Was I tricked by something when I used this spot meter? - Inconsistent measurements

[Vitor Delduque]

Hey,

Yesterday I was shoting a scene for a comercial and my L858 got me pretty good. I took some measurements with the spot meter from where the camera was (we'll 'rebuild everything later) and I got a underexposed footage (something around 1 stop) following the f3.5  sugested value.

OBS:

- I had some light hitting the spot's lens, but not in frame;

-I tought about recalibrating the meter, but I tried to measure different scenarios around the location and the L858 gave me 3 correct values out of 5. I used waveform to check;

-Last time I used my Sekonic it was perfect.. about a week ago;

-I'm new using lightmeters, so I hope it's just some problem with my interpretation of the situation, lol.

Any suggestion?

Thanks!

[Satsuki Murashige]

There are a lot of variables that you have to account for when using light meters on digital cameras.

1. ISO calibration can vary widely on many cameras. It will also change depending on the camera picture profile or gamma settings. You should shoot tests with an 18% grey card, white card, and a waveform to see where middle grey actually lands with your chosen ISO vs what the meter says.

For example, if you lit your grey card to f/4 and set T4 on the lens - middle grey is supposed to be at 41% (Sony S-Log 3) on the waveform and it’s reading at 30%, then you know the camera’s actual ‘ISO’ is lower. Open the lens until it reads 41%, and take note of how many stops you had to open up. This will be your calibration offset value for that camera and lens combination. If it’s a 2/3 stop difference, then your 800 ISO camera is actually 500 ISO, etc. 

Generally, cameras (and also film stocks) are not as sensitive as they claim to be on the box. ARRI Alexa is spot on, otherwise test and make sure. 

2. What lens? If it’s not a T-stop calibrated lens, then you will be underexposing slightly due to loss of light thru the lens. When in doubt, you can open the f/stop on the lens by 1/3 of a stop or so to compensate.

3. Avoid stray light hitting the spot meter’s lens, shade it with your hand. Any lens flare will alter your readings. 

4. When using a spot meter, the angle of the reading makes a huge difference. You can take readings before the camera is in place, but it’s a good idea to double check after the frame is up so you get a more precise reading from the exact angle. 

5. A spot meter is just a tool, sometimes it’s useful and other times not. Depends on the situation. When shooting film, you have no other choice for precise reflected readings. With digital cameras, a waveform or false color will often be more precise. When in doubt, I would trust the waveform. 

[Tristan Noelle]

If I had to guess, it has to do with the actual objects you metered. Some of the envelopes(?) should read white but the meter wants them to read gray, so it gives you a recommended darker exposure.  The spot/reflectance meter need to be interpreted more so than an incident reading.   To my eye, you probably should have exposed at an f/2, what the gray object to the left, is reading, or taken a reading from an 18% gray card at the proper angle. Honestly, an incident reading would have prob been more helpful in this specific situation. 
 

If you haven’t read up on the zone system, it’s a good guide to interpreting spot readings. IIRC, Blaine Brown has a write up of applying the zone system specific to Cinematography in his his book, “Cinematography” by Blaine Brown. 
 

https://en.m.wikipedia.org/wiki/Zone_System

-Tristan Noelle

[Vitor Delduque]

Hi, Satsuki!

Thanks for replying ?

1. Done. Using a white card, it matches perfeclty. Using different surfaces (leather, post-its, blue sheets, a cat, etc) it's kinda inconsistent. I've calibrated the l858 to 45% at ISO640 (Fuji XT-3). I've got values between 25-50%. 

Quick test: https://drive.google.com/file/d/1djiSs0jfr4ClqLpHmiziN_kJOgdF25Vj/view?usp=sharing

2. Cine Rokinon. The Calibration above was made with this lens.

3. Tried that today. Dark blue sheet lighted from brehind the camera. Got something around 25%.

4. I tried to stay as close as possible to the camera. The test above is the result.

5. That's the thing: is this kinda variation normal and the readings not aways precise? I'm wondering if there's something wrong with my tool ?. I really hope it's a lack of experience on my part because this I can fix with some pratice, hahaha!

Thanks again!

 

[David Mullen ASC]

Not sure you should be spot metering to a white card -- "white" can be any number of luminance values above a certain level.  If you're going to be spot metering a card, it should be an 18% grey card.

Based on your still life arrangement, the closest thing to 18% grey was closer to your f/2.0 reading.

[Satsuki Murashige]

6 hours ago, David Mullen ASC said:

Not sure you should be spot metering to a white card -- "white" can be any number of luminance values above a certain level.  If you're going to be spot metering a card, it should be an 18% grey card.

Based on your still life arrangement, the closest thing to 18% grey was closer to your f/2.0 reading.

Yes, I mentioned the white card specifically to check your highlights. Check exposure with the 18% grey card, absolutely. 

[Vitor Delduque]

7 hours ago, David Mullen ASC said:

Not sure you should be spot metering to a white card -- "white" can be any number of luminance values above a certain level.  If you're going to be spot metering a card, it should be an 18% grey card.

Based on your still life arrangement, the closest thing to 18% grey was closer to your f/2.0 reading.

Thank you both ?

I thought that every surface and color could be measured with the spot and it would be at 45% (or 41%, calibrated for Slog3) if the camera and suggested values were matching. Am I wrong about that?

I'm struggling a little bit to understand the different results I had. 

Thank you for all the help and patience!

[Satsuki Murashige]

10 hours ago, Vitor Delduque said:

Using different surfaces (leather, post-its, blue sheets, a cat, etc) it's kinda inconsistent.

That makes sense, since those objects are all different exposure values. So you should get different readings. 

The spot meter should be calibrated to 18% grey (Sekonic is notorious for calibrating to a different value in their meters, but that is a whole other topic). It assumes that you want to know what exposure to set in order to render the object you are metering at 18% grey on film. Therefore, if you are metering something other than an 18% grey card, your reading will require interpretation.

Let’s say you light a scene with an 18% grey card and a white piece of paper. The grey card meters at f/5.6, 1/60sec, 800 ISO. You meter the white card, which reads at f/16. Then your black cat wanders over and sits on the white paper. You meter her fur at f/1.4. Is your meter broken? No - it’s telling you that your grey card will be exposed correctly at f/5.6, that the white paper is 3 stops over, and your black cat is 4 stops under. Which is fine, you want the paper to look white and your cat’s fur to look black. But if you didn’t meter the grey card first, you wouldn’t necessarily know that.

[Satsuki Murashige]

1 minute ago, Vitor Delduque said:

I thought that every surface and color could be measured with the spot and it would be at 45% (or 41%, calibrated for Slog3) if the camera and suggested values were matching. Am I wrong about that?

Based on some Googling, it seems the middle grey target for Fuji F-log is 46%.

If you take readings of the white paper or your cat’s black fur, and set the camera and lens to those settings (and if your camera’s ISO is accurate) - then yes, they should come out at 46% on the waveform, or 18% grey. But then you have to decide if you want white or black objects to look grey? That’s where your interpretation of the meter reading comes into play.

If you want the white paper to look white, then you have to expose it brighter. You could arbitrarily decide that it should be 2 stops over, or 4 stops. But you could be more consistent with your exposures if you first knew how bright the grey card was. Then you have a reference to compare the white paper to. I think the spot meter is more useful this way - having already set your base exposure, you can check how bright the windows are in the background, or how dark the shadows are, and adjust them if necessary. 

[Vitor Delduque]

7 minutes ago, Satsuki Murashige said:

If you take readings of the white paper or your cat’s black fur, and set the camera and lens to those settings (and if your camera’s ISO is accurate) - then yes, they should come out at 46% on the waveform, or 18% grey. 

That's it. It's not aways coming out at 46%, or 18%.

If this is happening, even if I set the camera using a grey card and then measure something brighter like a white card, I won't be able to tell if the difference between them is real. 

Edited May 28, 2021 by Vitor Delduque

[Satsuki Murashige]

Just now, Vitor Delduque said:

That's it. It's not aways coming out at 46%, or 18%.

If this is happening, even if I set the camera using a grey card and then measure something brighter like a either card, I won't be able to tell if the difference between them is real. 

Have you tried metering an 18% grey card yet?

If you have, and it’s still not hitting 46 IRE on the waveform, then there is something else going on. But I would try this first before doing anything else. 

[Vitor Delduque]

16 hours ago, Satsuki Murashige said:

Have you tried metering an 18% grey card yet?

Hey,

I've tried it today using 10 different light conditions. I've positioned the card and camera, lighted the card, got the mettering, applyed the suggested settings on the camera (same ISO, adjusting aperture and S.s)... 10 times. All the shots stayed between 42 and 48%. The ones lighted with different CCT values were way more consistent: 44-47%. I used a portable RGB in 3 scenarios - because I was curious - and those were the ones that expanded the range to 42-48%.

Edited May 28, 2021 by Vitor Delduque

[Satsuki Murashige]

6 hours ago, Vitor Delduque said:

The ones lighted with different CCT values were way more consistent: 44-47%.

I’m not sure what you mean by ‘different’ CCT values, but 44-47 IRE sounds pretty accurate if the target for F-Log is 46 IRE. That would also indicate that the ISO calibration is accurate on your camera and should match the ISO values on your meter, which is great. 

Generally, a luminance (luma) waveform will be more accurate if the camera is white balanced to the light source because the RGB channels will be balanced. The RGB channels get averaged to display the luma signal, so depending on how far the color temperature is from neutral white balance, the IRE value may drift pretty far from 46%.

For example, if you light the grey card with a cool blue color temperature, the red channel would be lower that 46 IRE and the blue channel will be higher. You can see this if you look at an RGB parade which shows the IRE values of each individual color channel.

[Vitor Delduque]

Just now, Satsuki Murashige said:

1. I’m not sure what you mean by ‘different’ CCT values..

2. Generally...

1. It's just a poor English case, hahaha. Not the best, but it'll be improved soon.

2. Got it! Thanks ?

I couldn't understand what's happening to my readdins to make the sekonic suggest values that put the target at a lower IRE in some cases.. mostly when I try to measure dark surfaces. 

Thanks for your patience and all the content! It'll help me a lot.