Florescent and LED spectrum is nowhere near flat

[Warren Kawamoto]

Here's an easy test for everyone.
1. Get a color chart, one that shows vivid red, green, blue, cyan, yellow, and magenta.
2. Light it with tungsten 3200K, then white balance for 3200K.
3. Look at the results in post. (Preferably with a vectorscope) The colors should look fairly accurate, and can easily be corrected in post.

4. Light the same color chart with LED or flo, then white balance.
5. Look at the results in post. In most cases, you'll see some colors that aren't quite right.

If you try to color correct for those inaccuracies, you'll throw off the balance of everything else. Your conclusion will be that obtaining accurate color with LED or flo is impossible or near impossible.

[Bill Davis]

...then move on to PART TWO of the mental exercise...

LOOK at your scene. If you're shooting fashion and you have a bunch of purple and ecru scarves that MUST render color accurately - then bite the bullet and leave your limited spectrum choices behind.

But if you're doing a video interview with anyone OTHER than Elizabeth Taylor (those violet eyes?) then save the weight, power draw, and hassle of taking along gear that is over-heavy and inefficient since you can likely get a PERFECTLY acceptable interview shot - and save yourself a lot of hassles in the process — with Divas or even LESS expensive and power hungry LED solutions.

Sure you MAY knowingly devalue PERFECT color rendering. But sometimes that's a perfectly sensible trade, particularly if you know in advance that the the specific spectrum colors that a technology does not do particularly well are NOT a critical part of your shots.

THIS is the point of experience, folks. It's knowing enough to evaluate a scene - figure out a proper approach to that particular scene, and then to choose to do that which makes the most sense for THIS shoot in THESE conditions.

Or heck, go ahead and believe that your venerable Arri Tungsten kits, "portable" gennies and half a dozen sherpas is the ONLY way you can get great results on your next video gig. I just hope for your sake that it's not a gig titled - "Let's Go Climb Yosemite"

They knew better a few hundred years ago when somebody first mumbled "horses for courses."

Simple as that.

[Tom Morrow]

Bill, your "horses for courses" phrase has been echoing through my mind many times, and is I think a good way to approach these deficiencies. Everything has deficiencies, and it's a matter of choosing which ones you can get away with for a particular project.

I just got an SSD hard drive for my laptop, and love it, even though it has some glaring deficiencies like small capacity and high cost.

Guy has done some really great work in putting this information together in a digestable format, and I think it's useful especially for those who work with LED and flo lights, to understand what issues to look out for and work around.

I just bought a third LED panel from the same manufacturer as the first two. I was very glad that I had the opportunity to go into the store and look at these in person, because I discovered that the smaller size I was planning on getting had a significantly different color, different enough that it would have made mixing them problematic. This despite the fact that these were all lights in the same series from the same manufacturer. Even the store employee was surprised to notice the difference.

I instead got a third LED panel in the same exact size/model as the first two, and indeed the colors match. But the uniformity of the LEDs is much worse than the first two panels; some LEDs are brighter than others yielding a speckled quality, whereas the original ones I had are uniform. It doesn't affect functionality, but it goes to show how LEDs from different lots clearly are not the same, even when from the same source. Guy's paper goes into the issue of sorting and binning LEDs to achieve consistency.

No need for us to disagree. There can be serious issues and still be good enough for many purposes.

[Bill Davis]

Tom,

I agree with what you wrote wholeheartedly.

The big issue is that it's an age where many people are coming to video production with a "do it yourself" mentality - which is excellent. However, often this is combined with a "just point me to the piece of gear that will do this thing I want done automatically, and professionally, whenever I throw the switch" - and such things have always been fantasy.

THAT is the problem. People think that making video is done with things akin to an iPod, the magic of which is that it appears at first blush to be a stunningly easy machine to operate. Press a button and you get beautiful music. You do - and that instant satisfaction is sweet. But as you get to know it, you face the issues of music purchase, library management, interfacing it to something other than headphones, battery life - and the seduction of simple yeilds to the complexities of the reality that you're doing something much more complicated that it looks at first blush.

Lighting is the same. The new folks want "buy this LIGHT you can light anything." and that's the same as saying "buy this piano and you can play anything." It's true in a sense. It just neglects that small step of actually becoming a piano player!

[Charles Papert]

The piano player analogy is apt, but I would suggest that the iPod one is less so, in that the structure of working with an iPod is essentially finite. Yes, there is more to it than pushing one button, but it can be mastered by most in a fairly short amount of time, and once you learn the structure, it's a matter of simple repetition to achieve the desired results.

Lighting is an open-architecture pursuit; while there are basics to learn and apply, there will always be variations presented by circumstances that can and should inspire one to always keep trying new things. I don't doubt that the vast majority of shooters go about lighting headshot interviews exactly the same way every time, and that being one of the more simple setpieces in the book, it's not surprising. But at the same time, those who care to pay more attention learn that the exact height and angle of each source will have a specific effect on the results, and because each person is different, it's critical to understand this to make them look their best.

Agreed that there is a perception out there that there must be shortcuts to all of this, but fortunately some aspects of filmmaking are immune to the immature, entitled thinking that advances in technology have made filmmaking "easy", and lighting is one of these. There is no substitute for learning by doing and observing.

For me, the very fact that I have and probably never will "master" the art of lighting is what keeps me interested. Every setup I shoot, I draw on 25 years of personal experience but the results are rarely 100% predictable and I'll take away something new every day.

For the record, for interiors and night work I still light primarily with tried and true tungsten, but I like the space-saving and speed aspect of Kinos and the even faster and more flexible aspects of LED's.

Next time I'm shooting camera tests I'll have to experiment with each of my LED units and see how they fare with the Macbeth charts.

[Brian Drysdale]

You really need to test with the meduim you're using. Video cameras have traditionally been pretty blind towards the green in standard fluorescents, whilst with film the green does come out pretty strongly. You need to test with the camera you're using to see how it's responding with the greens, some of the new digital cameras may pick up green because they not limited to rec 709.

[Guy Holt]

Originally Posted by Tom Morrow

Everything has deficiencies, and it's a matter of choosing which ones you can get away with for a particular project.

No truer words have been spoken, but there are a couple of important differences between Kino Flos and LEDs that are pertinent to this discussion. One: the discontinuous color spectrum of Kino Flos can be easily corrected with gels, LEDs can not, making Kino Flos a better key source, in my opinion, for documentary interviews because they will render flesh-tones better. Once the green spike of Kino-Flo’s True Match tubes have been corrected by the application of minus green (magenta) gels, the resulting spectral distribution is nearly continuous and contains a greater proportion of the long wave length colors that are so critical to rendering flesh-tones accurately. As can be clearly seen in the side-by-side comparisons in my newsletter article, skin-tones are significantly altered by the steep drop-off of long wavelength colors in LED light sources. Kino-Flo’s high CRI True-Match tubes, on the other hand, contain sufficient light in that critical part of the spectrum to render skin-tones realistically. Once corrected for their green spike, Kino Flo True Match tubes provide a nearly full spectrum source. With Phosphor White LEDs, it is nearly impossible to correct for their deficiencies with gels. To see side-by-side comparison go to ScreenLight & Grip's E-Newsletter.

The second distinction is that while the out-put of both sources depreciate overtime, when you reach low light failure of a Kino Tube after 2500hrs you can simply replace it. The useable lamp life of Phosphor White LED luminaries designed for motion picture lighting applications are probably no more than 1500 hrs. Since these types of LED fixtures have no interchangeable parts that can be replaced after reaching low-light failure, after 1500 hrs the fixture can only be thrown away (see ScreenLight & Grip's E-Newsletter for more details.)

- Guy Holt, Gaffer, ScreenLight & Grip, Lighting Rental & Sales in Boston.

[Richard Andrewski]

We've had LED 600s out for 2 1/2 years now. No "low light failures" that we've heard of. The LED 600 is completely modular and if a customer wanted to upgrade or change out an LED panel, dimmer or battery interface for whatever reason they could even do it themselves or have us upgrade it for them for reasonable fees. I anticipate that as LEDs improve in output or spectrum, people may do just that.

As far as color is concerned, see post #19 in this thread. Bill Davis, who Guy told to "forget about Cool Lights", was showing some work he did with our LED 600 which he's had since they came out. Said he didn't even use any correction on the fixture.

http://www.dvinfo.net/forum/photon-m...irdness-2.html

I don't see anything wrong with that photo, where's the problem?

[Guy Holt]

Originally Posted by Richard Andrewski

“Most people using digital mediums today and LED fixtures are not having a problem, are white balancing and using filters and not seeing any big issues.”

Then why are online production forums filled with posts like the following:

“The few times I have been around LED's, the green spike or blue leaning color issues were unacceptable. Two weeks ago I was on a doc/interview shoot that was using some and the skin tones on the RED's monitor were very green.” Read more: Any truly color accurate LED panels out there? - Cinematography.com

The problems with LEDs go beyond color rendering. We have noticed that they cause interview subjects discomfort. John Rossetti, a well known DOP in London recently related on the Cinematographer's Mailing List : " It does not matter how many LED's there are in the head, its still a pointsource, from the talents point of view.... I have had some well known names refuse to be lit by them" (full account available at Message.) Having personally lit a number of Senators and even one President I know that the rendering of skin tones and the multiple shadows created by LEDs are on the radar of political media consultants. How are you going to look when the media handler for a Senator tells you that you cannot use them to light the interview.

Originally Posted by Richard Andrewski

“They (AMPAS) seem to want to say in the study that filters don't work and yet people using real film have been using this filter method for decades to correct on location lighting like parking lot or office flos, street lights, etc. How was it they got along till this study came out to tell them they're wrong?”

The industry didn’t “(get) along till this study came out to tell them they're wrong?” Anyone that has worked on dramatic productions, be they well crafted dramatic shorts, commercials, episodic television, or feature films shot on film or digitally will attest that they have NOT been “using this filter method for decades to correct on location lighting like parking lot or office flos, street lights, etc.” I have spent many an hour changing out office fluorescent tubes for Movietone tubes, changing out CFLs for PH bulbs, replacing sodium vapor bulbs with tungsten bulbs in streetlights, and wiring tungsten peanut bulbs behind lamp posts. Why? Because the color deficiencies of these light sources could not be corrected by simply slapping on gels. The AMPAS study only confirms in regard to LEDs what we have known for years about these other light sources.

Originally Posted by Richard Andrewski

“With digital you can see the results immediately if you want and correct anything right away.”

If only this were true. Unfortunately the reality is that camera white balance CAN NOT correct for the deficiencies of LEDs in every situation. Take Daryn Okada’s situation above. Had Daryn Okada been shooting with a digital video camera, he would have noticed the off color of the LED source immediately. But, given the limited spectral output of LEDs, his ability to remedy the problem would have been limited. If he white balanced the camera for the LED source, the background of the room beyond the doorway that was lit by tungsten lights would turn very green. In a mixed light situation such as this, the only alternative is to match the LED source to the prevalent tungsten source with a custom gel pack on the LED head. But, since gels rebalance the spectral distribution of a light source by passing the wavelength of the color that they are, it is not practical to use gels to correct for these deficiencies either because there is not sufficient light of those wavelengths to pass in White Phosphor LEDs to begin with.

If the actress in Daryn Okada’s shot were the model in the blue/cyan dress posted in my newsletter article, you can imagine what would happen when she stepped onto the mark lit only with the LED source by looking closely at the contrasting photos in my newsletter article. Since, under the circumstances, Daryn Okada could not white balance for the LED source (and it would not be feasible to match the LED source with a gel pack) the rich blue/cyan hue of her dress in the left photo would turn into the simple blue of the right photo. The vibrant skin tone of the left photo would turn into the flat skin tone of the right photo, and it would have an overall magenta cast to boot.

The bottom line is that color gel packs, camera white balance, or digital intermediate timing can’t bring out a color if it isn’t there to begin with. Simply by nature of their discontinuous spectral distribution, even high CRI Phosphor White LEDs will never accurately reproduce colors on screen regardless what you do on set or in post. Kino Flos, on the other hand, have a much more continuous spectrum. Once you filter out their green spike with minus green (magenta gel), they are able to render flesh tones more accurately than LED lights. For details, see ScreenLight & Grip's E-Newsletter for more details.

Originally Posted by Richard Andrewski

“Flos and LEDs are basically the same technology so why would you expect anything different?”

Only in so far as humans are basically the same as whales because we are both mammals. You make the same self serving reductionist argument on your website in regard to HMI and your CDM (Ceramic Discharge Metalhalide) fixtures. LEDs and high CRI Fluorescents couldn’t be more different. For one thing, the discontinuous color spectrum of Kino Flos can be easily corrected with gels, LEDs can not. Once the green spike of Kino-Flo’s True Match tubes have been corrected by the application of minus green (magenta) gels, the resulting spectral distribution is nearly continuous and contains a greater proportion of the long wave length colors that are so critical to rendering flesh-tones accurately. As can be clearly seen in the side-by-side comparisons in my newsletter article, skin-tones are significantly altered by the steep drop-off of long wavelength colors in LED light sources. With Phosphor White LEDs, it is nearly impossible to correct for all their deficiencies with gels.

Originally Posted by Richard Andrewski

“Cool Lights LED fixtures weren't a part of the AMPAS test--not sure why we get several mentions in Guy Holt's post”

Richard as much as answers his own question in the same paragraph when he states “the spectrums of most LEDs are similar but there will be some that are better than others.” The Litepanel and Arri LED panels used in the AMPAS tests being the better ones because they use more stringent criterion in the binning of the LEDs they use which accounts, for a large part, why they are more expensive. I mention Cool Lights LED panels among the others because they also use Phosphor White LEDs with the same deficiencies in color output.

Originally Posted by Richard Andrewski

“As far as color is concerned, see post #19 in this thread. Bill Davis, who Guy told to "forget about Cool Lights", was showing some work he did with our LED 600 which he's had since they came out. Said he didn't even use any correction on the fixture.

(http://www.dvinfo.net/forum/photon-m...irdness-2.html)

I don't see anything wrong with that photo, where's the problem?”

For one, he looks awfully pasty and pale for a Texan. The skin tone in Bill’s photo is consistent with the results one would expect when lit with an LED source. Absent the long wavelength colors above the 600nm cut off of Phosphor White LEDs, the flesh tone lacks “color” or vibrancy. I don’t deny that LEDs can be a valuable tool in certain circumstances. For a specific application, say where bulk and weight when traveling is a factor, a LED fixture offers the unique advantage of being compact and lightweight (but then so are Kino BarFlys.) Under the circumstances, for Bill this advantage to LEDs out weighed their shortcomings in color rendering. Bill says as much in this thread:

“… you MAY knowingly devalue PERFECT color rendering… sometimes that's a perfectly sensible trade, particularly if you know in advance that the specific spectrum colors that a technology does not do particularly well are NOT a critical part of your shots.”

For example if you were to look at the image of model in the cyan/blue dress lit by Phosphor White LEDs in isolation, you wouldn’t know that she doesn’t look like that and so wouldn’t think it was a “bad” likeness. But, when viewed side-by-side with the same image lit by a continuous tungsten source, you see what it should have looked like (the side by side comparison is available in my newsletter article.)

As Bill rightly points out, the value of experience is knowing how “to evaluate a scene - figure out a proper approach to that particular scene, and then to choose to do that which makes the most sense for THIS shoot in THESE conditions.”

When I read posts, like Bill’s (#32), from working professional DPs (as opposed to newbies and salesmen), I don’t hear them saying that “people using digital mediums today and LED fixtures are not having a problem, are white balancing and using filters and not seeing any big issues.” I hear them saying there is a problem. So choose the appropriate lighting technology for the job and clearly Phosphor White LEDs are not the best choice in applications where color rendition is critical. For instance, Bill himself says in this thread: “If you're shooting fashion and you have a bunch of purple and ecru scarves that MUST render color accurately - then bite the bullet and leave your limited spectrum choices behind.”

I, for one, wouldn’t try to light a table-top food/product shot with them. Because of their limited color rendering capability, food presentation that will look vibrant and colorful to eye under LEDs, will tend to look a little dull on camera. By comparison a full spectrum daylight source such as HMI or LEP will capture the vibrant colors.

Likewise, when I lit Ted Kennedy in his old age and his media handlers wanted to convey the impression of a robust and healthy elder statesmen, I fought to use an HMI interview kit, with its additional weight and cost, when the show producers wanted me to use a cheaper LED kit. To convince the Producers I brought the LED kit, and after seeing the results they came up with the extra money for the HMI kit because it was important to render a healthy flesh tone.

I also wouldn’t try to mix LEDs with a uniform continuous light source, such as a studio lit with tungsten fixtures, where if caught in isolation as in Daryn’s case, their color deficiencies will be quite noticeable and unacceptable in comparison.

I am not saying that you can’t use LEDs successfully. As long as you understand the limitations of the tools you have at your disposal you can do good work. But, in the case of LEDs it helps to have a thorough understanding of the technology to avoid bad situations like those discussed above. For this reason, I have put together an overview of the technology for our company news letter (available at ScreenLight & Grip's E-Newsletter.)

Guy Holt, Gaffer, ScreenLight & Grip, Lighting Rental and Sales in Boston

[Charles Papert]

Guy, I wasn't quite sure how to evaluate the RGB LED results in the graphs you posted--seemed like the spectrum was even choppier than the white ones. Have you tested the Zylight, for instance?

p.s. you get a special bonus for bringing your answer in under 500 words...!!

[Guy Holt]

Yes and Yes. Now what's my special bonus?

- Guy Holt, Gaffer, ScreenLight & Grip, Lighting Rental & Sales in Boston

[Charles Papert]

Satisfaction in brevity!! ha.

So but really--you no likey?

[Richard Andrewski]

As I said earlier, we already told people about all this on forums such as this one as early as 2008, our blog in 2009 at the release of our LED products, and we continue to say it today: LEDs aren't perfect, that's why we sell other products because no one product does it all. If someone comes to us and they aren't sure what they need for their uses, we question what's most important to them and recommend products based on those needs. If battery use and portability were of utmost importance, LEDs are the main recommendation. Particularly for ENG type uses. If higher CRI is of concern, there is always fluorescent or HMI. Some still like tungsten and far be it from me to talk them out of it. However, IMO tungsten, which is effectively 98 CRI, isn't necessarily the best for shooting digital today nor can we be totally sure its an accurate reference for lighting tests using digital mediums. Maybe it will be and maybe it won't--totally depends upon whatever camera you happen to be using and how well it's sensor and internal processing of images works with tungsten lighting. CRI would be bumping into the limitations of the sensor technology in that case.

Beating up LEDs because they aren't in the mature stage of development (like fluorescents for example which have been around for decades), is like beating up a circa 1990s Pentium because it wasn't a multi-core processor like what we have today. There is a march of progress and it may not be fast enough for some, but it is what it is and not everything happens overnight. Most breakthroughs only come with plenty of steps before. Same for a 4 core processor during its time. The I7 will be a 'has been' soon too. Still that Pentium was plenty useful for most that used it during its time. For those that needed more they paid a huge price for a multi-processor back then. Its all part of evolution and products gaining maturity and finally that lowering of cost which only happens if people buy into the development cycle and fund it buying whatever the current possibilities are (with whatever limitations they may have). In other words, there would be no high CRI LED (or LEP bulbs for that matter) if there wasn't anyone buying their predecessors. We first mentioned LEP bulbs back in 2008 on forums like this and were looking into it but they weren't quite ready for film and television use back then. Now they've gained some ground, so we'll look at it again. You can thank manufacturers and their customers buying those previous versions for uses other than that (like stage or street lighting for example--accepting the limitations it had) for the newer models which have less limitations. Some people may even be beating up on this technology (and providing really long write-ups) because they want to sell you one of their products. Internet sales copy is awfully long sometimes because they want to tell you about a problem you may not have even known you had and they just happen to have the solution for you...

As far as whether a high CRI LED can or can't render colors: when a true high CRI LED shows up, and its not all that far off, then they will render colors every bit as well as the best flos do today. The LED phosphors that can do this well have only recently shown up on the market. Thus, any panel that claims its high CRI (before the last year or so), is not really high CRI.

As I said many times here, my blog, other forums, if an LED fixture claims to be high CRI and you have color rendering or bias issues, its simply not using a high CRI LED. In other words, if you test it and proclaim to the world "this HIGH CRI fixture isn't rendering colors all that well, so I guess CRI as an index doesn't work!", you may have reached the wrong conclusion. The conclusion should have been "this fixture isn't really high CRI and therefore its not rendering colors all that well--I guess the specs were wrong." Or it could be "I'm using film so I have different issues to worry about than digital users."

In fact, if the fixture had a visible green bias you should have already understood it wasn't high CRI because a light that is high CRI wouldn't have visible green bias (however, a light that is "high CRI" for film use may well have a bit of magenta bias to over correct for the green spike). Again film and digital have such different behaviors that you must be more careful with discontinuous lighting for film use and the AMPAS study was mainly concerned with film use.

[Paul Cook]

All this knowledge is melting my head...I thought I was so clever when I found out about CRI many years ago - running around spouting the importance of high CRI lights to anyone who'd care to listen.

I notice that the article seems to suggest Kino's are an OK fluorescent but given the overall cost - I wonder if anyone has compared them to a simple electronic ballast and a couple of Phillips TL950's? They supposedly have a CRI (I know, I know) of over 93 and you could build a bank of 4 of those for under $200 - so if they are close to Kino's...?

I would also be curious to know if anyone has measured light from strobes for still photography. Where do they rank on the spectrometer? Something like a Nikon SB900...how does that measure on a spectrograph for colour? Do the studio strobes that cost thousands more output a better quality light or are they simply a more powerful light source? What about those cheap studio strobes from China that actually cost less than an SB900?

So much to know...so little time :-)

[Bob Grant]

Can someone please recommend a colour meter that works?

A couple of days ago I tried the now discontinued Gossen and at first I assumed it was either broken or out of calibration as both a tungsten lamp and a "tungsten" LED light read with the same "20 G" error. Giving up on that I checked out the range of available meters from B&H and felt the Sekonic C-500 worth a look but at least one user review notes it seems to have the same problem as the Gossen I assumed was on the fritz. The reviewer also asks a relevant question "why does it read a different CT value in film and digital modes?

The purpose of my question is to obtain a meter to be used to evaluate various lighting instruments. A full blown spectrometer is not out of consideration. My previous venture into a search for a suitable instrument for ploting filter attenuation was frustrating to say the least.