How to ETTR and get Thick Negative?

[Yousuf Abbasi]

I have often read on this forum that it is best to Expose to the Right (ETTR), and have a "thick negative." I think I understand the concept but could someone please clarify what is exactly the way to do this?

When I expose, I see my RAW histogram is generally a thick mountain concentrated to the LEFT SIDE, and basically nothing on the RIGHT SIDE of the histogram. Is this the "thick negative?"

I'd rather underexpose than overexpose, because I don't really like to deal with very bright highlights in post. Therefore I generally stay within the greys & blues, maybe pinks, in False Color mode.

Thoughts are appreciated

 

[Mike N]

I think it would help you if you also read this

http://reduser.net/forum/showthread.php?t=38946

and this

http://reduser.net/forum/showthread.php?t=39008

and maybe even this

http://reduser.net/forum/showthread.php?t=38224

The whole idea of this ETTR thingy is to get noisefree clean images. Following this path some people set their light meters to lower than 320 ISO numbers when shooting in tungsten light or low light conditions.

Having a histrogram concentrated mostly to the left is not a good thing with this camera in my opinion especially if that particular shot wasn't a night exterior with a lot of black in it.
A "thick negative" would mean that you have the histogram spreaded also to the right part.

The idea is to get as much of the image "covered with light" especially the shadowy areas and then crush the blacks a bit in post so you get an image that is as much noisefree as possible.

Also try to avoid to much purple in false color it usually means noise.

 

[Andrew Wilding]

The ETTR thing is more or less exposing higher (without clipping anything important) than the look you are going after, than bringing the signal down in post for as clean of an image as possible. This method does indeed produced the most noise free images, but at the cost of either some clipping or lighting within a smaller range of dr. Say that normally you have light values from 0 to 100 in a shot. Say a lamp at the top of the range, near 100, a face two stops below that, another face a stop below that, a wooden desk a stop below that, and some dark areas near to the 0 side o the scale. With ETTR, you say, screw it, the last stop or two down near 0 are too noisy for my taste. You instead decide to work within a range of 20-100, knowing that you will pull it down and stretch the histogram out in post. You still want to take advantage of the dynamic range though, so now the lamp is near one hundred, a face is a little less than two stops under that, the other face is a little less than a stop beneath that, etc etc. This is obviously a smaller range to work within, with smaller tolerances involved. An easier method is probably just to light with a little more fill than you would normally like and crush the blacks later. Keep in Mind please that I am no expert and this is just the way that I choose to look at the problem - I might be a total nutjob..

 

[Andrew Wilding]

some people set their light meters to lower than 320 ISO numbers when shooting in tungsten light or low light conditions.

But than you are sacrificing the top end of your dynamic range, no?

 

[Jeff Kilgroe]

ETTR is good, in moderation. One pitfall I see a lot of people running themselves into is trying to expose too much to the right. Only because ETTR is often preached heavily around here. They are essentially sacrificing dynamic range by doing so. It is much easier to dig detail out of shadows with the RED One than to try and recover from the high end. Noise isn't always a bad thing, it's about how much noise is acceptable.

 

[Steve Sherrick]

I've gone away from generalizing when it comes to exposing the Mysterium sensor and would rather say something to the effect of, light to the stock. In other words, do what you would do if you were shooting film stocks. You would test them, learn the dynamic range, learn the grain properties, learn the color corrections needed and what can be done in processing. Same strategy for RED. Learn it under a variety of conditions, keep track of your findings, and eventually know it well enough to know when you want to make certain decisions that will give you the image you are looking for. Simply saying ETTR doesn't say enough. Knowing what happens during the post process can help DPs immensely. RED has limitations. Dynamic range can be one of them, but it's not as simple as saying it has limited dynamic range.

Where I'm going with this is to say experience will trump generalizations.

 

[Meryem Ersoz]

When I expose, I see my RAW histogram is generally a thick mountain concentrated to the LEFT SIDE, and basically nothing on the RIGHT SIDE of the histogram. Is this the "thick negative?"

I guess the most important thing is, are you happy with your footage? Are your post people happy working with it?

The thick negative just means that you have the maximum amount of data hitting the sensor. As long as you're not clipping, you will be able to pull more information out of this data, on the high end or the low end.

On the post end, I would not be so happy with a thick mountain on the left side and nothing on the right side, because you would not be giving me much to work with in the way of highlights. But if you, on the DP or shooter side, are trying to create a specific look and are happy with this lighting environment, then, well, that's what you want.

The thing about RAW is that your post person can always crush blacks or adjust the white point, as long as you have exposed properly, and the information is there. But if don't give them the information that lets them have that flexibility, then you're defeating the advantages of shooting RAW and aiming instead for a specific image out of the camera.

 

[Andrew Wilding]

I've gone away from generalizing when it comes to exposing the Mysterium sensor and would rather say something to the effect of, light to the stock. In other words, do what you would do if you were shooting film stocks. You would test them, learn the dynamic range, learn the grain properties, learn the color corrections needed and what can be done in processing. Same strategy for RED. Learn it under a variety of conditions, keep track of your findings, and eventually know it well enough to know when you want to make certain decisions that will give you the image you are looking for. Simply saying ETTR doesn't say enough. Knowing what happens during the post process can help DPs immensely. RED has limitations. Dynamic range can be one of them, but it's not as simple as saying it has limited dynamic range.

Where I'm going with this is to say experience will trump generalizations.

 

[David Mullen ASC]

In some ways, it's very basic (and film follows the same rules):

You have "x" dynamic range of the sensor or film stock, and you have "y" dynamic range of the subject, usually wider than "x". You have to decide how much of "y" you can fit into "x", and what of "y" you can ignore or lose.

So one theory would be to split the recorded range evenly between highlight and shadows. Another theory would be to decide which end was more important in preserving detail in.

The other factor is noise versus clipping. A "strong" signal / "thick" negative has less noise, more shadow detail, but a higher chance of clipping bright highlights. A "weak" signal / "thin" negative has more noise, less shadow detail, but more highlight detail.

So those are the basic factors that go into making a decision on exposure, other than artistic intent and practical considerations (like shooting a documentary in available light, no matter how low.)

The problem comes from the fact that we aren't shooting still images that stand alone, we are often shooting a sequence of shots that have to intercut, and some angles may have a lot of contrast and others may not.

So this is a complicating factor, because if we vary every exposure, every shot, in some attempt to split the difference between highlights and shadows, we may end up with variations in noise shot-to-shot depending on where the key subject falls in that range, and how it ends up being exposed. Because ultimately we make color-correction decisions primarily on the key subject, and then secondarily on the deep shadows and brightest highlights.

 

[Sam Taylor]

Thank you David! "variations in noise shot-to-shot" I think that is really important.