1DC: 1DC exposure tests: best method for judging light exposure???

N

Nicholas Natteau

Veteran
I did some exposure tests on a WWII miniature action figure using my Canon 1DC.

To help judge exposure, I used three tools: the 1DC's histogram, a SmallHD DP7 Pro monitor and a Sekonic L-478D light meter. I shot in Canon C-Log 4K 24p at ISO 400, 1/50th shutter, shooting at two different f stops: 4.0 and 2.8

I'd like to know which of the two looks better exposed. The Sekonic lightmeter indicated that my f stop should be set to 2.8. Although this seemed too bright from just looking at the scene. So I then set the aperture to f/4
Please see the images here.

Am I overexposed on the waveform monitor at f/2.8 in C-Log? Or am I underexposed at f/4?

Will do more tests. Please see photos of test here:

http://www.flickr.com/photos/99291817@N07/sets/72157635110178191/

Thanks very much in advance for your feedback.
 
Nicholas Natteau said:
I did some exposure tests on a WWII miniature action figure using my Canon 1DC.

To help judge exposure, I used three tools: the 1DC's histogram, a SmallHD DP7 Pro monitor and a Sekonic L-478D light meter. I shot in Canon C-Log 4K 24p at ISO 400, 1/50th shutter, shooting at two different f stops: 4.0 and 2.8

I'd like to know which of the two looks better exposed. The Sekonic lightmeter indicated that my f stop should be set to 2.8. Although this seemed too bright from just looking at the scene. So I then set the aperture to f/4
Please see the images here.

Am I overexposed on the waveform monitor at f/2.8 in C-Log? Or am I underexposed at f/4?

Will do more tests. Please see photos of test here:

http://www.flickr.com/photos/99291817@N07/sets/72157635110178191/

Thanks very much in advance for your feedback.
Click to expand...

Exposure is usually subjective, but I would say the 2.8 looks more healthy. C-Log seems to favor the highlights more than the other picture styles, and you have plenty of room in the highlights looking at your waveforms and histograms, whereas the shadows look a bit crowded at f4.
 
Under exposing c-log shows its ugly head in post when you try to "correct" it. If you grade the two exposures, I would say the f2.8 is going to grade better.
 
I agree with you the screen looks too bright for the 2.8 ( including the face) . Experience told me is best to trust the light meter
 
View Assist always makes me think I am exposing hot, but it behaves different from the other picture styles. ymmv. Nice to have the wave form, but sometimes I am in guerrilla mode and going off the LCD back.

Even though VA is the "corrected" c-log, it is not crushing the shadows as much, and so if you do not see shadow info on the LCD, it is usually not there, whereas when highlights look hot, they are usually fine. That is my experience.


I would suggest Nicholas to grade the footage and see which one is better. It is easy to imagine shadow detail in c-log, that's what our brains do naturally.
 
I'm disregarding the images completely here (since there are too many variables the way they were captured) and going by the histograms. Under these circumstances I'd completely agree with James that the 2.8 exposure is healthier.

While a healthy Canon Log exposure CAN have an empty right side on the histogram, most of the time I want to hit at least IRE60, like in the 2.8 here. Everything above 80 is probably sky, a practical in the scene or a window. Reflected light should fall below IRE70.
 
Thank you very much, everyone, for all your very helpful feedback. Hi James, I haven't yet graded, but I'm just about to. And will post my results asap.

In my initial post, I forgot to ask about skin tones. In Canon C-Log with VA on, for talking head interviews, where should Caucasian skin tones fall on the IRE scale?

And on a white Lastolite Ez balance card, am I correct in assuming that the white should never be above 90% IRE on the waveform in C-Log with VA on?

Thank you again everyone.
 
Last edited:
Nicholas Natteau said:
Thank you very much, everyone, for all your very helpful feedback. Hi James, I haven't yet graded, but I'm just about to. And will post my results asap.

In my initial post, I forgot to ask about skin tones. In Canon C-Log with VA on, for talking head interviews, where should Caucasian skin tones fall on the IRE scale?

And on a white Lastolite Ez balance card, am I correct in assuming that the white should never be above 90% IRE on the waveform in C-Log with VA on?

Thank you again everyone.
Click to expand...

Depends.... always depends. Haha (hint a little above mid and tweak to intent. but if it is used as a rule in cinema, it will create a industrial or commercial look.)
 
Last edited:
On a speculative note:

C-Log (especially at the same exposure settings, aperture, iso, shutter) has more allocated to the highlights, so exposing skin I would go for lighter. Whereas on neutral, I would under expose skin a bit more, maybe at or below midway wouldn't bother me. This is especially true when using a light meter.

(while many do not talk about c-log being darker than neutral, even Shane Hurlbut talks about how muddy c-log gets when under exposed, but this is not a defect, just something to be aware of, as the highlights are fantastic.)

But with a waveform, I would start around mid way and go up or down depending on the subject matter. In C-log and soft light it is usually up, especially if you want that clean bright "skin tone" look in demo videos. But once you go over 70% (true 70% not c-log), you are getting into crispy territory. Since a face is not flat, you will see a range of shades, so mid is a good starting point for me, but that doesn't mean that I am not seeing peaks, just the overall face is in the middle. But you can see from my shooting that I am all over the place with exposure.


Disclaimer:
IRE readings for c-log 32% = 50%, yes? So, when I say "mid', mid = middle gray, in theory, but since C-log can be used with view assist fed to the waveform monitor there are just too many variables, so I say "mid". And then you sort out your own formula. Saves confusion hopefully, though I know it is less specfic.


Somewhere I heard about skin tone, and I tend to agree since every camera and person is different "throw out the old rules and just make people look nice". Dial in your own rules with each camera you use. May differ as much as 20ire. May be flexible on one camera, and specific on another. With 1D-C using soft light: in C-Log, err on over exposed, in Neutral err on slightly under.


All of these screenshots were shot in C-Log

The blue one is under, the yellow one is "standard", and the red one is over. I wanted the red to be a bit garish, and digital. (also, since the red one is near monochrome, I could easily use the green channel and then tint it red to get a fixed exposure, but I wanted to look a bit glaring. I used this example to show that "intent" is king, and non-standard looks can work)

Additional reasoning:
I do not mind the blue one being under exposed because she is back lit. Not to say backlighting, but the frame and situation dictates darkness on her face. The shadow on the face is motivated.

In the yellow frame, the subject is front lit with soft light. His face takes prominence in this shot, and as painters know, the brighter objects stand out. He definitely pops from the dark background.

In the red frame, it is extreme contrast, no way to get even exposure between the extremes, and also, I wanted an annoying or abrasive feel.
 
Last edited:
Thanks very much James, those examples were very helpful.

In your three examples were you shooting with a waveform monitor on set?
And if so (before you applied any color correction) what IREs were the lit part of your actors' faces registering?
 
Nicholas Natteau said:
Thanks very much James, those examples were very helpful.

In your three examples were you shooting with a waveform monitor on set?
And if so (before you applied any color correction) what IREs were the lit part of your actors' faces registering?
Click to expand...

These were from a test that I was working with the in-camera exposure tools (V.A. and histogram), so I did not get an exact IRE reading. Which is why I used these examples... the rules are guidelines. (ok, I also used them because they were already uploaded online, and it was easier than digging through footage, haha.)

I do know that the yellow one was somewhere around mid for the average illumination on the face, and brightened a bit in post. The yellow one could have opened up another half stop or more, but was close enough, and it gave me more color in the lamps behind him (I could see the lamps peaking from the histogram). And the Red one is actually clipping in the red channel in the brightest parts of the face, so it is off the charts as far as rules go.

Next time I get to editing, I will grab a more standard frame. Anyway, hope this all helps.
 
I know this doesn't help practically, as it is better to learn theory then break rules in general, but good words none-the-less:

"There is no formula. The formula comes out of you. So, whether it’s a top light or whether it’s some other thing. It just happened to be — that’s what was necessary to do this particular movie or this particular scene. So, I did it. Bottom line is, the design behind all of that, or the thinking behind the design of all of that came out of Marlon Brando, because Marlon had this makeup stuff he was using, so top light seemed to be the most effective way of dealing with him. You don’t really want to see his eyes. There was a big Hollywood rush about, “You can’t see his eyes.” That’s right. You can’t." Gordon Willis
 
This may be more telling about the latitude in regards to skin: one shot is over exposed by 3stops and the other under exposed by 3 stops and then they are corrected in post and look like this. while they both struggle, the end result of correcting the over exposed image is better than the under exposed and corrected one.


On the one hand, this is a front lit scene and is bias towards needing to look bright, but both shots looked bad before fixing, and both have been corrected and brought back as much as they could without adverse effects.

Here is a test shot I did which was well under exposed:

Left (original c-log) , Right (corrected)

My test shot above was in HD, not 4k, so it less forgiving, but noise and lack of detail are evident enough when it is pushed.
 
Last edited:
Sadly, I do not know yet if Shane Hurlbut and team were using C-Log. I can only assume at this point that they were, as Shane seems to prefer it and used it almost extensively on "The Ticket". Also, the behavior of the latitude in this test seems to be more like C-og, but I could be very wrong.
 
Last edited:
If you are accustomed to shooting raw (any raw stills or motion cam, except Arri maybe) or with other Canon picture styles, you will most likely have developed the habit of under exposing. C-Log is different on the 1D-C and does not have near the same shadow detail at the same exposure settings.

My general rule with C-Log, is if I do not see actual shadow detail (seems pretty straight-foward), then it is not there. With a waveform I look at my mids and shadows far more than I ever did with other digital cameras. In that way it is more like film.
 
Last edited:
James, thanks again very much for sharing all this information. Very kind of you to take the time to explain it as well. I will be doing more tests and uploading shortly.
 
Nicholas Natteau said:
James, thanks again very much for sharing all this information. Very kind of you to take the time to explain it as well. I will be doing more tests and uploading shortly.
Click to expand...
No problem, I can only hope that my experience translates into something that works well for you too.

Exposing exactly with the 1D-C has proven trickier than any other professional camera I have worked with, because it has more variables and no built in waveforms or zebras.

There are variables when using an external monitor for exposure, for example, you have to make sure what is coming through the hdmi. If it is c-log, then all of the equations have to be brought into account (32%-middle grey), if it is view assist, you have to learn that by experience, as there are no specifics on what View Assist is really doing.

Then if you are shooting in Neutral, well now you lose some highlights and gain some shadows. Now your light meter is way off.

At least with a waveform or false colors you can see what is clipping regardless of percentages.
 
Last edited:
If you are doing some tests, could you grab a dark shot comparison between C-Log and Neutral. Get them really close to complete black, and see which one has more shadow detail. I am the only one that has mentioned this, but look at this test:


shadowdetail_eos by onesielaff, on Flickr

Both images have been pushed to show maximum detail. they are both crops from 4k footage. Which one has more detail? A or B?
 
You must log in or register to reply here.
Back
Top