how do you practically shoot at 5000K?

[yort16]

greetings,

i have a fair amount of experience with the red camera as a first assistant. all of the dp's that i have worked with have balanced the color temperature in camera (even though i know it is applied only to the meta-data) as opposed to filtration. more and more i am seeing that the daylight chip will loose latitude and detail if this is done. most of the shows that i am working on are typically between 5 to 13 million dollar pictures that are classified as independents so fixtures are available. it is ingrained at these tier pictures about night interiors/exteriors as being in a tungsten mode as it has been for many years. hmi's have definite color variations from lamp to lamp and do not have the same characteristics as a tungsten fixture in regards to softness. simple set design also relies heavily on practical tungsten lights on camera for motivation. filtration, such as an 80a is not really practical based on a 2 stop loss. for the cost, the camera looks amazing in a daylight environment. i wish the same for tungsten. any thoughts?

 

[Daniel Browning]

For the cost, the camera looks amazing in a daylight environment. I wish the same for tungsten. Any thoughts?

Tungsten light has four times as much red light as it does blue light. 400%. Quadruple. That's a lot of missing light. Even before it gets to a camera, the light itself is that way.

The camera manufacturer only has one way to make daylight/tungsten equal: reduce sensitivity to red and green relative to blue. Rob Peter to pay Paul.

No one is working on that sort of solution, that I know of. But RED and everyone else *is* working on reducing read noise and improving QE accross all colors. That will improve tungsten (read noise improvement especially), but it will also improve daylight. A rising tide raises all ships.

Personally, I don't think we'll ever get *equal* performance in Tungsten as daylight, but I think that eventually the performance of both will improve so much that we'll ge able to do much more with less light.

By the way, since English is obviously not your first language, I wont say anything (else) about the poor grammar.

 

[donatello b]

you could go with a 80c (-1 stop ..approx +1200-1700k ) filter .... try to keep practicals 2800k and above ..
rate camera ISO around 200 interior ...200-250 exterior ... 320 max (exterior night) ...

or forget the 80c ... add 1/3 - 1/2 blue on tungsten lights ...or use combination of 80c with 1/4 - 1/2 blue on lights ...

do test ... if you are going for a 2k out ..make sure you test taking 4k clips down to 2k and view projected ...

 

[LEON]

you could go with a 80c (-1 stop ..approx +1200-1700k ) filter .... try to keep practicals 2800k and above ..
rate camera ISO around 200 interior ...200-250 exterior ... 320 max (exterior night) ...

When you say ISO around 200 interior you mean this taking account
of the 1 stop loss because of the 80c filter or without ?

 

[jahno]

as a first ac too, i always set the WB to 5000K and explain the DP why. most of them trust me

so we're always shooting @ 5000K.
what i'm interested in is the right way for the post.... the keep saying "he we don't need to fix the WB in redcine cause we'll do a complete color grading in speedgrade or what ever." does this work? or would you advise them to correct the WB in RC?

 

[Joel Kaye]

he we don't need to fix the WB in redcine cause we'll do a complete color grading in speedgrade or what ever." does this work? or would you advise them to correct the WB in RC?

I'd advise testing the two workflows to see if there's any visible difference. Speedgrade normally supports RAW formats but some people have wondered if the SDK gives the same access to the file that RedCine or Scratch has.

But I have to ask if you're going to fix it in RedCine then why not just put it in the camera that way to start? It IS just metadata so nothing happens to the RAW recorded file regardless of what you set the WB at.

I really don't have any clue as to your logic for not putting the WB balance somewhere close to what the DP would like to see when the footage is finished. I've never seen anyone contend that there was less dynamic range shooting at one WB vs. another. Perhaps you're saying the preview monitor shows more crushed blacks and clipped whites when set to 3200 as opposed to 5600?

I've never really noticed anything like that, but I always rely on the histograms to tell me what's really happening.

 

[Nick Gardner]

I think that it has been made very clear, time and time again, that this fear of tungsten is unfounded.

Shoot tungsten, expose correctly, enjoy great images. Let other people panic unnecessarily.

 

[Blair S. Paulsen]

Tungsten tips

Tungsten tips

If you have plenty of footcandles then putting an 80(x) in front of the lens can yield a smoother image, test if you don't believe that.

What if light levels are limited by circumstance? I leave the key lights alone and put 1/4 or 1/2 CTB (blue) gels on the fill lights. The idea is to "feed the blue beast" in the shadow areas where noise issues are most likely to present. In some cases I have been able to match the lighting ratios gaffers and DPs are comfortable with in 35mm negative acquisition with much less shadow noise - that said, I still recommend flatter ratios on set then let the colorist deepen the shadows to taste.

Note: You may still want to crush Zone 1 where the blue channel gain is most problematic but with the right toe curve I have seen much better detail to noise ratios in Zones 2 & 3 with this method. as always, VMMV.

Cheers-

 

[Daniel Browning]

Great advice, Blair.

Note: You may still want to crush Zone 1 where the blue channel gain is most problematic but with the right toe curve I have seen much better detail to noise ratios in Zones 2 & 3 with this method.

Another option is to desaturate on a curve, with full colors in the highlights, slightly desaturated midtones, and very desaturated shadows. It removes attention from the blue channel noise without totally removing detail. Our eye works the same way: the lower the light level, the less saturation we detect, until vision is completely monochromatic in the darkest light.

 

[Blair S. Paulsen]

Back at you Daniel. Letting the chroma tail off in the low end seems like terrific way to quell noise. This could work particularly well when the composition has a large percentage of the image area dark, with only the subject in the fat part of the exposure.
On a related note: I like punching the saturation selectively just as a creative choice.

 

[Alexis Hanawalt]

I think that it has been made very clear, time and time again, that this fear of tungsten is unfounded.

Shoot tungsten, expose correctly, enjoy great images. Let other people panic unnecessarily.

I've had really bad problems with Tungsten when exposed correctly. All that blotchy noise...

 

[Nick Gardner]

If you have noise, you are not exposed correctly.

I have shot under all types of lighting, I shot a feature where 90% of our interiors were lit with good old mole richardson tungsten units, night exteriors with maxi brutes, mini brutes and 10ks, no problems with noise.

If you under expose, and try to pull it up in post, then there is noise.

Tungsten light is not a problem.

Nick

 

[Blair S. Paulsen]

Fear not, respect yes

Fear not, respect yes

OK, I'll buy the "don't fear the tungsten" bit, it has been overblown enough to qualify as classic RedOne FUD [DIGRESSION ALERT: sounds like an interesting thread or magazine article "Classic RedOne FUD vs Mythbusters"] but like many things there is some basis in truth.

For the purposes of focusing this thread on what I perceive as the salient issues I suggest we stipulate the following:
Zone 1 can be crushed into full black no matter what the lighting color temp.
Zone 4 and above are rarely polluted by enough noise to be an issue.
Noise tolerance is dependent on subject matter, what the shots are being inter-cut with and personal taste.

Within these limitations the issue is noise presentation in Zones 2 & 3 (in a rough approximation of the Ansel Adams Zone system, you can substitute IRE units or some other scale if you prefer) both in terms of character and amplitude (S/N ratio).

Due to the physical characteristics of silicon sensor manufacture and other issues beyond the scope of this discussion CMOS sensors like the Mysterium are more sensitive to particular parts of the spectrum. To compensate for this non-linear spectral response gain is added, most notably to the blue channel. It should be noted that the debayering algorithms that Graeme has developed are assumed to be using the full RAW data set to intelligently interpolate blue channel information even from pixel sites that do not have a blue mask. If that is incorrect then perhaps Graeme will see this post and clarify. In any case, the key point is the blue channel gets pumped up and that process can create noise.

Tungsten sources are naturally deficient in blue spectrum output so the artifacts of the gain stage are more pronounced. The relative softness of most edge detail in Zones 2 & 3 makes the edges of the noise "grains" stand out a bit more. Daniel and others have used a variety of strategies to de-emphasize these "grains". The depth and breadth of post-processing options for DNR is a thread (or book) of its own so I will leave that be.

The lower the signal level, the more you can discern the background noise inherent in the physical design of the sensor. Next generation sensor tech like the Monstro may significantly alter this equation - can't wait to see!

All that said, my core advice remains the same:
Feed the blue beast in the shadows.
Use plenty of fill light and reshape the curve in the toe to steepen the contrast ratio in the finished image as desired.
De-saturate the shadows in the grade to keep the blue channel noise down.

YMMV. Cheers-

 

[Blair S. Paulsen]

Not all tungsten is created equal

Not all tungsten is created equal

Nick's post about shooting a ton of tungsten scenes mentioned working with serious pro gear like Mole-Richardson instruments. Old or new these are the real deal. In my experience the difference between the amount of blue energy at a true 3200 Kelvin vs the 2950 or so that often happens on location is dramatic. BTW, working with a proper genny, proper distro, etc to keep the voltage up counts - on Nick's shows I'm guessing that is well handled by G&E. I left my slide rule in my other overalls :innocent: but I believe there is a logarithmic shift going on that ends up asking a lot more of the gain stage.

Evin, and others I have discussed this issue with, feel like moving the color temp up even a few hundred degrees can really help (he likes 4000 when he can get it and will give away up to a full stop with on lens filtration to get there, I will leave a more in depth explanation to him).

It is my contention that when Nick and others say "proper exposure" it takes into account numerous factors that a good DP manages in terms of an overall lighting plan. A simple example would be off key skin tone rendition vs off key surface reads as noise in skin tones can be more problematic.

 

[I Bloom]

filtration, such as an 80a is not really practical based on a 2 stop loss.

You are loosing 2 stops of orange light, but not 2 stops of blue light. The problem is not enough blue light and your light meter doesn't differentiate.

So... if you actually have enough blue light then the 80a is just protecting highlights on the red channel.

In the Red world tungsten lights are just blue deprived. The camera is naturally more sensitive to the blue part of the spectrum (as all digital cameras are) so it does better with low light when it's blue.

I hope this makes sense.

IBloom

 

[Evin Grant]

I've begun using a lot of "Stupid lights" as Starraro cal them. 5000K balanced CFL light bulbs that are cheap and come in 100W, 150W and 300W equivalencies. These lights are wonderful and cheap, those lights plus a Kino coffin, 2x 1200W HMIs, my 200W Joker two litepanels and a 3 ton grip truck pretty much round out my short film lighting package.
All daylight and under a grand a day to rent. Here's a photo of the a stupid light keying a shot...

And here's a 300W CFL keying from inside a Chimera lantern. This is an all 5000K scene keyed with a $30 bulb!

Here's a frame grab...

 

[donatello b]

"When you say ISO around 200 interior you mean this taking account
of the 1 stop loss because of the 80c filter or without "

ISO 200 with the 80c filter on camera .. it's little less then 1 stop loss ( 1 stop would be ISO 160) .. 200 works for me ...

 

[Dustin Cross]

Evin,

Is this 85w CFL what you are using?
http://www.filmtools.com/eiko-85-watt-fluorescent-lamp.html

I have a few of these and they are awesome.

I use a lot of CFLs, Kino Diva 400s, Coollights 150s, and Jokers 400 & 800.

From my experience with Tungsten you need to over expose a little to be safe. Under exposed or even properly exposed gets blue noise.

I have had a couple DPs light a scene with Tungsten and I show them the noise on the monitor, so we relight with daylight and the noise is gone. The noise isn't bad when properly exposed, but it is there. I still use Tungsten light with my Red, but only when I have to.


Mahalo,
Dusty

 

[Douglas Underdahl]

I'm using CFLs, too, inside some custom translucent umbrellas that work like Chimera Pancake lights. Here's some shots from a recent comedy (that's John "Cha Cha" Ciarcia from The Sopranos). Lots of light, don't get hot, low power consumption - CFLs are excellent.

 

[Dustin Cross]

Doug,

Do you have any pics of the custom translucent umbrellas?

Dusty