Fluorescents, on the other hand, are all over the map, because the spectrum of light they generate depends on the particular mix of phosphors that coat the inside of the glass tube, and lamp manufacturers have all sorts of recipes. The best of them have a CRI in the low-mid 90s, and those are rare. Cheap ones have CRI ratings down around 60, which is a quantitative way of saying "god-awful." Halogen is amongst the best light to do color matching with, but it's an expensive way to light the whole shop. High-CRI fluorescents are a more practical choice.



From contributor A:
Sorry to hear about your problems. Although this is your first experience with this, it's not an uncommon problem.

The other problem you have is when you do a color match, the pigments you use are more than likely different from what's on the submitted color sample or part. That is why you need the sample to be approved on the job site with the correct lighting. Once you have it approved, then it doesn't matter what lighting you have in your production area, because the pigments are the same, the color will match under all lighting.



From contributor V:
If you can't match it in natural light, then the best light is the same light as where the piece is going. Fluorescent kills red. If you can, throw up some sky lights.


From contributor J:
Just to clarify/correct: street lights may also be metal halide or high-pressure sodium types. These are all very efficient lights with poor color rendering characteristics.


From contributor L:
4' full spectrum fluorescent tubes are readily available at Home Depot, etc. You want to look for the high Color Rendition Index (CRI) in the 90's. You can find these numbers listed on the packaging sleeve. Look for between 5000 - 6500K (Kelvins). The original pigments used and the pigments you are using are surely from different mines and will reflect light and color differently. Colors in the room and halogens are all able to throw the mix, but high CRI should be your best bet to get this out the door.


From contributor D:
You may want to either build or buy your own light booth, with a host of different light sources. We actually have two light booths, one we bought and one we built. We use both these booths, plus we set up different lighting in different areas of the building. Now that it is getting warmer, they will be outside checking their matches a lot more.


From contributor T:
I'd rather not debate whether some flood lights are halogen, and having been bitten by the halogen snake, I may be oversensitive to it. The point is: different lights emit different colors and that will affect color matching, as you know. You may not want to go out and check your match at high noon, but natural sunlight is the truest source of white light. A match in true white light will be a match anywhere. An alternative is the lights your diamond seller uses over his display counter, but I suspect they're not cheap. A more economical alternative lies in the responses of contributors O and D. If you can create a match that is good under two different light sources, it is probable that it will match under all light sources. If it matches under three different light sources (say daylight fluorescent, incandescent, and halogen), you'll have even a higher probability of a good match in all sources. So I would either build myself a light box with switchable light sources, or I'd light my color matching bench with switchable lights. It will get you closer to where you want to go, but it's no guarantee and doesn't replace the need to check it again on-site. However, it will improve your chances of getting it right the first time.


From contributor A:

Regarding this light issue... You will never match a color under all light sources unless the pigments used in both the submitted color sample and the control are exactly the same. I have matched at least 1 million colors, give or take a few, and have never, and I mean never, matched a color submittal under every light.

My best suggestion is this. If the color match is for residential, then match in natural first, then incandescent. If it's in a business, match under florescent. But it sure helps to know what the product is in before the work is done.



From the original questioner:
Thank you all for great advice. We busted out the halogens, took it out in the sun and used fluorescents when it came back to shop. 3 completely separate looks. We are starting from scratch on this one. From now on, I will add an extra flat fee for the time it will take to do field matching. The table is an opaque green look in fluorescent and transparent red/brown/amber in halogen. They definitely used an all-spray color system on the piece and we used penetrating stain and glaze.


From contributor I:
Okay, wood guys, here's the answer from the auto-body world which has to match colors a hell of a lot harder to match than wood tones. I do both, so trust me on this. Go and buy a 3M Sungun lamp, which matches sunlight very closely and is used in auto-body to achieve color matches on spray out cards. Cheaper than changing every light in the shop.


From contributor E:
After years of searching, the closest thing to outside sunlight/color match for us is a 4ft Phillips F32T8/TL865 Plus (Alto Collection). It's a 6500k daylight bulb for T8 fixtures. We get 25 bulb boxes for around 104.00 shipped from bulbs.com (or Google the part number). We use 24 in the booth (10 fixtures). I could not find these local or in any HD or Lowes type store.


From contributor B:
Would a skylight be possible in your spray area? We put a 4x8 skylight in the middle of our paint room four years ago. It cost about $300 and took about one hour to install. It has turned out to be one of the smarter things we've done.


From contributor R:
I think the new T5 bulbs are even better than T8's, and they put out more light for less money!