DV Info Net

DV Info Net (https://www.dvinfo.net/forum/)
-   Sony XDCAM PMW-F3 CineAlta (https://www.dvinfo.net/forum/sony-xdcam-pmw-f3-cinealta/)
-   -   Getting that S-Log exposure (https://www.dvinfo.net/forum/sony-xdcam-pmw-f3-cinealta/508291-getting-s-log-exposure.html)

Gregory De Tennis June 6th, 2012 12:16 PM

Getting that S-Log exposure
 
Here is another spot shot using S-Log going to a PIX240. Still getting the exposure figured out. The next spot that I am editing right now should be better yet.

For this spot I more or less used the zebra's for exposure. This was before I learned the method of using my zebra's to expose skin at 70% and then backing down one or two stops or by setting my white at 90%-100% and then backing down one or two stops.

The next spot I post will have those setting used. But for now here is this spot.



Richard Crook June 7th, 2012 04:14 PM

Re: Getting that S-Log exposure
 
That looks really nice. I just shot a Harley Davidson spot and used the spot exposure to set grey card (or skintones in some shots) to 38%, with zebras at 100, histogram centered or to the left if highlights deemed it so.. Seemed to work nicely in the grade.

Dennis Hingsberg June 7th, 2012 06:37 PM

Re: Getting that S-Log exposure
 
Zebras are a poor way to determine exposure, a far more accurate in-camera method is using marker function built into F3 and aiming at skin area to see what you have, or grey card as suggested by other poster. For Slog, 30-40 produces nice results from what I've seen. Without Slog I'm keeping skin around 50 or lower.

I actually only use a light meter now to determine and set my exposure to get consistent results.

Doug Jensen June 7th, 2012 07:27 PM

Re: Getting that S-Log exposure
 
I disagree, Zebras are great way to judge exposure when used properly -- particulary on bright white or an 18% gray card because those surfaces are consistent and you can know EXACTLTY where the zebras should appear. But using skin tone to judge exposure is always a bad idea because everyone's skin shade is different. Would you put Obama and Romney at the same exposure? Obviously not, that's why judging anything by skin tone just means you're making an educated guess. But white and gray -- those are conistent and zebras can be every bit as accurate as a waveform monitor when they are used properly.

Dennis Hingsberg June 7th, 2012 11:10 PM

Re: Getting that S-Log exposure
 
Why would you use zebra strips on a gray card or white paper when the F3 has an amazing built-in marker function that will display the actual IRE value of whatever the marker box is hovering over? This is almost as good as a waveform monitor which can be used to show you exactly what values you have throughout a given scene - not to mention FAR more accurate than using zebras to view the overall exposure of a scene.

Also just using zebras can be subjective ie: just how many lines need to appear to indicate you are at a specific IRE? And once you back down "a bit" just how much did you in fact back down, what IRE are you at now? Good luck repeating that from shot to shot or day to day. All that = inconsistency in your shots.

Even if you do manage to use zebras to set your "white" you would still adjust your IRIS or LIGHTING accordingly to the skin level of your subject in the scene, same as you might open up if you are shooting in a dark room with black walls and you want to add detail to the shadows in the shot. Overall "good" exposure takes in account what you are shooting, not just what is white and making sure it is set to "100".

And that is the concept behind exposing for "skin" tone, it simply means with lighter or darker subjects you would bring light levels up or down sublty as required.

Doug Jensen June 8th, 2012 04:48 AM

Re: Getting that S-Log exposure
 
Dennis, you have just made my point because that is my definition of just guessing at the exposure. I prefer to be more scientific than that because I want to nail the exposure perfectly.

Please don't assume that your way of using zebras is the same as my way of using zebras because I would never do what you have described in your last post. After 30 years of using zebras I think I know what I am doing by now.

But if your way works for you, great. There is more than one way to skin a cat. All I am saying is please don't make a misleading statement that "zebras are a poor way to determine exposure", because that is completely false and does a disservice to inexperienced people who may read it. Zebras can be an excellent way of setting exposure if they are used properly -- like any tool.

BTW, if the "marker function" (it's actually called Brightness Display) is so handy, then I wonder why you are only using a light meter now? I have a 758c meter for times when I need to check something, but it is totally useless in a run & gun situation or anytime I have to be moving fast. Zebras, and sometimes the Brightness Display, provide me with much more information than the light meter could ever dream of.

Dennis Hingsberg June 8th, 2012 02:36 PM

Re: Getting that S-Log exposure
 
Doug, my first comment "zebras are poor way to determine exposure" was in relation to Gregory's post and the method he had described and used. You then came along describing a different method (your method) so I believe we are in fact discussing two different approaches using zebras.

My comment to Gregory was not at all to criticize his work because in fact the final composition looks amazing and I feel it looks absolutely fabulous! (Proof that using zebras on skin method can work!) But as soon as you stop down from 70% zebras on skin you are REALLY just adjusting the levels to what you feel "looks good". Again, that is great, but it can also be pretty ambiguous. So my suggestion is why not take a reading with your light meter, or waveform monitor, or built-in brightness function (I called it marker function stolen from HVX cams) to see what level your skin is at so you can repeat it as needed? In Gregory's case say for example he was at 70% then stopped down a bit and the skin level reading ended up being 50%. That is quite scientific, so why not use that as a reference point going forward from shot to shot if possible?

As for your method, I would be very enlightened if you would describe it in more details since you agree that you wouldn't expose Obama and Romney at the same exposure yet you mention using zebras to set your white source, but no where did you mention how you might adjust your levels according to your actual subject matter? I'm sincerely intrigued. Please explain.

I mentioned use of the marker function/brightness display for those people who have absolutely no intention of using a light meter since I know people are not big fans of them.

Doug Jensen June 8th, 2012 08:06 PM

Re: Getting that S-Log exposure
 
Dennis, I need to clarify something I wrote yesterday because when I read it tonight is is confusing the way I worded it.

What I wrote: "Would you put Obama and Romney at the same exposure? Obviously not . . ."

What I meant to say was: "Would you expose Romeny and Obama's skin tones at the same place with zebras or a waveform monitor? Obviously not . . ."

I'm sorry my wording was confusing because I would absolutely expose each of their faces exactly the same. All things being equal, if f/2.8 is the right exposure for one of them, then it would be the right exposure for the other. My exposure is determined by incident light, not the lightness or darkness of the subject. That is why judging expsore by skin tone means you're just guessing.

For example, let's say we set up a headshot of the darkest skinned African we could find. And we both agreed that we had exposed the shot perfectly at f/4. Now the African leaves and we replace him with the whitest albino you could imagine. Would you change your exposure? I certainly would not. The lighting is exactly the same so why would I change my exposure?? One subject is supposed to look darker than the other -- because they are darker than the other. If you try to make every face medium gray and mitigate their natural differences, that is wrong.

Another example. Suppose the albino and the African are standing side-by-side in the frame and we agree again that f/4 is the right exposure for the two-shot. Now one of them walks away and only one is left. Would you change the exposure? Not me. The lighting has not changed.

Another example. Suppose we have a shot of a dark blue box sitting on a table top. Now we replace the dark box with a bright yellow box, but everything else about the shot is the same. Would you change your exposure? I would not. The dark box is supposed to look darker than the yellow box -- because it is. But if I was judging me exposure by looking at the zebras or Brightness Level meter on the box -- then I might be tricked into choosing the wrong exposure. I would set me exposure with a white or gray card, and the lightness or darkness of the box would be irrelevant.

The whole idea of exposing for a gray cards (white cards are the same principle) or using an incident light meter is that your exposure is not dependent on WHAT you are shooting. It is dependent on the actual lighting of the scene.

On the other hand, if you judge exposure by looking at skin tone with zebras, the Brightness Level meter, or even waveform monitor, now you are judging your exposure by an a surface of unknown reflectance and have to guess based on how dark or light you think the subject's skin is.

Using zebras, the Brightness Level meter, or a waveform monitor to measure the light hitting gray card or a white card is the same thing as using an incident light meter. You are measuring the lighting of the scene from a known surface, and not guessing based on how dark or light you think the subject should look.

For example, let's say you think skin tone should be exposed at 38%, 50%, 70% or whatever. That values doesn't matter. What matters is that number, whatever is it, can't possibly be correct for both the albino and the African. At least one of them is going to be exposed wrong.

But if I use an 18% gray card or a white card, I can reliably set me exposure because I know precisely how that surface should be lit. Obviously I might want to adjust the exposure for dramatic or creative reasons, but that is not what we are talking about.

If you really want a great illustration of why zebras can't be used for skin tone, get yourself a DSC Cambelles chart and tell me which of the four girls you are going to use as your standard -- yet they all must have the same exposure because they are on a chart.

If you'd like to come to one of my workshops this summer, I'll be happy to demonstrate my methods of setting exposure in person. :-) The next one is on June 28th at AbleCine New York.

Setting up and Shooting Great-Looking Interviews with LED Lights – New York

Dave Sperling June 8th, 2012 10:09 PM

Re: Getting that S-Log exposure
 
Hi Doug,
I agree with you that zebras are not a magic exposure wand for faces of varying brightness, but I disagree with you on some of your Light Skin / Dark Skin analysis. I often find that the dark skin may need a 'special' light to bring up a little extra detail, or the pale skin may need to be taken down a bit. I find myself adjusting the light level on the subject to get a more pleasing ratio to the background -- I've been doing this since my days in film. When I have extremely light and dark subjects in the same frame, I'm often adding a light, or taking light off with a single or double net.

Doug Jensen June 9th, 2012 05:13 AM

Re: Getting that S-Log exposure
 
Hi Dave,

I don't disagree with you. I might adjust the lighting for a particular subject too (for all kinds of reasons), but I'd try to keep the overall exposure the same. Like I said, there's more than one way to skin a cat.

Alister Chapman June 9th, 2012 09:23 AM

Re: Getting that S-Log exposure
 
One problem is that the zebras don't go down far enough to be of any use with S-log and a grey card. White is less useful as it's in the very heavily compressed part of the S-Log curve so exposure is less accurate and more prone to inconsistencies.

Quote:

Originally Posted by Doug Jensen (Post 1737381)
The whole idea of exposing for a gray cards (white cards are the same principle) or using an incident light meter is that your exposure is not dependent on WHAT you are shooting. It is dependent on the actual lighting of the scene.

For log, this is true as you will be able to adjust the scene for mood etc in post, all you are doing is trying to ensure as much of you scene fits within the dynamic range of the camera, then artistic choices can be made in post. For standard gammas with minimal post, I don't believe this is at all correct because I believe that exposure is equally dependant on what you are shooting as well as the measured light levels. After all you would not want a mid day scene to be exposed with the same brightness range as an evening scene. I would adjust my exposure differently for your dark skin/albino example if I felt it helped the aesthetics of the shot. It has to be remembered that video cameras and monitors have much more limited dynamic ranges than our own visual systems. Say we are in that room with the dark skinned and lighter skinned people, while we observe the room around us at essentially the same exposure because of our large latitude we will easily see both faces against the background. Shoot them with a limited dynamic range camera and it's often beneficial to cheat a bit to avoid the dark face becoming to dark compared to the background and vice-versa.

Doug Jensen June 9th, 2012 11:50 AM

Re: Getting that S-Log exposure
 
Alister, obviously there are creative reasons why one might want to light a scene darker or lighter than normal. And that is why I wrote in my last post: "Obviously I might want to adjust the exposure for dramatic or creative reasons, . . ."
So we are in agreement on that point, but I don't think that is really what we are discussing here.

If you're going to adjust your exposure differently for the albino vs. African, what are you basing that difference on? I think it just comes down to the fact that you're guessing. Maybe you have enough experience to make a good guess, but it's still guessing. I on the other hand, would not expose differently for those two faces, and I would set the exposure based on a known value and not the tone of their skin. That's all I'm saying, and that 30 years of experience doing it that way, and never grading to "fix it in post", shows that it works.

The point of my original post was to take issue with the ridiculous statement that zebras are are a poor way of determining exposure. I know that statement is wrong -- whether you agree with my particular techniques or not. Zebras are excellent tool for judging exposure if someone knows how to use them properly.

Alister Chapman June 9th, 2012 12:36 PM

Re: Getting that S-Log exposure
 
It's not guessing any more than when you drive a car and turn the steering wheel to go around the corner. Most people that have driven a car for some time can judge with incredible accuracy how much they need to turn the steering wheel to drive around a corner that they may have never seen before. You could say they are guessing how much to turn the wheel or you could say they are using their skill and judgement in turning the steering wheel. I don't think we drive cars through guesswork, I think we use a learnt skill. Either way the result is a phenomenally accurate end result considering all the variables involved, something that computers find hard to match.

I say that when someone tweaks exposure to improve the look of a shot that they are using skill and judgement which goes beyond mere guessing. Exposure is not a purely technical exercise where "X" scene needs "Y" exposure. It is a creative, expressive exercise that combines both matter of fact technical skill with creative interpretation. If it was a purely technical exercise then every scene in every movie would be exposed in exactly the same way with "X" exposed at "Y" and that would be incredibly boring. When I adjust my exposure to make an image look nicer, it's not a guess, I know it looks nicer.

While typing this a TV ad came on with an actor in a large number of scenes. In almost every scene the exposure of his face is completely different ranging from deep shadow to brilliantly bright. None of them looked incorrect, they each matched to mood of the scene, there was no way you could have done that using zebras on his face. Zebras do work and can be used for gauging exposure when your working with what I would call a classic average scene. But so much of what we do involves scenes that are not your classic average scene. Then how do you expose? Do you guess or use your judgement and skill?

Dennis Hingsberg June 9th, 2012 01:07 PM

Re: Getting that S-Log exposure
 
In theory I can definitely agree that what is white 100 in a scene to a camera, that anything else placed in its view really is relatively lit the same as you would see it in real life and should not be changed. But, I would still change the lighting (subtly, such as 1/4 stop increments) depending on WHO sat down. It is artistic decision as Alister has put it, and you don't have to - but if you were hired to shoot the BET awards and didn't open up your IRIS in accordance to what your white 100 level was, I guarantee you would be fired for your work. :)

I'm mainly just saying that once you find the lighting or look you like, use your cameras brightness function, a light meter, or waveform monitor to keep the expsoure consistent throughout your shoot. I just believe there are far better tools for doing this than "zebras".

P.S. I do use my zebras to show me what's clipping in a scene though, one last sanity check i do before pressing record.

Alister Chapman June 9th, 2012 01:51 PM

Re: Getting that S-Log exposure
 
1 Attachment(s)
Actually white is an interesting case in point because while the absolute maximum brightness you can show on a screen is 100% (ignoring superwhites for the moment), that does not necessarily meant that you should expose white at 100%. Lets say you have a room with a white wall and a lamp in the middle. Where do you expose the white wall? Certainly not at 100% because the lamp will be invisible against the 100% white wall. In practice, the nominal correct exposure for a white card with standard gammas is 90% to allow for brighter than white objects.

Now if we take that same room and introduce a desk with a sheet of white paper on it, lets imagine that the sheet of paper is registering at 65% on a waveform monitor. In the overall scene it will appear quite grey and dull compared to the brighter white wall. Now if you replace the white walls with black walls what happens? That same piece of paper will appear much brighter because now it is one of the brightest things in the scene. Same exposure, but the paper will appear to have different relative brightnesses. If you were cutting between two shots like this and the paper was your key subject you might find you need to reduce (darken) the exposure of the second darker shot so that the papers relative brightness remains the same. A reflected light meter looking at the scene average would tell you to increase (brighten) the exposure, an incident light meter would tell you to keep the exposure the same, but in order to maintain consistency between shots you would need to stop down a little which goes against everything your meters, zebras or whatever would tell you.

See the attached image. look at the grey squares. What do you see, does the middle grey square appear slightly brighter than the top one? Well it's not it's exactly the same brightness (65%). What about the bottom grey square, it should appear to be roughly the same brightness as the top one, but it isn't, it's considerably darker at only 45% but the different backgrounds alter the way it looks. It's things like this that mean that exposing "x" at "y" won't always work.


All times are GMT -6. The time now is 09:58 AM.

DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network