Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Makeup showing up with greenish hue

Well said. Indeed not specific to Red.

In a past scenario, the Red camera was blamed when the problem was using unauthorized CF cards.
 
Skintone/make up issues show up on film also.
Not necessarily the same but it's generally a combination of light type (spikey is worse) vs makeup types vs spectral responses of the medium.
 
We had this issue on the first episode of the show I'm working on. It was obvious one some people and non existent on others. The makeup department changed to a different kind of makeup at least for the problematic skin tones. Problem has not reappeared since. I do not know what the specific brands were.

And regarding different tint values at different kelvin settings, that is definitely the case, which is why it is so important that you white balance either in camera or in post before final color correction. Doing so will help minimize these skin tone issues somewhat. I know this is a totally separate discussion but it is mildly related because if you don't get the tint setting correct, you are almost never going to get the skin tone contrast to come out right.

Tim
 
Tim Sutherland said:
We had this issue on the first episode of the show I'm working on. It was obvious one some people and non existent on others. The makeup department changed to a different kind of makeup at least for the problematic skin tones. Problem has not reappeared since. I do not know what the specific brands were.
Click to expand...
I've noticed this with certain CCD pickups as well. X Factor on Fox here in the U.S. has had certain judges with very yellow fleshtones, which is something I've never seen to this degree. I think it's a combination of pickup, exposure, lighting, camera settings, and makeup.

Careful testing can avoid this problem. We can solve it to some degree with color correction (pulling a soft secondary key), but that only works with a finished show, not in a live situation.
 
Ah...

See...

The sensor prefers daylight (Like every CMOS or CCD generally) ...I'm glad you found this. I've been using min 1/8th 1/4 or halfs for anything tungsten with great results.

David



axel ebermann said:
I also do not know why this is not discussed more. Skin color is absolutely significant.
I just got a 80C filter (half blue), makes a big difference. and brings my iso back down
to 400...
Click to expand...
 
the 80C definitely helps. so does using 'red color 2'. this color space
definitely gives a much 'cleaner' look.

the 80C is by the way also very helpful whenever shooting greenscreen under tungsten conditions.
we did some tests and were able to pull much cleaner keys.
 
Deanan said:
Skintone/make up issues show up on film also.
Not necessarily the same but it's generally a combination of light type (spikey is worse) vs makeup types vs spectral responses of the medium.
Click to expand...

hey dean,

indeed but that does not seem to be the problem here. we did test shoots with 35mm stills and on my Red and the problem
was definitely camera / Cmos related.
 
we did some extensive test for a feature a year ago, and our conclusion was that makeup plays the key role.
We had a great make up artist, who tried out different kinds of makeup products and the difference was huge !
Unfortunately I can´t post any pictures but believe me it was an eye-opener.

I´m not an makeup artist, but i´ll try to explain the background for that issue.
The artist told me that there are basically two types of make up one is kinda "fat" based the other one is silicone based,
the "fat" based is mainly in use for theater and stage since it "adds" to the face and is better for large expressions.
The silicone based is more "suited" towards movie and TV, since it does´t add so much but just gently smoothes out skin.
Problem with some of the silicone based brands is that there are green pigments in it, since green is contrary color to red, and therefore is used to pale down reddish skin.
We tested using different cameras (alexa, Red MX, 5D, SI2K, and generally one can say that CMOS Cameras with a Bayer pattern are more sensitive to this green pigments in the makeup than film and CCD Cameras (F23 in this case).

We had the best experience with the two Brands: BEN NYE as well as NUMERIC PROOF PARIS

This was by no means a scientific test, but served our purpose very well, and showed us how much makeup takes a role in getting natural skin tones.


maybe that can help somehow.
Manuel
 
axel ebermann said:
hey dean,

indeed but that does not seem to be the problem here. we did test shoots with 35mm stills and on my Red and the problem
was definitely camera / Cmos related.
Click to expand...

I disagree. I've seen it on film, dslrs, ccd, alexa and sony cameras.
One the the premiere colorists also mentioned he sees it on film also.

The cause comes from metamerism between the makeup pigments and skin.
When the makeup combined with a light source will have different spectral
reflectances that are then are received by the capture medium's (film, ccd, cmos, eye)
spectral response.

The greater metameric mismatches, the more the color of the skin vs the makeup start to shift away from each other.

Some pigments are very prone to this and others vary alot depending on the light source.
 
Deanan is correct - it's a metamer, in that it's a complex interaction between the spectral characteristics of the light, the pigment and the camera's spectral sensitivities. And yes, I've seen such issues on numerous camera systems over the years.

Graeme
 
As a stills beauty photographer (shot for Pantene,L'Oreal,Max Factor etc) I traditionally with continuous light sources have used flicker free HMI and meticulously colour temped and filtered each light individually. Any light that showed a green spike of more than +9 (Minolta meter) without filtration was sent back. In fact we used to ask that the lamps were checked before they went out to us. On film (so long as we batch tested and took this approach) we never had a problem - until digital. The initial cameras that came out were shocking and have slowly got better. However in the last year or so where I have been shooting, less on my Phase back and more on the Canon, budgets tightening, meaning more use of tungsten the problem has come back. I think the problem is a spectral/sensor compatibility issue the further you move from daylight with artificial light. I think it's also software related eg a Canon RAW file in Capture showing green tints has much less in Aperture. I have found that if I 1/2 blue all my tungsten lights as a starting point it is less likely to happen. What's weird about it is shooting beauty on a beach at sunset I don't seem to have a problem so that would suggest that under very warm daylight it's maybe not as simple as staying away from very low colour temps. It's certainly not something that just effects Red.
 
It looks like to me that the green is also in the back ground and is wrapping around the talent from behind. the front on skin tones look fairly ok,i don't think this has anything to do with make,i think the make up is pretty good really. either way it is a very quick fix in redcine.
Regards
warwick
 
Agree that this is an Alexa and Sony issue too for sure. Hope you RED geniuses can one day be the first to make a sensor that responds as closely as possible to the human eye - heck, or better :)

The other issue is that the human eye and brain's color response is not just a simple R, G, B bell curve. It's really sensitive in some areas, and inaccurate in others.

Reproducing that on a foreign sensor technology seems to be really tricky - although film managed it somehow... lots of things were not accurate about film, but some of the subtle stuff it managed to get right.

People who think it can be fixed entirely in post ("because it's RAW") don't understand the problem:
- RAW sensor data doesn't mean you captured every spectral wavelength emanated for every pixel.
- RAW sensor data means you captured how much a sensor responds to three wavelength bands (approximating red, green and blue), created by color filters.

Maybe those color filters' wavelength bands aren't quite right (they overlap too much or too little, they let other wavelengths in that they shouldn't - eg IR contamination, etc).

If they aren't right, you can compensate in post, but if the info isn't there - or is ambiguous - you can't restore it in post. Unless the info was there originally and the R3D decode is somehow blending the channels and losing the info? Doubt it though - the RED folks are smart.

OK - maybe you could get a little closer to fixing it in post if you could tell the software "this bit is skin with makeup, lit by a LED light", "this bit is a skin-colored pot in sunlight" and "this bit is skin without makeup lit by an HMI" and then it could compensate, knowing the spectral response for each material & light combo vs the RED's sensor characteristics vs the human eye. Which is kinda what we do now in the grade with secondaries - just by hand.

But personally I would happily trade off EVERYTHING - have lower resolution, lower latitude, higher noise - for a sensor which (combined with clever processing) can capture every nuance the human eye sees - and either doesn't see or can later filter out the stuff the human eye doesn't see either (presumably makeup's IR response and / or metamerism).

I don't know if the answer is to have more color filters (instead of Bayer, have "far red", "near red", green, blue)? More pure color filters? Or it's something that will naturally get better as CMOS technology advances? And how the heck did film get it looking so good?

Bruce Allen
www.boacinema.com
 
We just finished shooting a tv commercial that was shot on 2 epics. The scene was 10 people sitting on a round rotating turntable acting as a call centre. The whole set was lit with 5 space lights. One space light was in the middle of the set and the other 4 were evenly spaced around the set with black skirts on them.

Half of the people had a green sickly tinge to them and half didn't, we changed the people of 5 times so there was a total of 50 people. We had 3 make up artists. Another strange thing was we were also getting a magenta tinge on some peoples necks.

Space lights are tungsten soft lights for those who don't know what they are.


We did some testing afterwards and found that certain brands of makeup under soft tungsten light go green. But regarding the magenta tinge on peoples necks is still a mystery as all of the people where wearing white polo shirts and the set was white, the had white apple computers sitting in front of them, and we had a few floppies for negative fill, so there was nothing red anywhere near anyone!

We fixed everything in Davinci but it was a bit of a struggle as the set was rotating, thank god for Davinci's tracker or we would have been F^%Ked.
 
Deanan,

How can you disagree with a test result that you never saw ? That strikes me as a little strange....


I did not say it only occurs on the Red One. I said Cmos chip cameras. DSLRS, the Alexa and most modern cameras
are Cmos based.
I have not owned / shot on a CCD camera in a while. But I shot tons of footage on CCD chip cameras and cannot remember this being an issue.

My aim is not to bash any camera system - but to have a good understanding how all the factors
as in light/camera/make up interact and how to avoid this very unpleasant effect.

And as I am a proud Red owner, this camera is on the top of my list when it comes to tests.
And looking at this thread it is an issue that deserves some attention.


Deanan said:
I disagree. I've seen it on film, dslrs, ccd, alexa and sony cameras.
One the the premiere colorists also mentioned he sees it on film also.

The cause comes from metamerism between the makeup pigments and skin.
When the makeup combined with a light source will have different spectral
reflectances that are then are received by the capture medium's (film, ccd, cmos, eye)
spectral response.

The greater metameric mismatches, the more the color of the skin vs the makeup start to shift away from each other.

Some pigments are very prone to this and others vary alot depending on the light source.
Click to expand...
 
awesome information !!! thanks for sharing.


Manuel Wenger said:
we did some extensive test for a feature a year ago, and our conclusion was that makeup plays the key role.
We had a great make up artist, who tried out different kinds of makeup products and the difference was huge !
Unfortunately I can´t post any pictures but believe me it was an eye-opener.

I´m not an makeup artist, but i´ll try to explain the background for that issue.
The artist told me that there are basically two types of make up one is kinda "fat" based the other one is silicone based,
the "fat" based is mainly in use for theater and stage since it "adds" to the face and is better for large expressions.
The silicone based is more "suited" towards movie and TV, since it does´t add so much but just gently smoothes out skin.
Problem with some of the silicone based brands is that there are green pigments in it, since green is contrary color to red, and therefore is used to pale down reddish skin.
We tested using different cameras (alexa, Red MX, 5D, SI2K, and generally one can say that CMOS Cameras with a Bayer pattern are more sensitive to this green pigments in the makeup than film and CCD Cameras (F23 in this case).

We had the best experience with the two Brands: BEN NYE as well as NUMERIC PROOF PARIS

This was by no means a scientific test, but served our purpose very well, and showed us how much makeup takes a role in getting natural skin tones.


maybe that can help somehow.
Manuel
Click to expand...
 
Might be a different problem, but in the "hobbit" blogs I noticed the make up artist commented on a issues like this, and she said that she put a bit of red into the base (she had the actor hold up his hand which didn't have makeup and it looked pale). Also the makeup artist I talked to, showed me some cool new transparent bases that de-gloss really nice(if it's just a simple touch up).
 
axel ebermann said:
Deanan,

How can you disagree with a test result that you never saw ? That strikes me as a little strange....


I did not say it only occurs on the Red One. I said Cmos chip cameras. DSLRS, the Alexa and most modern cameras
are Cmos based.
I have not owned / shot on a CCD camera in a while. But I shot tons of footage on CCD chip cameras and cannot remember this being an issue.
Click to expand...

I'm not sure what test you're talking about.
I'm referring to all the tests I've been involved with.

I've worked extensively with both ccd and cmos on the camera development side.
The silicon technology is not relevant with respect to the spectral sensitivity and metamerism.
Both are similar in sensitivity.

The dyes on the sensor or through the filters in a 3 chip along with the visible pass
filter make up the bulk of the system response.
 
fair enough. thanks for the insight.

Deanan said:
I'm not sure what test you're talking about.
I'm referring to all the tests I've been involved with.

I've worked extensively with both ccd and cmos on the camera development side.
The silicon technology is not relevant with respect to the spectral sensitivity and metamerism.
Both are similar in sensitivity.

The dyes on the sensor or through the filters in a 3 chip along with the visible pass
filter make up the bulk of the system response.
Click to expand...
 
You must log in or register to reply here.
Back
Top