Underexpose in camera or in post?

[Tyler Clark]

What is a better practice when looking for an underexposed look? (Thinking fincher, social network, dragon tattoo, etc.)

 

I work with the c300 and fs7 alot using Clog and Slog3.

 

Underexposing in camera on the C300 results in alot of muddiness and on the fs7 alot of noise. The 5D on the other hand seems to handle it well a stop or two under and Ive seen tests with the RED EPIC that 2 -3 stops under works quite nicely.

 

is it on a per camera basis?

 

 

[David Mullen ASC]

Didn't we already have this discussion or was this thread on RedUser already?

 

You partly answered your own question, it depends on the camera and recording codec. It also depends on the amount of bright highlights and darkest shadows in your image because while just exposing normally and bringing down the image in post is often safer, and obviously will give you less noise problems, your clip point will be different than if you underexposed the image.

 

Also, if your scene contains objects that emit low levels of light naturally then with at a lower artificial lighting level, these self-illuminating objects will expose brighter in comparison, so the look will be different compared to working at a higher lighting level and bringing down the image.

 

Keep in mind that Fincher has a lot of high-end noise reduction processing at his disposal; knowing you will have that tool in your back pocket might affect your exposure decision-making process.

 

I think the general answer is just be conservative with your underexposure, don't avoid it completely but don't overdo it either because you can always make it even darker in post.

[Robin R Probyn]

C300 so called LOG isnt a "real" log gamma.. its more like hyper gamma 7/8 in the Sony Fs7/F5/55.. they do this because its only 8 bit..

The Fs7 is 10 bit in XAVC..and Slog3 is almost identical to Arri Log C.. but no NR in camera.. will be with the next V3 for Fs7.. if it ever arrives :)

 

Common wisdom for Slog seems to be never underexpose, and actually better to over expose a stop or 2.. and correct in post.. also if your scene is pretty dark and doesnt have a large DR in the first place.. shooting Log you are throwing away alot of data.. and getting a very small amount in the shadows.. better to shoot with one of the hyper gamma,s ..

 

Good article about this here

 

http://www.xdcam-user.com/2015/12/deeper-understanding-of-log-gamma-experiments-with-a-waveform-display/

Edited December 15, 2015 by Robin R Probyn

[Tyler Clark]

Didn't we already have this discussion or was this thread on RedUser already?

 

You partly answered your own question, it depends on the camera and recording codec. It also depends on the amount of bright highlights and darkest shadows in your image because while just exposing normally and bringing down the image in post is often safer, and obviously will give you less noise problems, your clip point will be different than if you underexposed the image.

 

Also, if your scene contains objects that emit low levels of light naturally then with at a lower artificial lighting level, these self-illuminating objects will expose brighter in comparison, so the look will be different compared to working at a higher lighting level and bringing down the image.

 

Keep in mind that Fincher has a lot of high-end noise reduction processing at his disposal; knowing you will have that tool in your back pocket might affect your exposure decision-making process.

 

I think the general answer is just be conservative with your underexposure, don't avoid it completely but don't overdo it either because you can always make it even darker in post.

 

 

C300 so called LOG isnt a "real" log gamma.. its more like hyper gamma 7/8 in the Sony Fs7/F5/55.. they do this because its only 8 bit..

The Fs7 is 10 bit in XAVC..and Slog3 is almost identical to Arri Log C.. but no NR in camera.. will be with the next V3 for Fs7.. if it ever arrives :)

 

Common wisdom for Slog seems to be never underexpose, and actually better to over expose a stop or 2.. and correct in post.. also if your scene is pretty dark and doesnt have a large DR in the first place.. shooting Log you are throwing away alot of data.. and getting a very small amount in the shadows.. better to shoot with one of the hyper gamma,s ..

 

Good article about this here

 

http://www.xdcam-user.com/2015/12/deeper-understanding-of-log-gamma-experiments-with-a-waveform-display/

 

Dont believe that was me Dave, im not on REDuser (unfortunately considering I never have budgets for RED). I do appreciate the advice and tips everyone!

[Mark Kenfield]

I'd recommend making your exposure decisions in light of the specific conditions of post-production you'll be working with on a shoot. If you're not going to involved in the grade - I'd strongly advise you to expose to precisely the levels you want in camera, as this gives them less room to muck things up and push the image in post (I've been burned by that enough times already).

 

When you do know that you'll have some control in post, I think it's well worth considering a more restrained approach. For example, limiting your underexposure on faces to say two stops under - a level which is dark, but still leaves plenty of detail and information to push things a little either way in the grade.

 

More and more these days, I prefer to light consistently to my Monitoring LUT, this just puts everything close to where you want it across the entire project. And keeping things simple seems to be the best possible way to achieve solid results in the oftentimes disjointed post processes pictures are put through these days.

Edited December 15, 2015 by Mark Kenfield

[Adam Frisch FSF]

I always do it in camera, for many reasons.

 

1. Everything looks better underexposed.

2. Clients get used to the dark look and will then later grade it that way when you don't have control.

3. By underexposing, the clients can only lift it so much in post before it falls apart. This protects my vision.

[Miguel Angel]

Hi there!

After reading Adam's answer I was wondering if any of you use a waveform to expose the image or just false colour and a light-meter and if you expose through a waveform, would you expose using the LUT applied (REC709, Commercial, REDCOLOUR3, 4, etc) or using ArriRaw or RedLog?

 

There is also the question of where would I put the black, skin tones and highlight levels while exposing through an Astro in the Lut / Raw?

 

Maybe something like:

 

- Black levels: 5

- Skin tones as in a normal exposure for them: 35

- Highlights: 70

 

And if you have a super dark scene, would you use a waveform too?

 

Thank you very much!

[Stuart Brereton]

I rarely use a light meter any more, and I've never gotten the hang of false color, so I rely on a properly calibrated monitor. I almost always view with a LUT applied, so that I can see what the picture looks under the least forgiving circumstances. If it's all good in REC 709, then I know that there's more info to work with in Log, so I have room to manoeuvre if needed. I'll check Log if I'm worried about highlights, and I check the waveform if I'm worried about shadows.

[Robin R Probyn]

I do the same as Stuart.. use REC 709 800% on a Sony.. MLUT.. also never liked the false colour thing..

 

If your using a Sony F5/55 there is a not very well known setting in options only,thats very useful.. its called Hi/Low function.. set it to one of the assign switches and you can see very quickly where your high lights and low levels are with each push.. even if you have a LUT in the VF.. very handy for a very quick check of what your Slog is seeing..

[Miguel Angel]

Thank you!

 

I am thinking about buying a 9" Flanders Scientific monitor because I love using false colour for setting up the exposure and the lightmeter for ratios but I have seen a lot of people as of lately exposing with a waveform and I was curious about what you guys were doing

 

Thank you

 

Have a lovely day!

[Jesse Frank]

I think it was me who posted about the RED user. I was getting googly eyed about the RED raven, but still out of my budget.

[Aapo Lettinen]

I mainly light by eye even if having a decent monitor with correct LUT, then checking with waveform or histogram that the exposure is correct. I usually have 95% zebras on at all times so it is easy to see if something is close to clipping.

 

never been a fan of false colour but it can be handy at times. I am used to incorrectly calibrated monitors and general unreliability of digital systems so I mainly treat them just like film: lighting by eye and doing minimal checks with meters, then adjusting it in grading to match the look seen by eye on set.

monitors are mainly for framing etc. and not colour tools for me. they are, however, handy if you have lots of diffusion filtration etc. and they are useful for directors because they are not used to "imagining the look over the image"