[UnknownVT]

Although actually I think blue or green gels are worse.

Blue and green don't look that good on skin tones -
but do not pose the technical difficulties in photos in areas such as AF, CA and JPG compression.


this is a shot from the same night/venue with blue and green LEDs only on the main subject - there is definitely no problems in definition - either in the original photo or even when compressed in JPG to my normal quality

Blue only tends to have some difficulties -
but no where as much as magenta made up with Red and Blue LEDs - for all the reasons given in this thread so far.

This lovely artist had the misfortune to have have her entire set lit with blue (LEDs) only:

This photo technically is kind of OK - just not very flattering to her.

However with the much maligned technique of chimping -
I saw this problem almost straight away, and was able to mitigate it somewhat:

a little psychedelic and Andy Warhol-ish, but better, at least to my eyes -

ah! you cry - wasn't she lit entirely by blue LEDs -
so where do the reasonable flesh tones come from -
post-processing in magical RAW?
No way! even RAW can't do that -

I "lit" the scene with the K-x built-in flash to compensate.

BUT that does not look like a "Flash" shot, hopefully some are saying......
well it's -2/3 stop flash compensation - which on its own didn't do it -
I had to dial in -1 stop overall compensation -
to capture the in-situ blue lighting and merely fill-in enough to light the face to get my "pound of flesh"-tone that is....

[dragonfly]

Here's an example of some unflattering lighting:

By desaturating and reducing the vibrance I was able to compensate for the lone pink light. Some minor tint changes were made to try to make the flesh tones appear normal.

I only did this in a couple of minutes, so the results could be better, and I'm not a pro, but the results are a lot closer to what I remember seeing than the camera.

The image isn't as colorful as the scene was in real life, but most viewers won't be aware of that. It looks realistic to me, and that is important.

I think you get more options and more latitude with RAW files, which is especially important with weird scenes like these.

[UnknownVT]

I think you get more options and more latitude with RAW files, which is especially important with weird scenes like these.

First well done on the shot.

Without a doubt RAW has more flexibility -
all I was saying is that RAW cannot work miracles -
if the same shot under red and blue LEDs only -
even RAW cannot recover the skin tone that wasn't there in the first place
(one needs some green in the mix minimum to do anything -
other than artificial tinting, then one can do that in JPG too)

Your shot had the saving grace of having a pink light in the mix -
plus it may have been some gel'd tungsten/halogen (I am only guessing) -
so the overall spectrum was wider and more continuous -
that's why you were able to salvage the shot -
that is not taking anything away from the good work you did.

Most of the time I am not complaining about the colors par-se -
I tend to prefer to present the shot as I see them
and do as little PP as possible -
BUT magenta light made up of red and blue LEDs - plays havoc with photos -
both as a shot and JPG compression -
so I canNOT easily present the way I see it -
that was the whole point of the post in the first place -
not that I wanted to present realistic flesh/skin tones under any and all circumstances.

However I accept that is just my personal take and preference in photography.

Here's your colorful Red JPG shot manipulated to get approx the same skin tone -

I left in the EXIF data to show that it was the colorful Red JPG I manipulated to get similar skin tone -
so this is from JPG - one can also tell it was the Red JPG I manipulated as the upper cheek and forehead has lost detail because of the original Red shot's JPG compression

[dragonfly]

OK, now I see what you're saying. This advance in stage lighting is another element to learn to adjust to. Or stay away from.

[RolloR]

try to set a lower exposure, use M mode and not Av since narrow spectrum lights trick the light meter of the camera.

halogen gels, leds, etc in the mix of lights:

try not to recover the natural skin tones, use the colors to your advantage


this one's a bit over-exposed, but still ok imo




magenta tends to be a "clipped highlight", so lower the exposure a bit



and example of over-exposed "magenta lighting:"

it clips nastily, and you guys are right, it gets even worse when compressed to JPG. The only real solution I have tried is to lower the exposure a bit. Or wait til the lights change color lol. Magenta's not very pretty anyway

[Marc Sabatella]

Blue and green don't look that good on skin tones -
but do not pose the technical difficulties in photos in areas such as AF, CA and JPG compression.

True. I guess I was thinking more in terms of how easy / possible it is to recover anything resembling natural skin tones from the image via WB, hue, and saturation adjustments. And to make matters worse, my experience is that it takes a slow shutter speed to get usable exposure for the same "apparent" intensity of light if the light color differs so strongly from the local color (eg, blue/green light on skin). And trying to complete desaturate a blue- or green- lit image of skin tones to create a B&W image tends to really robs you of light to a great extent than desarutating a red-lit image of skin tones, meaning you're left with a really dark exposure that needs more pushing than otherwise to be usable - meaning also noisier results. Although I haven't done a really controlled comparison to back up these impressions. Fortunately, blue and green lighting tends to not be as common, and when I have encountered it, it's mostly been in situations where the lights rotated, and a few seconds later I had orange or yellow or magenta.

[UnknownVT]

when I have encountered it, it's mostly been in situations where the lights rotated, and a few seconds later I had orange or yellow or magenta.

Ha-ha! you must have the same light person....
when they see us point the camera -
they change the lights!

---

Last edited by UnknownVT; 03-17-2010 at 02:46 PM.

[UnknownVT]

Without a doubt RAW has more flexibility -
all I was saying is that RAW cannot work miracles -
if the same shot under red and blue LEDs only -
even RAW cannot recover the skin tone that wasn't there in the first place
(one needs some green in the mix minimum to do anything -
other than artificial tinting, then one can do that in JPG too)

I'm kind of about to eat my own words about RAW -

I have this shot that I took in paired DNG and JPG -

which has the magenta (red and blue LEDs only problem)
the outlines shows the areas I concentrated on.

First I used ACR 5.6 (Adobe Camera RAW) in PhotoShop Elements 7.0
(note: ACR 5.6 is the latest version which supports the Pentax K-x)


Left: As Shot settings ............................................................. Right: set White Point on the shirt.


Left: As shot settings ............................................................. Right: set White point on shirt of bassist as previous sample.


One can see even with "manual" white point selection there isn't that great an improvement and this is with ACR 5.6 the latest version that officially supports the K-x.

This is why I said RAW can't do that much better.

But I thought I'd play around with Pentax Digital Camera Utility 4.11 (SilkyPix)
and I was surprised by the results -


Left: Camera Settings ............................................................ Right: set Grey Point on bassist shirt.

WoW! I didn't think this was possible, since I thought the main subjects lit with mainly red and blue LEDs would lack anything in the middle of the spectrum, and especially since ACR didn't improve things that much.

Harder test was the drummer -

Left: Camera Settings .......................................................... Right: Grey Point on bassist shirt as sample above.
No this does not show any great quality -
but I am really impressed in the ability to get something that even resembles skin tone here.

Not that I really want to present realistic flesh tones in a shot such as this - but I am really surprised that the Pentax Digital Camera Utility/SilkyPix (4.11) can actually manage even this much.

Here's the overall shot that was balanced by Pentax DCU/SilkyPix -

straightforward adjustment to brightness contrast and a bit of sharpening but I had to "deselect" the bassist's shirt to bring up the brightness in the rest of the pic

[falconeye]

But I thought I'd play around with Pentax Camera Utility 4.11 (SilkyPix)
and I was surprised by the results -

Thank You for your experiment.

I now tried with some of some of my own stage shots and can confirm your findings.

LR 2.6 seems to be limited to 2000°K color temperature. If red or magenta is the dominant color, LR will fail. Partly, as has been said, because of a lack of green.

SilkyPix has the same limit (2500°K actually), but ignores its limit with the gray color tool. As a result, it can push colors which only have a very low amount of green into gray.

In my own photos, much more fresh colors and more realistic skin tones have been the result. Better color contrast meant sharpoer looking images as well.


There is one caveat though!

With little green only, very dark images are the result and I had to push +2EV, which in my case meant to push ISO 1600 to ISO 6400 (K-7). I then found the SilkyPix noise reduction to be insufficient.

Therefore, this is my recommendation for an important false color stage shot workflow:

1. Shoot at less ISO than you would do normally (1 to 2 stops better).
2. LR, check that white balance indeed clips at 2000°K.
3. SilkyPix, gray tool white balance (a gray card will help) and exposure push.
4. If 1. was impossible, now run a decent Noise Reduction (Dfine, NN, etc.)
5. Reimport into LR.

[Entropy]

Some of those pictures look like exposure for one channel was clipped, so the reccommendation of someone earlier to drop exposure a bit makes sense.

I also liked one of the other ideas that someone showed an example of - use a strobe for fill. If the strobe is only used for fill and not primary lighting, it won't have that obvious "flash photography" look, especially if you move the strobe off camera.

One thing, the color temperature of nearly all stage lights is much lower than typical strobes. Put a CTO gel on the strobe! I've had good results of mixing stage ambient and strobe by putting a CTO gel on my AF-540, although in many cases I probably could use an extra 1/4 CTO, stage lights are often VERY orange even when ungelled.

[UnknownVT]

I now tried with some of some of my own stage shots and can confirm your findings.

LR 2.6 seems to be limited to 2000°K color temperature. If red or magenta is the dominant color, LR will fail. Partly, as has been said, because of a lack of green.

SilkyPix has the same limit (2500°K actually), but ignores its limit with the gray color tool. As a result, it can push colors which only have a very low amount of green into gray.

Thank you for your valuable input.

I have numerous attempts in both ACR (Adobe Camera RAW) 5.6 and Pentax DCU/SilkyPix of changing the color temperature and even playing with the "fine tuning" in DCU of giving max green balance and maxing out amber - but all those led to unsatisfactory results similar to the ACR results I showed above.

Then I hit on using the Grey Point in DCU/SilkyPix and not expecting much - Boy! was I surprised - I was eating crow.... (raven, corvus, corvidae - and all manners of those pesky cawing black colored birds..... )

I really didn't think such a thing was possible - but there it was staring me in the face - well actually I was staring at the results.

I have played with various RAW processors including LightRoom (3 Beta) - and so far Pentax DCU 4.11 (SilkyPix) is the only RAW processor that seems to be able to recover to a reasonable semblance of real flesh tones even with extreme lighting such as the magenta made up with narrow bandwidth red and blue LEDs - where even the very best in class RAW processors like ACR (Adobe Camera RAW) 5.6 and LightRoom (3 Beta) could not (although I strongly suspect LR uses the same technology as ACR - since they are both from Adobe)

Some of those pictures look like exposure for one channel was clipped, so the reccommendation of someone earlier to drop exposure a bit makes sense.

This is the difficulty with extreme lighting such as magenta made up with very narrow bandwidth red and blue LEDs - either of which can be considered monochromatic (no, not black and white - but of a single wavelength) which would give absolutely no leeway in color adjustments other than color substitution.

The K-x exposure was about as optimum as I could get it - the clipping one is seeing is partly the JPG compression (which loses details in the magenta areas) and the difficulty of presenting magenta on the screen - if one looks at the close up details there is plenty of detail albeit somewhat lossy due to JPG compression.

If anything, as falconeye pointed out above, one might have to increase exposure somewhat - as a recovered to flesh tone shot was pretty dark - like his own experiment confirmed - but as you point out if one does that an as-is shot would have the colors clipped.

So I think (without the benefit of further experimentation) that it is probably better to use normal exposure and up the brightness etc PP if one wants to recover to "flesh tones" even if it is sub optimal otherwise an as-is shot would be clipped badly - please remember all the samples in my experiments on RAW in post #23 above were from the exact same shot - all except the first scene setting one which was the paired JPG of the same shot - all the rest were from the exact same DNG/RAW.

I also liked one of the other ideas that someone showed an example of - use a strobe for fill. If the strobe is only used for fill and not primary lighting, it won't have that obvious "flash photography" look, especially if you move the strobe off camera.

Yes, I get this advice about off-camera flash a lot - it is pretty generic (without meaning any disrespect) - but this is in-situ fill-in that tries to seem like it is not a flash shot. Please note the flash compensation was -2/3 stop and the overall compensation was -1 stop - the flash really is only just filling in, and the direction/diffuseness is not that critical - that is why I used the humble on-board flash which many would automatically pooh-pooh, or advise me to use an off-camera flash - I have a setting on my compact that I use a lot that's "P" for exposing for the ambient conditions (with a slowest shutter speed of 1 sec) and -1 2/3 stop flash compensation - most of the time people cannot tell it has flash-fill and I even had a photographer come up to me to ask how I am managing to shoot without flash!!!

One thing, the color temperature of nearly all stage lights is much lower than typical strobes. Put a CTO gel on the strobe! I've had good results of mixing stage ambient and strobe by putting a CTO gel on my AF-540, although in many cases I probably could use an extra 1/4 CTO, stage lights are often VERY orange even when ungelled.

This may be right for traditional gel'd/filtered tungsten/halogen lights -
but it is not correct for LED lighting - which is what this thread is about -
colored LEDs have very narrow bandwidth - basically monochromatic, as in single color -
even when the light appears to be "white" to our eyes (made up with equal proportions of RGB) BUT the spectrum is discontinuous - has very strong peaks -
so strictly speaking there is no color temperature** and the CRI (Color Rendering Index) is very low -
despite all that, LED stage lighting is becoming more and more prevalent since smaller LED units offering millions of colors are becoming affordable that even small clubs are now using them -
and guess what? most find magenta is flattering and "cool" looking -
but it plays havoc with photography!

An aside ** color temperature and CRI (Color Rendering Index) - there is a popular line of light bulbs from GE called "Reveal" now available as CFL (Compact Fluorescent) one will notice that the GE Reveal CFL gives a color temperature of 2500K and CRI=70
compare this to the GE standard SoftWhite which tries to imitate tungsten - with color temperature of 2700K and CRI=82!
How come the Reveal which is clearly more "white" with more blue in the light gets a lower color temperature than a SoftWhite imitation tungsten?
this is I think due to the discontinuous spectrum of Reveal - which also leads to its lower CRI.

[Marc Sabatella]

Just curious what you find about using blue or green light, then. Your results with PPL more or less match mine with ACDSee Pro - red (incuding orange or magenta lights) require me to peg the WB slider to the left, and might also require a hue shift in the color tool, but I can often recover somethng at least as good as what you show here. With blue or green, I've got plenty of room on the slider to play with with, but the results just never look very convincing.

[UnknownVT]

Just curious what you find about using blue or green light, then. Your results with PPL more or less match mine with ACDSee Pro - red (incuding orange or magenta lights) require me to peg the WB slider to the left, and might also require a hue shift in the color tool, but I can often recover somethng at least as good as what you show here. With blue or green, I've got plenty of room on the slider to play with with, but the results just never look very convincing.

This is a hard one for me to answer.

I tend to prefer to present my photos as I saw it - be it psychedelic or Warhol'ish and all - it's not that I'm a stickler for documentation - it's just that I feel my photos should reflect the gig as I saw it.

So I tend to shoot JPGs and accept what I get - and mostly I am happy, even if I know full well I could have squeezed more optimal quality out of RAW -
but then I am mostly posting photos at 600x400 - that hardly merits anything extra - I could (and have) gotten away with a 2Mp p&s with no manual control - much less a superlative Pentax K-x with its outstanding image quality.

I know how to manipulate things in RAW mostly via ACR or Pentax DCU/SilkyPix - even though recovering the skin tones from magenta light was a big surprise to me - so I do not know everything and definitely acknowledge there are many who are far more experienced than me on RAW.

For shots "as-is" (in either JPG or RAW) there are exceptions when the camera is fooled by the lights - eg: if there is strong tungsten "white" or amber light included - it could upset the AWB to rendering the scene way too blue - my guess is that cameras do not yet understand the very narrow almost monochromatic LED lights (not really their fault - it's our eyes that are deceived) those are probably the only occasions when I wished I shot in RAW - most other times I am quite happy with the JPGs
- and the K-x produces very good JPGs.

So with stage LED lighting we now face quite a lot of problems that were not there when older filtered/gel tungsten/halogen lights were the norm.

For example a blue filtered halogen may look the same blue as a LED - but there is a lot more frequencies in the tungsten light - as Entropy rightly pointed out - I used to use tungsten white balance at most gigs I shot with the K100D - now with LED lighting more and more pervasive I am using AWB - basically to cope with the changes in light color which is almost inevitable due to light persons changing for change sake. Or if anything I'd use daylight balance - neither is ideal.

After all that digression - I agree with you anything that an attempt to recover to real skin tones even via the more flexible RAW is just not convincing - probably due to discontinuous spectrum of the lights (even with filtered tungsten - by definition filtering eliminates some frequencies) this is really aggravated with LED lights where the colors are literally monochromatic - and although they look good to our eyes - plays real havoc with photos -
magenta light made up of red and blue LEDs is a very good case in point.....

...... and I've managed to come full circle!

But I suggest trying out the supplied with camera Pentax Digital Camera Utility/SilkyPix updated to whatever version you are allowed for your camera and see if that makes any difference - like I said I was really surprised and chargrined by using the grey-point for the white balance and recovering almost acceptable skin tone and balance - I really did not think that was possible.
So perhaps with a less extreme lighting maybe Pentax DCU/SilkyPix might do even better?

[Ben_Edict]

Some of those pictures look like exposure for one channel was clipped, so the reccommendation of someone earlier to drop exposure a bit makes sense.

I also liked one of the other ideas that someone showed an example of - use a strobe for fill. If the strobe is only used for fill and not primary lighting, it won't have that obvious "flash photography" look, especially if you move the strobe off camera.

One thing, the color temperature of nearly all stage lights is much lower than typical strobes. Put a CTO gel on the strobe! I've had good results of mixing stage ambient and strobe by putting a CTO gel on my AF-540, although in many cases I probably could use an extra 1/4 CTO, stage lights are often VERY orange even when ungelled.

The low clour temperature is sur true for traditional halogene lighting, but not for LEDs. In the true sense of the definition, these do not even have any colour temperature, as they do not have a contious light spectrum. And "white" LEDs have a much cooler colour than halogene bulbs.

Ben

[falconeye]

In the true sense of the definition, these do not even have any colour temperature, as they do not have a contious light spectrum.

Because this notion of color temperature keeps getting repeated, please let me comment about it.

"In the true sense of the definition", only black bodies heated to a given temperature radiate light with a temperature (which is the black body's temperature due to thermodynamic equilibrium).

Because our dear star's "sun" surface is such a black body (heated to ~6000°C), this notion is very convenient.

Because the spectrum of the black body radiation is defined for all frequencies, one can compute the temperature from the ratio of intensities for any two frequencies. If one is adding a third frequency (color) then one ends up with having 2 temperatures from two adjacent intensity ratios.

Because image processing software does use three colors indeed (RGB), we get a "temperature" and a "hue" measure. Where a "hue" means that the light didn't have a defined temperature in the first place.

If the light source's spectrum is discrete or continous has nothing to do with all of this.


E.g., many light sources with a "known" color temperature have a discrete spectrum, like fluorescent lamps.



Further reading: Color temperature - Wikipedia, the free encyclopedia