THE SWEET SPOT IN THE NEGATIVE

Our four-part series dissecting the RED raw image is presented in order of, what I consider, importance.  I believe the second most important ( and sometimes misunderstood ) area is to not only exposing the image correctly, but hitting the sweet spot in the negative.

Again, I'm not going to tell a DP how to light, how to expose, or how many stops over or under they should be.  I'm only going to offer some thoughts on what has helped us create good looking images from the RED.

There is a clear difference between “what the camera is capable of” and “hitting the sweet spot”.  And you don't always want to hit the sweet spot.  Sometimes you're going for a very specific look and you want to intentionally overexpose the negative, or whatever else.  But most of the time, you are trying to create a well exposed, organic, film-like image with little grain and not too many blown out highlights.

Below are some grabs from a film shot by Ben Kufrin.  We did tests on this film before they shot and helped show them the ideal range to expose within.  I was very impressed with the results, its spot on.  I use these grabs as an example of a kind of negative I consider a perfect RED exposure.  I've posted the corresponding waveform readings and histogram of the raw negative.

RAW

WAVEFORM OF RAW

GRADED

RAW

WAVEFORM OF RAW

GRADED

RAW

WAVEFORM OF RAW

GRADED

These represent what I consider the sweet spot of the negative.  Notice anything?  The RAW's seem a little dark, don't they?  Thats a good thing.

If you're trying to achieve the most organic, rich image with the RED, its important to watch your high end very carefully.  Notice how nothing is either clipping or crushing in the RAW.  There are no blown out highlights or crushed blacks.  That means that we are going to be able to preserve detail both in the high end and low ( like film ) and most of the information is clumped in the lower middle.  That's where the skin tones live.  So we're also going to get a nice amount of tone and detail in the skin.

if the exposures were any hotter overall, we'd be clipping in the highlights and the skin would develop a plastic-looking loss of detail.  If we were any darker, there would be bad noise in the shadows.

This is a sample ( not shot by B. Kufrin ) that shows what can happen when you are underexposed in RED :

GRADED

WAVEFORM OF RAW

Notice not only the noise, but vertical compression lines ( or something, hard to know what is creating that but it appears without fail in all underexposed footage ).  That's very hard to get rid of, if not impossible.

Long story short, watch your highlights, watch your low end, and try to get a nice gathering of information in the middle of the exposure.  When we pull that RAW file apart in the DI, it will be the difference between possibly being able to fool the audience into thinking its film, and the audience knowing its digital.

NOTE : THIS SERIES DEALS WITH THE PRE-MX RED SENSOR.  EVERYTHING HAS CHANGED ( FOR THE BETTER WITH THE MX SENSOR AND MUCH OF THIS SERIES DOES NOT APPLY TO THE NEW COLOR SCIENCE )

I am never going to tell a DP how to shoot.  Its not my place.  But I get a lot of DP's asking me why their Red images look rust-colored.  Or clipped.  Or like they are made out of jello.  So lets address some of those questions.

This deals with what I consider to be the areas of most concern when it comes to creating great looking images with the red.  Although we will link to more in-depth ( and more technical ) discussions, this will not be an all out tech-fest.  This is more of an overview, more theory.  I'm trying to keep it as understandable as possible, especially if you're not a DP or an engineer.  Lets start with COLOR BALANCE.

COLOR BALANCE

You might be surprised to see that my #1 most important aspect of the red image is the camera's native color balance.  But I see more confusion about this than anything else ( and a lot of rust-colored material ).

As you may or may not know, the Red's sensor is daylight balanced ( at around 5000K ).  Regardless of how you manipulate the METADATA, you cannot change the inherit nature of the chip.  And since RED doesn't do any white balancing in-camera, your RAW NEGATIVE will show you exactly what kind of light was hitting your subject.  This means that if you hit it with an HMI or sunlight or anything that can produce light around 5600K, you get a nicely balanced negative :

RAW ( LIT BY HMI )

GRADED

As you can see in the RAW and in the graded finals, the negative isn't skewed heavily to one color cast and is somewhat neutral to begin with.  This is under DAYLIGHT balanced light.

But when you hit it with tungsten light ( pretty much every other kind of light ), you're depriving the blue channel of information.  All thats left is a negative that is heavily weighted towards red / orange.  And its not a pleasing looking "straw" type warm, its more of a rust looking warm :

RAW ( LIT UNDER TUNGSTEN )

RAW ( LIT UNDER TUNGSTEN )

Also, if you're hitting the subject with tungsten light and the scene is exposed dark intentionally, or unintentionally, you will probably see more noise in the darkest areas :

RAW ( LIT UNDER TUNGSTEN )

It is still possible to dial blue into these images in the DI, but the result is less organic and in my opinion, always looks somehow more digital, more manipulated :

GRADED

Compare that to a negative that has a good balanced amount of information across the three color channels :

RAW ( EXTERIOR DAYLIGHT )

GRADED

We find it much easier to create rich, organic images when we receive RAW files that are lit with the understanding that the sensor really performs best under light that is around 5600K.

Now, I realize this is not always practical onset.  You're rushing, you've only got tungsten lights, no time to gel, no time to filter in front of the lens ( or you can't afford to lose the stops by filtering ), we've all been there.  Its good to know, though, what you're giving the camera in those situations, and what you're eventually going to have to deal with in the DI.