Measured incident versus reflected light in camera calibration question

[plankton]

It's raining here in Northern California (States) so what better activity to occupy ones time than to compare the incident light on a 18% card versus the reflected light measured by ones camera?

Here is the setup. I have a 18% grey card in a stand 3 feet from camera. To camera left at 45/45 degrees is a flash head in soft box (24" square) at 1/4 power.

Camera and Sekonic L-308s are setup for shutter priority, ISO=200 and 1/60 sec shutter.

The sekonic measures f/11 incident light. I set the camera in manual mode with f/11 aperture, iso=200 and 1/60 shutter.

Results:

I take a picture expecting the vertical bar mid way on histogram...well because it is 18% grey card and the sekonic told me the perfect light setting would be 60/f11. Well, what I got was a bar 1/3 to the left or dark side of middle. If I open up the camera from f11 to f9 the bar moves nicely to the middle of the histogram - where it should be.

Both my camera and sekonic setup for 1/3 stop increments, so dropping the apeture from F11 down to F9 (F11->F10->f9) is 2/3 of a full stop (F11->F8 being 1 full stop). So, there is a 2/3 of a stop difference between my sekonic and camera. Wow.

The 308s doesn't appear to allow me to adjust the setting to match camera. Darn.

So, if the sekonic is to be trusted then am I shooting 'dark' by 2/3'rds if I measure light by my camera alone?

PS The white balance was set to 'Flash'.

Scott

[Colin Southern]

Couple of points ...

1. Light meters generally don't measure 18% - and a lot of grey cards aren't 18% either (they tend to get used more for white balancing). It's more like 12 or 13%, but the number is so widely used we tend to just save an argument and keep using it.

2. The histogram you're looking at is after the RAW to JPEG in-camera conversion, and it's likely that the camera has applied a tone curve anyway.

In my experience, sekonic meters are deathly accurate; I use the 758DR which you can profile to an indifidual camera, but after spending $200 on the targets I've found that I didn't need to change a thing.

PS: I'm a bit confused as to your testing methodology - where do the strobes enter into the equasion? If they're manually set to 1/4 power, are you triggering them, or just relying on the modeling lights?

PS: THom Hogan explains it better than I do ... http://www.bythom.com/graycards.htm

[plankton]

"PS: I'm a bit confused as to your testing methodology - where do the strobes enter into the equasion? If they're manually set to 1/4 power, are you triggering them, or just relying on the modeling lights?"

Hey Colin - Yes, I'm using flash to light the 'grey' card and switch between sekonic triggering the strobe via cable and camera via wireless slave. I set the power to 1/4 to get a mid aperature that's all.

[Update]. Wow, just read the Thom Hogan article and that prolly explains it. I did buy a 18% card and like most people noticed that the camera puts it slightly to the left of center on the histogram. If Nikon et al are really using 12%'ish iso standard (luminance) then that may vary well account for the difference (on the camera). How odd.

Thanks.

Scott

[Colin Southern]

In all honesty Scott, I think the best rule is "so long as the highlights aren't blown, don't worry about it"! With todays cameras there's so much clean detail captured well below what our monitors can display (lets call it "safety margin") that minor under-exposure just isn't a real-world issue (assuming a RAW capture, and a mostly reflective scene). Over-exposure is a bigger problem (much smaller safety margin).

[Policar]

One other question: Did you pop the lumisphere out or leave it recessed? When metering flat art you should leave it recessed or you may get too bright a reading (and thus underexpose).

I don't think 18% gray is meant to correlate exactly with 127/127/127 gray, anyway. 18% gray is more of an analogue phenomenon. If a linear eight-bit digital sensor placed 18% gray at 127 it would theoretically only be able to see one stop over 18% gray (i.e. caucasian skin tone would be pure white) so there's obviously some behind-the-scenes trickery going on in the ADC. And a photo of a white wall shows a line on a histogram clearly left of center (at least with my older dSLR). When I light video, I place skin tones at key around 50 to 65 IRE, and I'd guess 18% gray around 35-45 IRE, which I gather correlates pretty closely with what calibrated film scans return.

Don't worry about it. Sounds pretty normal to me.

[Colin Southern]

If a linear eight-bit digital sensor placed 18% gray at 127 it would theoretically only be able to see one stop over 18% gray (i.e. caucasian skin tone would be pure white) so there's obviously some behind-the-scenes trickery going on in the ADC.

Umm, no. You've left gamma conversion completely out of the equasion, plus, mid gray on a printed page is approx 2 stops down from a highlight, and 2 stops up from a black (4 stop total dynamic range) whereas a typical DSLR has an 11 to 12 stop dynamic range, so they really don't equate like that.

[Fredrik FW]

Hi there,
this is a very interesting subject and I think the question put forward is asking for an extensive answer. I take a chance in giving a comment and will try to be short although that is unfamiliar to me.
I have been through the process of trying to understand metering since this is imp. for understanding the Sone System by Mr Ansel Adams which I used when doing B/W film. I am still coming back to the subject and are often scouting on the internet, and that was how I found this forum. The tutorials has a very good and compressed discussion of the matter. I can only rec the site Cambridge in color tutorials, happy reading.

Grey Card sometimes called the neutral card is the ref card you should use for setting your exposure time, so where ever you have a detail in your subject that item will be correct exposed, all other light values in your subject will get an exposure with ref to your "grey card" exposure. Using the grey card for WB is hard for me just coming from B/W film photographing. WB is complicated and therefor I have problems to se how a Grey card could help in setting White Balance in that case it could be named GB (Grey Balance). So for WB I use the Exposure Disk. Which also by the way contain a Grey Ref.
Check out Sone System on the internet, although it is intended for B/W it is helpful for color.

I do not have a camera where I can calibrate the light meter, but what I have done is finding out the exposure toll. of the light system of the camera. In my testing I have used the Sone System extensively, and of course in manual mode. This helps me in setting how much I could over expose a subject and still have good representation in the high lights when working with the picture later on.

If you will do a color check you set your camera in manual for measure of light from a uniformly colored item, adjust to 0-deviation reading on your light meter. Make an exposure, look at the R, G, B representation in the histogram, one histogram for each. They should fall exactly under each other with a 50% reading on the horizontal value scale. Repeat after you have done a correction any deviation

I hope I have not mixed up, in that case I hope for correction and that it will enlighten us further. Your subject is very interesting
since light is what we are recording. And don't forget Ansel Adams, a great photographer and explorer in light recording.

Kind regards from Fredrik FW.

[Policar]

Umm, no. You've left gamma conversion completely out of the equasion, plus, mid gray on a printed page is approx 2 stops down from a highlight, and 2 stops up from a black (4 stop total dynamic range) whereas a typical DSLR has an 11 to 12 stop dynamic range, so they really don't equate like that.

I know (except I've yet to see a dSLR that really has a 12 stop range). I was being hypothetical (no modern camera has an 8-bit ADC). There's no neat correlation with prints, though. Prints compress the dynamic range and lose some shadow detail; you don't suddenly lose 8 stops of information, most of it persists but is compressed.

All I'm saying is: cameras don't "see" 18% gray as 127/127/127 (I've seen linear sensor output--it's horrifying) and they don't record it like that either. It's best to treat 18% gray as some sort of analogue phenomenon and once things get digital not worry about it. 18% gray (in the digital world) is darker than "digital mid gray" and should be for any linear representation.

[Peter Ryan]

It's raining here in Northern California (States) so what better activity to occupy ones time than to compare the incident light on a 18% card versus the reflected light measured by ones camera?

Here is the setup. I have a 18% grey card in a stand 3 feet from camera. To camera left at 45/45 degrees is a flash head in soft box (24" square) at 1/4 power.

Camera and Sekonic L-308s are setup for shutter priority, ISO=200 and 1/60 sec shutter.

The sekonic measures f/11 incident light. I set the camera in manual mode with f/11 aperture, iso=200 and 1/60 shutter.

Results:

I take a picture expecting the vertical bar mid way on histogram...well because it is 18% grey card and the sekonic told me the perfect light setting would be 60/f11. Well, what I got was a bar 1/3 to the left or dark side of middle. If I open up the camera from f11 to f9 the bar moves nicely to the middle of the histogram - where it should be.

Both my camera and sekonic setup for 1/3 stop increments, so dropping the apeture from F11 down to F9 (F11->F10->f9) is 2/3 of a full stop (F11->F8 being 1 full stop). So, there is a 2/3 of a stop difference between my sekonic and camera. Wow.

The 308s doesn't appear to allow me to adjust the setting to match camera. Darn.

So, if the sekonic is to be trusted then am I shooting 'dark' by 2/3'rds if I measure light by my camera alone?

PS The white balance was set to 'Flash'.

Scott

From my prior reading of this issue I understand that the light meter in all cameras are set to a standard 13%. That's the norm. The 18% grey card was somethning Ansel Adams used as part of his Zone system but Kodak printed it as a standard measure, as dsicussed above.

[Colin Southern]

From my prior reading of this issue I understand that the light meter in all cameras are set to a standard 13%. That's the norm. The 18% grey card was somethning Ansel Adams used as part of his Zone system but Kodak printed it as a standard measure, as dsicussed above.

The interesting thing is though, since 18% reflects more light than 12%, why isn't the spike on the histogram to the right of centre, rather than the left?

[Peter Ryan]

The interesting thing is though, since 18% reflects more light than 12%, why isn't the spike on the histogram to the right of centre, rather than the left?

I agree Colin. In my tests I get the same result as Scott and always thought it should be slightly to the right. I have not been able to explain it and like you suggested as long as you are not blowing out the highlights of the total scene it should be ok. I decided I did not need to lose sleep over it and have moved on but it is still an unanswered question.

[Colin Southern]

I agree Colin. In my tests I get the same result as Scott and always thought it should be slightly to the right. I have not been able to explain it and like you suggested as long as you are not blowing out the highlights of the total scene it should be ok. I decided I did not need to lose sleep over it and have moved on but it is still an unanswered question.

At the end of the day, a reflective scene (or which "18% gray") is in the middle only covers around a 4-stop range ... the camera is capable of around 3 times that - so I'm guessing that manufactures have just popped the exposure a little lower into the dynamic range to add a little safety margin.

One of these days I'll get around to profiling my light meter, and will investigate further at that stage (maybe!).

[Peter Ryan]

At the end of the day, a reflective scene (or which "18% gray") is in the middle only covers around a 4-stop range ... the camera is capable of around 3 times that - so I'm guessing that manufactures have just popped the exposure a little lower into the dynamic range to add a little safety margin.

One of these days I'll get around to profiling my light meter, and will investigate further at that stage (maybe!).

I always thought my Nikon shot slightly underexposed and having tried this myself I decided to set my EC at +.7 as a starting point in the summer months. In winter I tend to drop it back to zero.

[plankton]

Great discussion all. My solution after reading all the comments was to pop the tab on another Asahi Super Dry beer and shoot some more pics ... I do like my simple Sekonic and it does give me a great starting point on exposure.

@Peter - I read in a wedding photography book that the old rule was to meter a wedding dress then over expose by +2 for nikon and +1.7 for cannon to set the white dress white. Not a scientific solution but I've seen these numbers floating around for some time and now I'm starting to understand why...tis where the camera manufacturers have set their mid-tone logic.

Scott