DV Info Net

DV Info Net (https://www.dvinfo.net/forum/)
-   Photon Management (https://www.dvinfo.net/forum/photon-management/)
-   -   Florescent and LED spectrum is nowhere near flat (https://www.dvinfo.net/forum/photon-management/493073-florescent-led-spectrum-nowhere-near-flat.html)

Tom Morrow March 14th, 2011 03:06 AM

Florescent and LED spectrum is nowhere near flat
I've been researching light spectrums for LEDs, florescents, and tungsten lights. I'm aghast at how non-flat the spectrum is for even the best of the LEDs and florescents. Even the best of class (Litepanels and Kino) are nowhere near as smooth as tungsten light when you look at the graphs.

In the case of flourescent there are huge spikes in the blue and green, and good LEDs have broad peaks in blue and then a dip, then another broad peak up towards red. Non best of class can have all kinds of nonlinearities in addition.

This echoes my experience, that I can get white to be daylight white fairly easily with tungsten or strobes, but with my medium-grade LEDs and florescents I'm going to need to do quite a bit of experimenting to find just the right gel to match daylight.

For those of you who successfully use LEDs or florescents, I am curious, are you able to mix them with daylight and still get whites, or do you custom-balance at camera or in post to get them white? And do you find that once balanced they give just as good color rendering as tungsten?

In other words, even if you get the RGB balance spot on, is it even possible to get really great color with spikey or peak/valley spectrums like these? I would think that if something in your scent happened to reflect the same wavelength as a flourescent spectrum it would glow unnaturally bright. The one thing I can't figure out from the graphs I see is how much energy is in those spikes... I know they are really high, but just how wide and how much area is underneath them? That is hard to evaluate.

It would seem that because LEDs generally don't have those huge florescent spikes, they should give better CRI once RGB balanced. But then why do LEDs seem to always be said to have lower CRIs than florescents by those tho know? Is it just the RGB balance that needs correction on the good LEDs or is there something I'm missing?

After buying and testing a few LED and florescent lights I'm considering selling them and going to tungsten because this problem is more intractable than I had imagined. I know part of the issue is that I'm not willing to spend for the best (Kino and Litepanels), but is this industry still so immature that medium grade stuff just can't be mixed with real daylight to critical standards?

I'm using ephotoinc LED 1200's, a Flo-light FB2500, and an assortment of 45w, 65w, and 85w CFLs.

Bill Davis March 14th, 2011 01:07 PM

Re: Florescent and LED spectrum is nowhere near flat

The major driving force behind both fluorescent AND LED lighting is not color spectrum. It's energy efficiency. And there are TWO very important elements to that.

In the field, the VASTLY superior energy efficiency yeilds significant cost savings in powering them - whether on location or on a set. In fact, LEDs are so energy efficient, that area lighting via simple lightweight battery powered units is finally possible.

In the studio not ONLY do you get direct energy savings by virtue of the minor power draw, but UNLIKE tungsten units, the vast majority of the energy in BOTH of these technologies goes into producing light - rather than HEAT - a factor that eats up MORE energy and drives up direct costs when it comes to studio HVAC.

Yes, there are spectrum issues. However, the judicious use of filtering (subtractive magenta for counteracting some of the green) - and more important, the ability of modern editing suite software to "re-color balance" in post - makes it perfectly possible to generate very good on-screen results with these amazingly efficient light sources.

If perfect color balance or absolute color reproduction is your holy grail - you may never be satisfied with these light technologies. If lights that do a decent job, sip rather than gulp power, and are much, MUCH, easier to pack and travel with than any tungsten kit on the planet are more important to your practice, than there's a place for all of these.

As in all lighting - horses for courses.

Tom Morrow March 15th, 2011 04:09 AM

Re: Florescent and LED spectrum is nowhere near flat
Yes, I am closing in on getting the color balance back to white with gels and perhaps some tweaking in post. But this hassle is nontrivial. And yes these are amazing energy savings.

But I rarely hear people talk about the way that skin tone looks more sickly, and green casts can still come about.

The spectrums can be seen overlaid on each other by going to this page:

Solid State Lighting Project | Science & Technology Council | Academy of Motion Picture Arts & Sciences

Then starting the "Summary" video under "Symposium videos" on the middle right of the page, and going to 1:37 in that video.

You can see that LED and Florescent have much different slopes and peaks than the film responses (designed to match human vision) which video tries to emulate.

The case study videos underneath also demonstrate the sickly skin tones that can come about, especially with lighting source number three, which is probably florescent.

Mike Watson March 15th, 2011 09:25 PM

Re: Florescent and LED spectrum is nowhere near flat
Looking at charts and graphs drives me crazy. I can't stop focusing on stuff like green spikes.

The same thing happened to me when I was looking at the 5DMkII when it first came out. How can anyone shoot with a camera with so much judder? I looked at youtube videos by the dozen where the photog was zoomed in on a brick wall and would swish pan back and forth - the lines between the bricks were so curved that you couldn't tell it was a brick wall anymore!

I ended up buying the camera anyhow - I've never replicated the problem in real life. Turns out I don't do a lot of swish panning in front of brick walls.

I was totally worried about the green spike when I moved to Kino Flo, I used to gel everything with 1/4 minus green. One day left that roll back at the shop. Thought we'd have to cancel the shoot. Decided to shoot it anyway just to see. Could hardly tell the difference.

Didn't even worry about gelling the LEDs. Love the flexibility to run them off 12V or in an office without a generator mucking everything up. The best move I've ever made.

One man's opinion.

Giroud Francois March 16th, 2011 02:56 AM

Re: Florescent and LED spectrum is nowhere near flat
technically fluorescent AND LED lighting are the same technology, so you can expect the same behavior.
What you get is is the light emitted by some layer of phosphor activated by some IR or UV light.
you can expect the led to be better only if you mix several leds with different spectrum (intentionally or being lucky).
for my home made led lightsm i use a mix of warm white (yellowish) and white (blueish) so the average is better than only one type of led.

Dean Sensui March 17th, 2011 04:13 AM

Re: Florescent and LED spectrum is nowhere near flat
Just FYI, even real daylight doesn't have a "flat" spectral content graph.

The best lighting tools get close, but not precise. Knowing where the deficiencies are, and your purpose, can help you make the best decision.

Oren Arieli March 17th, 2011 11:38 PM

Re: Florescent and LED spectrum is nowhere near flat
Just out of curiosity, where are you getting your spectrum graphs? I wish it was standard procedure for manufacturers, like including lumen rating and estimated bulb life.

Dean Sensui March 18th, 2011 12:01 AM

Re: Florescent and LED spectrum is nowhere near flat
I have a spectrometer that will create a graph of a light's spectral distribution.

I'm able to compare different light sources and see where they might be excessive or deficient in terms of wavelengths. It makes it a little easier to find the right filters sometimes.

John Nantz March 18th, 2011 10:52 PM

My LED work-in-progress experience (long)
Tom, if it's any consolation, I've been going through the same problem you are. As for background, I'm what one would call a hobbyist or amateur and got really bitten by the video bug last summer when I was given a free video camera. Since then in order to improve my videos I've acquired literally hundreds of dollars of equipment and the more I get into this the worse things are looking, and sounding (there's audio too). The past couple weeks has been my lighting nightmare from hell which is basically the same thing you're dealing with.

Here in the Great Northwest we're experiencing winter - you know, rain, rain, and more rain. And, it's cold and windy outside. This will continue basically until August when I can finally get outdoors and capture the video I've been dreaming about. But for the time being I decided to use my video lighting (three Cool Lights CL-600LED Floods) to take still pictures of stuff in the house in order to de-clutter the place. Now the problem starts: The pictures were too green. I tried editing in iPhoto but getting the right color temperature, tint, contrast, saturation, and exposure is, shall we say, a royal pain.

Plan A: I found that by using the camera flash I could mostly drown out the green but this caused other problems such as reflections from the subjects (items intended for sale on crags list and eBay).

Plan B: I found a Lowel L-light fixture for sale on crags list and since they were close by and it was cheap so I bought it with the intent of flooding the scene (dining room table with my items on a foam white board) using a fluorescent EiKO 5500K "daylight" bulb which is supposed to be one of the best (I was told). Well, this didn't work either because the pictures actually came out more green! It may be that the fluorescent bulb needed to be on longer to warm up more but I didn't test this aspect.

Research: A call to Lowel about what bulb they would recommend led me to their technical person who was a lot of help with explaining my problem but didn't have a particular bulb recommendation. The next call was to EiKO and I talked to Rick Porter, National Sales Manager for North America Photographic Bulbs. This was very enlightening and while I've been a little long-winded here, what I learned would easily double or triple what I've written. In summary, bottom line, for me the best solution would be to go with tungsten. Why? because (1) it is very consistent in the 2600K to 2650K range, (2) it is a known science, and (3) it has repeatable results.

Depression: So here I sit with three LED "daylight" light fixtures. What to do? Switch to Tungsten?

Masochist: Naaaa….. let's try one more thing. While the current pictures are for crags list and don't need to be that perfect, this IS laying the groundwork for my future videos.

Plan C: I didn't get any filter packs with the lights so that option was out until I had digested all this new information. Rather than switch to a new system I decided to order filter packs. They came yesterday and within 30 minutes I had run test pictures using no filter, 1/4-minus green, and 1/2-minus green. The results showed that the filters really helped but I'm not very satisfied because the 1/4-minus looks a little green and the 1/2-minus looks a little pink. Maybe that's because of my computer screen or the light I'm viewing the computer screen in, but that was yesterday. This morning the 1/2-minus seemed to look better.

Continuing Problem: The green spikes are clearly visible during the color editing process and I'm not very adept at removing them. One can increase or decrease certain things like the OVERALL color temperature, tint, etc. but, at least in iPhoto, not INDIVIDUAL, and especially multiple individual, areas of the color spectrum, and it's not good with dealing with spikes. Maybe a different and more expensive program ("modern editing suite") will but that's not on my option list. If I'm having this much trouble with still life I can only imagine what it's going to be like with video.

Plan D: Go to DVinfo.net and read what others have to say.

Bill Davis, Post #2, said it well. The LEDs have their advantages but a table-top "studio" (aka "wife's dining room table") in the hands of a novice is probably not one of them. Mixing light types someone said is not a good idea either. What Bill said about HVAC loads is not a problem here, well, at least not he AC part.

I can identify with Mike Watson, Post #4, about the green spikes driving one crazy - I can identify with that.

Currently I'm monitoring the local craigs list ads for some tungsten lights with soft boxes but in the meantime progress is continuing with my "close enough" product shots of "stuff" for selling but I'm very concerned about the next step which will be my dream video effort.

I want: a spectrometer like Dean has!

Oren Arieli March 19th, 2011 01:18 AM

Re: Florescent and LED spectrum is nowhere near flat
I'd love to have a high-tech spectrometer too....but there is a free alternative I found here:
A CD spectrometer

David W. Jones March 19th, 2011 09:23 AM

Re: Florescent and LED spectrum is nowhere near flat

Originally Posted by Tom Morrow (Post 1628034)
Yes, I am closing in on getting the color balance back to white with gels and perhaps some tweaking in post. But this hassle is nontrivial. .

I guess I'm just a little old fashion, as I have been gelling lights for over 30 years and have never thought that getting the right look was a nontrivial hassle. Unless you are a hobbyist it's part of the job.

With using fluorescent or LED lights, I simply use my old trusty Minolta Color meter II to quickly determine the appropriate gels needed to bring the lighting instrument into spec. It may be a 1/8 minus green along with a 1/4 CTO for example. For what it's worth... I own a variety of different lighting instruments, and have found that LED and Flo lights can be used to great effect once you learn how to balance them.

Good Luck!


Charles Papert March 19th, 2011 11:18 AM

Re: Florescent and LED spectrum is nowhere near flat
I can't tell you what the technical difference is between high and low end flos and LED's, but the bottom line is that the products packaged for the film industry photograph just fine. Most features and TV shows utilize both technologies and the skin tones are obviously acceptable there.

I use LED's from companies like Litepanels, Zylight and Litegear--the last being a fantastic system developed by an IA gaffer based around color corrected LED's which is a great problem solver for tight spaces. Just like the other manufacturers, these LED's are a lot more expensive than the similar looking variants but are suitable for lighting talent. I haven't seen information on whether or not they present spikes under a spectrograph, but all I know is that they look good in the final product, which is all that counts to me.

Paul R Johnson March 19th, 2011 11:53 AM

Re: Florescent and LED spectrum is nowhere near flat
I've kind of got used to all this over the years because I do theatre work where the main sources illumination is tungsten halogen at 3200ish, however theatre lighting tends to run at almost any point on the dimmer scale, rather than the TV method of a bit over and a bit under .7 on the dimmer - which is a manageable colour shift. The trouble for me is that we also now have lots of discharge sources at 6000ish and it's always difficult to know what to colour balance to. My usual method depends on if the show has follow spots, and white balance on their colour temp - this leaves everything else with a predominant red cast - but now we add in led lighting and it's yet another weird colour - the spikes aren't really a big problem, but they mean that it's almost as if they have a bit of gel on them. For us theatre types we can cure this fairly easily, because our LED fixtures are usually RGB (or even RGBA) so we can adjust the 'white'. We do now have some white LED theatrical fresnels, but these are a nuisance to try to return back to white. In most cases, the shift isn't too noticeable - especially if they're bright because they tend to squash the whites anyway when they are used in multiples.

The only time I see problems is with mixed face light.

John Nantz March 19th, 2011 12:32 PM

BINGO! (1/4-minus *plus* 1/8-minus)
Oren (Post #10): That was a great link to the CD spectrometer. Talk about a low budget setup and this has to be it. The examples were very interesting and example (12a), the Moon where he used a telescope was a good one because it shows that the light can be manipulated/concentrated and the spectrometer can still work.

By the way, tonight, March 19, 2011 (or maybe tomorrow night?) the moon will appear it's largest for over a decade. Details should be googleable.

Speaking of Internet research, I looked up David Jones' Minolta Color Meter ii (or II) and found a number of hits, including one at Minolta Color Meter II Review where he tells how his II is making a comeback. Used ones on eBay run in the $300 to $400 range and the III in the $700 to $800 range. Now that's getting up to what I paid for all my lighting!

Another item of interest is the daylight color setting that is designed into the camera. While talking to Rick Porter (see post #9) I learned that camera manufacturers use different daylight norms in the design of their cameras, and even within various models of the same manufacturer. Then add the computer program that is used to view the digital data and the computer itself, and now you have a real assortment of how a picture can be viewed, and this is assuming someone doesn't have eye issues.

Also, David Jones (#11) said he might use a 1/8-minus along with the 1/4-minus and based on my test results I think that will be my solution too.

Jad Meouchy March 19th, 2011 03:27 PM

Re: Florescent and LED spectrum is nowhere near flat
Great points all around with regard to the absolute accuracy of lighting, but it's missing the point. With more efficient lighting, the scene/room/studio stays a cooler temperature (and I'm not talking color temperature). That means my talent doesn't sweat as much and I don't need to slow production to do touchups every other take.

And less chance of a fire or injury means lower liability exposure, cheaper insurance, etc. A slight shift in color won't get noticed by 99.999% of the audience and if the performances are better as a result, then the product has improved.

I'm not crazy about LED/fluoro lighting from an optical standpoint but it makes up in other areas.

All times are GMT -6. The time now is 12:23 PM.

DV Info Net -- Real Names, Real People, Real Info!
1998-2020 The Digital Video Information Network