How can I improve my lighting?

[Bruce Watson]

Another attempt!

I was going to comment that lighting and grading go hand in hand in the digital era. Video, more than film, wants you to get a good exposure in camera that provides the detail and hue/staturation information you need to make the shot work, then grade it to where you want it to end up, in post. And, to recommend that you get and read the Color Correction Handbook by Alexis Van Hurkman, which is one of the best written technical books I've come across. An invaluable resource; it will answer questions that you haven't thought to ask yet.

But.... it looks like you are already headed down the path that Van Hurkman might indicate. Keep going, you're headed in the right direction IMHO.

 

[Brandon J. Garcia]

My only comment at this point would be have you checked this on a 'big screen', at least a 32" HD display for any 'wyrd' artifacts of the contrast changes. That is 'banding' in smooth gradient areas, encoding artifacts, etc.

Another attempt!

Original

Graded

Original

Graded

If you have to grade this much to get the look that you want, you have a problem.

The original image is super flat.
Looking at your other grabs, it looks like your ratios are pretty flat as well and very underexposed.
It's ok to have an underexposed image, but you need to have at least some part of the frame exposed properly. If you want high contrast then use hard light hitting the back of the head and side of the face and expose for that hard-light and you can use as much fill as needed. It doesn't even need to be hard light, could be soft light, but it has to be exposed properly.

It looks like you are not using enough light and sometimes positioning them kind of weird.

What are you using to light?

It's important to remember that cinematography isn't about adding light - it's about adding shadow. Shaping your light.

Take this screen grab from a film I shot.
We hit the back of his head with a 650w Arri fresnel light, while blasting a 1K open face through the window and supplementing that with a fill of a 4x4 bank kino bouncing off the floor, just giving enough fill.

I exposed for the highlights so all other light comes in with more contrast but still exposing the image properly.
The amount of contrast depends on the story and your aesthetic, but you get the idea.

 

[Azmyth]

If you have to grade this much to get the look that you want, you have a problem.

The original image is super flat.
Looking at your other grabs, it looks like your ratios are pretty flat as well and very underexposed.
It's ok to have an underexposed image, but you need to have at least some part of the frame exposed properly. If you want high contrast then use hard light hitting the back of the head and side of the face and expose for that hard-light and you can use as much fill as needed. It doesn't even need to be hard light, could be soft light, but it has to be exposed properly.

It looks like you are not using enough light and sometimes positioning them kind of weird.

What are you using to light?

It's important to remember that cinematography isn't about adding light - it's about adding shadow. Shaping your light.

Take this screen grab from a film I shot.
We hit the back of his head with a 650w Arri fresnel light, while blasting a 1K open face through the window and supplementing that with a fill of a 4x4 bank kino bouncing off the floor, just giving enough fill.

I exposed for the highlights so all other light comes in with more contrast but still exposing the image properly.
The amount of contrast depends on the story and your aesthetic, but you get the idea.

Apparently everyone has a different set of "rules" for how this should or shouldn't be done. One minute I hear "don't try to get the look in camera, shoot flat and do it in post".. and the next minute I hear "Oh, no you want to try to get as close as you can in camera".

I'm exposing with the histogram on the camera. I don't have a light meter or any of that fancy stuff. I think I am just going to light until I'm happy with how it looks, and quit worrying about anything else. Everytime I think I'm doing it "right" someone says I'm not. So either, I'm not.. or there really is no right or "wrong" way to do it. No offense to you or your advice, I just keep getting "conflicting" ideas.

 

[legrevedotcom]

I think this is where one should break in and stop the doctrines and advise you to listen to all advice, sift through what you can use and toss the rest (or just keep it in the back of you head for something else)... There will always be people on either side, all depending on who they worked with, who they read, what teachers they had and what lights they prefer to use.

But keep in mind... what you fix in camera, you don't have to worry about in the edit. Maybe a mixture of the two conflicting sides is the answer for you?

 

[Azmyth]

I think this is where one should break in and stop the doctrines and advise you to listen to all advice, sift through what you can use and toss the rest (or just keep it in the back of you head for something else)... There will always be people on either side, all depending on who they worked with, who they read, what teachers they had and what lights they prefer to use.

But keep in mind... what you fix in camera, you don't have to worry about in the edit. Maybe a mixture of the two conflicting sides is the answer for you?

I came to the conclusion myself that lighting to get good exposure, and darkening it in post was still better than trying to get the exact or close to the exact look in camera, because the image will be degraded so heavily that if I did need to do some more post work on it.. I couldn't.

I've seen tutorial after tutorial, of "day for night" where one turns a daytime exposed shot into a night scene.. this seems to be perfectly acceptable (even to hollywood) since I see it done all the time.. The person who posted the watchmen grabs showed that as well.. those images were heavily altered from their originals.

I really liked how my grabs looked, especially in comparison to what I had when I first started testing. It may not be perfect, but I am thinking it really is starting to become "subjective".

 

[fde101]

Something to bear in mind is that your camera will capture a finite range of intensity values, ranging from a minimum brightness level to a maximum one.

If you expose the camera to light that utilizes the range of the sensor (without blowing out on either end), you will have all of the intermediate values available to maximize your flexibility in post to the point of being able to get dark or light images with a minimum of quality compromise.

However, if you make the image too dark or too light while capturing, you are not making full use of the camera, and you are throwing away information which otherwise could have been used in post.

Both perspectives are correct: try to get the overall look of the image as close as reasonably possible to the final result (colors, diffusion, etc.), but expose it in such a way that you are getting maximal use of the sensor range (without blowing out or crushing desirable shadow detail) rather than allowing the image to be too dark or too light.

You then have much more information to work with when simply dropping or raising the intensity levels in post. This is MUCH MORE IMPORTANT when working with equipment such as the GH2 or many other consumer/prosumer grade equipment, where you only have 8 bits to work with in the first place, but is still true with equipment that offers 12+ bits of color resolution...

 

[Azmyth]

Right I totally agree.

What I did was lit so that I had proper exposure (according to the histogram on the camera). I shoot in smooth with the proper settings turned down 2 notches for a flatter image.

 

[j1clark@ucsd.edu]

Take this screen grab from a film I shot.
We hit the back of his head with a 650w Arri fresnel light, while blasting a 1K open face through the window and supplementing that with a fill of a 4x4 bank kino bouncing off the floor, just giving enough fill.

I exposed for the highlights so all other light comes in with more contrast but still exposing the image properly.
The amount of contrast depends on the story and your aesthetic, but you get the idea.

What camera were you shooting? And while I don't know the details, I'll bet that Azmyth does not have this type, or amount of lighting available.

I also tend to think shooting 'flatter' with the GH-1 is better, and make adjustments in post. Boosting contrast does have certain consequences, so getting the 'ratios' close works best.

This is not a 'fix it in post' situation, but where best to apply a 'tone/mood' effect, given certain limitations of camera and available lighting... either natural or artificial...

 

[j1clark@ucsd.edu]

Apparently everyone has a different set of "rules" for how this should or shouldn't be done. One minute I hear "don't try to get the look in camera, shoot flat and do it in post".. and the next minute I hear "Oh, no you want to try to get as close as you can in camera".

I'm exposing with the histogram on the camera. I don't have a light meter or any of that fancy stuff. I think I am just going to light until I'm happy with how it looks, and quit worrying about anything else. Everytime I think I'm doing it "right" someone says I'm not. So either, I'm not.. or there really is no right or "wrong" way to do it. No offense to you or your advice, I just keep getting "conflicting" ideas.

They are not so much rules... as philosophical points of view... One is to get 'in camera' as nearly what is desired, as possible. The benefit here is 'less post processing', that is quicker delivery. The other philosophy is 'get close', but use post for effects, and have available as much data so options are available.

This all has to be in the context of a camera and it's limitations. The same held true for film... Want a 'night scene', with ISO 64 film... required one or two stops of 'push process', result 'grain'... hence the popularity for shooting 'day for night'... but then there were 'rules', no shots of the sky that would counter the 'night' setting, for example.

These days even with film, the use of Digital Intermediaries has allowed film shooters to shoot less 'in camera' and use more post to produce a desired mood. (Never did get an answer where those 'before and after' shots of "Watchmen"(2009) came from...). But those are pretty clearly shot 'flat' with the expectation of post-processing for not only mood, but CGI enhancements.

The main thing is to learn what your camera will do, and how to get what you want out of it... whether it's 'in camera' or camera+post. And this was the same for film. One would use a specific film stock, and 'learn' what it was capable of and one tended to not change willy nilly, without 'retesting' and 'relearning'...

 

[Kevin Marshall]

Going flatter for the sake of grading is all well and good, but even then, there has to be some semblance of placement and ratios. You can overexpose a scene intended to be darker, or flatten the ratios a bit for exaggerating in post, but there's gotta be something there to work with.

 

[David Jimerson]

Yes. And even when you're lighting a certain way with the intent to grade in post, you should already know what you want before you shoot. A combination of lighting and grading can be very, very effective, but only if you plan it ahead, light it appropriately, and THEN work with it in post. The idea of just lighting flat so you can make these decisions in post is, IMO, a n00b's errand.

 

[j1clark@ucsd.edu]

Going flatter for the sake of grading is all well and good, but even then, there has to be some semblance of placement and ratios. You can overexpose a scene intended to be darker, or flatten the ratios a bit for exaggerating in post, but there's gotta be something there to work with.

This topic has had a couple of threads going, so I don't recall which thread I have posted to... but basically one of my early suggestions was if one wanted a 1:3 lighting ratio, on a 'fair skinned' person, it would be better to start with an exposure that placed the 'light side' at one stop above middle, grey, the the 'dark side' would be 1 stop below middle grey, and if one wanted overall dark, then 'pushing' the light side down to 'middle grey' would maintain the ratio, and with adjusting the curves, one could maintain detail in the shadow while achieving the overall 'dark' look.

In this situation 'flatter' means, for me, not having an absolutely 'white' object, and perhaps having no 'black points' either, not that there is no lighting difference across the face, for this example.

The problem with increasing 'contrast' by 'big amounts' is that this enhances noise, which becomes objectionable, as well as certain 'banding' when the resulting histogram looks more like a comb than a smooth haystack... So, this means small adjustments for effect.

 

[j1clark@ucsd.edu]

Yes. And even when you're lighting a certain way with the intent to grade in post, you should already know what you want before you shoot. A combination of lighting and grading can be very, very effective, but only if you plan it ahead, light it appropriately, and THEN work with it in post. The idea of just lighting flat so you can make these decisions in post is, IMO, a n00b's errand.

I don't think I indicated that there should be no ratios of lightings, but based on what I've found with my test shots of the GH-1, that 'flatter' is better... other may have different opinions. And I should probably be clear, 'flatter' for me is having more exposure in the shadows, while limiting highlights, especially if the desired result is a 'dark/mood' shot. The post processing may include some non-linear curves adjustment to keep the ratio on the face, while getting the shadows 'dark'. And of course... to much manipulation will result in visible artifacts...

The original shots the OP gave had blank black shadows. Their later efforts at least had 'some detail' in the shadows, and could be worked with in terms of maintaining that detail while still achieving the 'dark/mood' shot desire.

 

[Kevin Marshall]

This topic has had a couple of threads going, so I don't recall which thread I have posted to... but basically one of my early suggestions was if one wanted a 1:3 lighting ratio, on a 'fair skinned' person, it would be better to start with an exposure that placed the 'light side' at one stop above middle, grey, the the 'dark side' would be 1 stop below middle grey, and if one wanted overall dark, then 'pushing' the light side down to 'middle grey' would maintain the ratio, and with adjusting the curves, one could maintain detail in the shadow while achieving the overall 'dark' look.

In this situation 'flatter' means, for me, not having an absolutely 'white' object, and perhaps having no 'black points' either, not that there is no lighting difference across the face, for this example.

The problem with increasing 'contrast' by 'big amounts' is that this enhances noise, which becomes objectionable, as well as certain 'banding' when the resulting histogram looks more like a comb than a smooth haystack... So, this means small adjustments for effect.

I agree with that, especially about banding (in the case of dslrs). For what it's worth, my post wasn't directed at you, but at clearing up confusion that OP expressed with hearing differing views, and qualifying them.

 

[David Jimerson]

I don't think I indicated that there should be no ratios of lightings

No, I wasn't referencing you with that, just the general idea which is out there.

but based on what I've found with my test shots of the GH-1, that 'flatter' is better... other may have different opinions. And I should probably be clear, 'flatter' for me is having more exposure in the shadows, while limiting highlights, especially if the desired result is a 'dark/mood' shot. The post processing may include some non-linear curves adjustment to keep the ratio on the face, while getting the shadows 'dark'. And of course... to much manipulation will result in visible artifacts...

That's planning for grading with a look in mind, which is what I mentioned. Shooting with detail in shadows and then lowering the curves in post is time-honored and exactly the kind of thing I had in mind.

 

[Azmyth]

Right and I agree with all of those things, and thats what my "intention" was in my tests. To light in such a way that there was still detail in the shadows, that I could then darken in post to get the desired mood. Not to "light flat" and fix it in post. My goal is the get the "look" and "mood" I want and to maintain as much detail in the shot as is possible.

 

[mail4joeg]

Lots of comments, good. They usually overlight and then bring it down in post. This is to avoid noise. You don't want noise on the faces.

Another thing I noticed was a difference of color temperature on one side vs. the other in your still. It might be better to use a single temperature for all the lights, as the sensors are very sensitive to variations like that. It's tricky. Smaller differences might work better.

 

[bruce southerland]

I really liked how my grabs looked, especially in comparison to what I had when I first started testing. It may not be perfect, but I am thinking it really is starting to become "subjective".[/QUOTE]

I agree. You're on the right track--keep experimenting. If you can, do a test with your camera to find the high & low limits. How dark you can expose &
still hold detail without noise-and how light you can go without clipping. If you know this, it will be easier to get the lighting closer to where you want it before you do the grading.

 

[Brandon J. Garcia]

Right and I agree with all of those things, and thats what my "intention" was in my tests. To light in such a way that there was still detail in the shadows, that I could then darken in post to get the desired mood. Not to "light flat" and fix it in post. My goal is the get the "look" and "mood" I want and to maintain as much detail in the shot as is possible.

Look.
Every DP or gaffer has there own creed, so to speak, when it comes to lighting a shot.

Nothing is absolute and you are doing the right thing by reading what people have to say and taking what you like and leaving out what you don't. There are rules, but the rules are there because they are meant to be broken.

The way I learned, was kind of the old school way, where you try and get the image as close the final look as you can. But sometimes, depending on what is needed and the circumstances, you can't shoot this way. So it's good to be flexible and have a balanced knowledge on how to achieve a look multiple ways. That gives you the ability to adapt to almost any shooting condition.


Trying to to match your shot as much to the final product is the what makes cinematography the most fun for me. Creating an awesome image that when you look at it in your monitor, you KNOW that it looks good. You should light to how you feel, at least in the narrative realm. Believe it or not, the grab that I showed you has no color correction and that is something I am very proud of.

Does that mean lighting the way you did and color correcting it like you did make that WRONG? No.
But what I immediately noticed is the DRASTIC change in the image - it was pretty much nowhere near the original. It totally changes the mood and tone of the image, rather than accentuating it, in my opinion. So when I see that, it says to me that you're are trying to 'save' a bad shot and make it compelling by crushing the blacks so much. That's just what I see, does that apply to everyone? Absolutely not.

Color correction, in my opinion, should take a great image and make it even better, not save it from mediocrity - although sometimes that's necessary. If you want to get better at lighting, you just need to keep doing it and practicing. Maybe theres a way that you can keep the same amount of light on his face, yet cut down the light on the wall and thus increase contrast and make the picture look more dynamic? If you keep your ISO down, then having grain in the shadows shouldn't be an issue with most cameras.

The question you should be asking is not how I can I get better at lighting, but how can I get better at creating shadows?


P.S. An additional tidbit I'd like to add. When you're shooting with lower end cameras, I.E. DSLRs, they compress the color space to 4:2:0 ... which means the dynamic range you have to play with in post is VERY LITTLE. So, relying on color correction for your image in post can be risky.

 

[gonzo_entertainment]

Term "flat" getting thrown around a lot "flat" as in the way the camera captures it for post latitude (cinestyle or other flat profile for example) fine. "flat" as in flat lighting, yucky poo (unless it's stylistically something you want). A lot of the posted shots seem to have nothing but a key light, which, sure enough, is going to look very flat and boring on most cases no matter how you grade it.