Reading mode: Light Dark

Exposure and 18% grey cards...

Go to last post
Martin Ocando said:
I understand, but isn't the grey card supposed to reflect less light than the white card, so in theory should appear differently toned, even at different exposures?
Click to expand...
No. Ideally the meter would read the greater amount of light reflected by the lighter target and adjust the exposure so that they have the same tone. Remember, the purpose of a meter is to adjust exposure for the total amount of light reflected by the subject. The meter doesn't know whether that is from a white card or from a gray card illuminated by a brighter light source. All it can do is try to change exposure so the amount of light gathered is the same.

Dave
--
http://www.pbase.com/dsjtecserv
 
Right. There is no color information. Just luminance.

Thanks
dsjtecserv said:
Martin Ocando said:
I understand, but isn't the grey card supposed to reflect less light than the white card, so in theory should appear differently toned, even at different exposures?
Click to expand...
No. Ideally the meter would read the greater amount of light reflected by the lighter target and adjust the exposure so that they have the same tone. Remember, the purpose of a meter is to adjust exposure for the total amount of light reflected by the subject. The meter doesn't know whether that is from a white card or from a gray card illuminated by a brighter light source. All it can do is try to change exposure so the amount of light gathered is the same.

Dave
--
http://www.pbase.com/dsjtecserv
Click to expand...
--
Martin Ocando
-------------------------

 
Federico said:
when I take a picture of the white side (filling the whole frame, no exposure compensation) I don't get the same shade of grey I get when shooting 18% grey. The white surface is of course underexposed but still lighter than what I get with 18% grey. The histogram is shifted to the right.
Click to expand...
I see what you mean (using a WhiBal card), although the histogram shift between the grey and white images is very slight for me, less than the shift given by 1/3 of a step of compensation. I see the same effect with spot, partial, or evaluative, so it's nothing to do with the metering mode. It's consistent in each case, so it's not simply the effect of the meter vascilating between two values for the dependent parameter (e.g. 1/640th... 1/800th... 1/640th...).

I noticed that the luminosity and channel curves for the grey image are all "wider" than those for the white image, suggesting that, although the grey card looks like a single tone, it is made up of a broader range of tonal values than the white card. (That could also be caused by things like the white and grey sides having different textures.) So I'd bet it's that different in the quality of light from the two cards that's causing what we're seeing.
 
Federico said:
Yesterday I bought an EzyBalance Calibration Card to do some exposure tests with my camera and flash.
Click to expand...
It's a nice tool to have.
Federico said:
What is not clear to me is the way the camera meters on the white side in P/Av/Tv mode.
Click to expand...
The white side is 90% reflectance, which is 2 1/2 stops difference from 18% (or 3 stops from 12%). This side is used to meter in dimly lit locations.
Federico said:
I have to figure out what the guide number of an LED flash system I'm working with is. In order to do that I need to create some reference exposures to compare to the LED flash's output.
Click to expand...
The good news is that the process is a lot simpler than it first appears. Do this...

Find and average size room (most manufacturers state a "room with average reflectance"). Light bouncing off of walls and ceiling will affect your flash strength (guide number). Very few manuf adhere to any ANSI standards when speccing Guide Number. So just choose a typical room like one you might be shooting in.

Now set up your grey card 5 feet from your flash.

Set the flash to Max output.

Set your camera to Manual Exposure mode (metering mode doesn't matter).

Set your shutter speed to the x-syc speed of your camera or slower (I usually like to set it at half the x-sync, or even better yet 1/60 sec). This has to do with flash duration.

Set your ISO to 100.

Dim the ambient lights.

Now pick any aperture and fire off a shot. Check your shot and adjust your aperture until the grey card is at 18%...

If you can "see" the grey card spike in your histogram, then shoot and adjust until it is just to the right of middle.

Otherwise open the image in Photoshop and use the Eyedropper on the grey card (it should read 128 - 128 - 128). 18% grey.

Now read the Aperture from the frame with the correct exposure, and muliply that times the distance (5 feet). This is your flash's Guide Number in feet (at ISO 100).

Example: f16 x 5 feet = guide number of 80 (feet).

or

Example: f8 x 5 feet = guide number of 40 (feet).

or

Example: f5.6 x 5 feet = guide number of 28 (feet).

Now if you have a Really powerful flash, you can do the same test at 10 feet (then multiply the Distance times the Aperture to get the guide number).

Example: f22 x 10 feet = guide number of 220 (feet).

To use this newly found information (guide number), just divide whatever distance you will be shooting at, into the guide number to get the proper aperture. Given that you keep your ISO at 100.

Example: Your guide number is 80, and your shooting distance is 20 feet, (80 divided by 20 = f4). Shoot at f4 from a distance of 20 feet away. If you distance changes to 10 feet, then you'll be shooting at f8. If your distance is 5 feet, then you'll be shooting at, voila... f16).

In the studio, this becomes Very easy to do. Even when out and about, it gets to be second nature very quickly.

Now say for instance you want to raise your ISO to 200, well just multiply your guide number by 2. Easy.

Or if you want to raise your ISO to 800, just multiply your guide number by 8 (this gives you a lot more range!).

If your flash has manual power settings (for instance if you want to shoot at 1/2 power), then just divide the guide number in half, and adjust your aperture appropriately.

Clear as mud?

Like I said, it's a lot easier in practice. Just keep in mind that room reflectance can affect flash strength significantly. Similarly - shooting outside, or outside at night even moreso.

Well, good luck to you. Hope your project goes well. I've just purchased some new flash modifiers and need to do the same tests myself!

Cheers,
R2

--
Good judgment comes from experience.
Experience comes from bad judgment.

http://www.pbase.com/jekyll_and_hyde/galleries
 
Karl Gnter Wnsch said:
As for the white card not being far enough underexposed, that's due to the subject database the camera has, it probably evaluated the scene as being a shot in fog or on a snowy slope and factored that into the exposure.
Click to expand...
So it sees a foggy scene, then a second later when it sees an identical scene but one Ev step darker, it says, oh no, that's not a foggy scene, so I won't apply a fraction of a step of compensation to that one? I don't think so.
 
Fernando

Your original assumption should be correct but as some others have mentioned, the camera's artificial intelligence must be adding some form of correction to the exposure. In real life, its probably exactly what you'd want. For testing, not so much.

If you haven't already, do try the different metering modes and shoot this on manual to see if you can get the same shade of gray on both 18% and 0% gray.
 
R2D2 said:
Now say for instance you want to raise your ISO to 200, well just multiply your guide number by 2.
Click to expand...
Oooops, it was getting late when I posted last night.

If you raise your ISO from 100 to 200, then your guide number gets multiplied by 1.5 (not 2). Inverse Square Law.

So if you raise your ISO from 100 to 400, then your guide number doubles. From 100 to 1600, then your guide number quadruples. Etc etc.

Similarly with adjusting flash power manually. 1/4 power will cut your guide number in half. 1/16 power and your guide number gets divided by 4. Etc etc.

Thanks,
R2

--
Good judgment comes from experience.
Experience comes from bad judgment.

http://www.pbase.com/jekyll_and_hyde/galleries
 
I think the problem is something along these lines:

The 18% grey card is not, necessarily 18% grey in colour but 18% reflectance. The card reflects 18% of all light giving it a grey appearance. If it reflected more blue and less red and green it'd appear light blue, though it might still reflect a total of 18% of the light falling on it. 18% grey ensures that the reflected light always matches the spectrum of the incident light (this is what means you can use it for white balance).

The white side seems like it should be, intuitively, 100% reflectance (ie mirror like) but in reality is probably more like 90% reflectance.

Not sure what help that information is though.
 
Hi R2, very nice post, it comfims the theory I have in mind about how to handle flash power is correct (including your correction post about ISO speed).

And I'm happy because this gives me the opportunity to go a little bit more in detail. ;-)

I had in mind a slightly different way to evaluate the guide number. The main problem I see in you workflow is assuming that 18% gray card is equivalent in RGB to [128,128,128]. Googling a little bit I've found so many differet values for RGB or CMYK that now I don't know which one is correct. My understanding, as you and most of the posters say in this thread, is that 18% grey is a "surface" reflecting 50% of the incident light.

But: does it translate in RGB 128 128 128? Probably yes as if you enter it in photoshop this is the equivalent of HSB (0,0,50) and CMYK (47,37,37,18). B=50% and K=18% sound like nice numbers... ;-)

My solution to this problem is using a flash whose guide number is well known to create a reference taking a picture of the grey card in manual mode (controlling shooting distance/f number/ISO speed). The exposure time is set to a value which makes ambient light negligible.

I used for this test both the pop-up flash (GN=13 in meters) and my 430EX (GN=43 in meters).

The result (shooting 1 to 3 meters away from the target) is usually consistent: same exposure trying different combinations of ISO-f/-distance.

Well, the value I read in photoshop is usually around RGB(104,104,104) which is darker than 128,128,128. In order to avoid pixel to pixel variation I sample the color after applying a 10 pixels gaussian blur to a selection of the grey card and use 5x5 pixel average for the eyedropper tool.

Now, what happened? Candidates:
  • camera's ISO sensitivity is wrong (but in 60D's review they claim its accuracy is within + - 1/6EV);
  • f/ not accurate? (Lens: 15-85 f3.5/5.6 IS USM);
  • both flashes have wrong GN? O_o
  • 18% grey card is not 18% grey...
  • ...?
Also, during my tests I have also noticed that using the pop-up flash at short distance (around 30-60cm) the exposure is much darker than expected...I suspect this might be caused by to the pop-up flash not having a reflecting parabola which adapts to the focal lenght like in 430ex. Not sure about that, I have to investigate...

Actually it is not that important whether I get 100,100,100 or 128,128,128, I just need to know which is the reference value. If let's say my 430ex' GN is correct I can assume that the problem is elsewhere (grey card, actual ISO sensitivity...) and use the output I get with it as a reference...

The main problem I see now is the different result at low distance as LED flashes are not that powerful so they're supposed to be used at short distance...

Lot of tests to do, any other comment is welcome.

Cheers,

Federico
 
Did some more tests...discouraging results.

Setup: 60D+430ex+50mm F1.8 II

The manual of my 430ex states that at 50mm the guide number is 34 (ISO 100, in meters).

I therefore expect approximatelly the same exposure for the following random combinations:

Flash power 1/1:
F22-> 1.5m
F13-> 3.3m

Flash power 1/2:
F18-> 1.3m
F8-> 3m

Flash power 1/16:
F8-> 1m
F5-> 1.7m

Flash power 1/32:
F5-> 1.2m
F2-> 3m

Flash power 1/64:
F3.5-> 1.23m
F1.8-> 1.95m

Well, the results are erratic, only the fourth picture (1/2 F8) shows RGB 128,128,128.

Most of the others are darker and a couple are much lighter (1/32 F2 and 1/64 F1.8). The bottom line is that whatever the numbers are, the exposure changes moving from one setting to the other...
Not nice.
 
Federico said:
Hi R2, very nice post, it comfims the theory I have in mind
Click to expand...
And unfortunately it really is just theory. In practice, published guide numbers are often way off (you have to do your own empirical testing). Kudos to you here.
Federico said:
The main problem I see in you workflow is assuming that 18% gray card is equivalent in RGB to [128,128,128]. Googling a little bit I've found so many differet values for RGB or CMYK that now I don't know which one is correct.
Click to expand...
If you use sRGB and your monitor gamma is the standard 2.2 (PC), then you can use 128-128-128. No need for CYMK etc.
Federico said:
My understanding, as you and most of the posters say in this thread, is that 18% grey is a "surface" reflecting 50% of the incident light.
Click to expand...
Actually it reflects 18% of the light. And the white side reflects 90%. "Medium Grey" comes from the old Printing standard, and does not denote 50% reflectance.
Federico said:
My solution to this problem is using a flash whose guide number is well known to create a reference taking a picture of the grey card in manual mode
Click to expand...
And there is where you will run into problems using that approach. Manuf specced guide numbers can be waaay off, and most seem to cheat! Their idea of a room of "Average Reflectance" must be one made up completely of mirrors!
Federico said:
  • both flashes have wrong GN? O_o
Click to expand...
Yes, the most likely scenario, unfortunately.
Federico said:
I suspect this might be caused by to the pop-up flash not having a reflecting parabola which adapts to the focal lenght
Click to expand...
No. The physics govern the propagation of light (inverse square law again). A zoom head will only extend the coverage either to the edges (wide angle), or increase the reach (long distance). A simple flash with a fixed reflector will adhere strictly to the formula (at least when on-axis). The likely culprit again is an inflated published guide number.
Federico said:
The main problem I see now is the different result at low distance as LED flashes are not that powerful so they're supposed to be used at short distance...
Click to expand...
Oh yeah, they have a long ways to go yet. However IMHO LED flashes and LED "Hot? Lights" will become very prevalent in the coming years.
Federico said:
Lot of tests to do, any other comment is welcome.
Click to expand...
Also keep in mind that the manual power settings with most flashes can be off quite a bit too (again, need to do your own testing to come up with your own guide numbers). You'll have much more luck (if your setup permits) adjusting your flash power by changing the flash to subject distance (the physics won't lie).

Best of luck!
R2

--
Good judgment comes from experience.
Experience comes from bad judgment.

http://www.pbase.com/jekyll_and_hyde/galleries
 
18% gray is actually 118,118,118 - not 128,128,128.
R2D2 said:
R2D2 said:
The main problem I see in you workflow is assuming that 18% gray card is equivalent in RGB to [128,128,128]. Googling a little bit I've found so many differet values for RGB or CMYK that now I don't know which one is correct.
Click to expand...
If you use sRGB and your monitor gamma is the standard 2.2 (PC), then you can use 128-128-128. No need for CYMK etc.
Click to expand...
 
R2D2 said:
KLO82 said:
18% gray is actually 118,118,118 - not 128,128,128.
Click to expand...
You are right. Those were the values for my old monitor/printer, DOH.

PV .4635 x 255 = 118.1925

Thanks,
R2

ps. The OP's figures should look a little better now.
Click to expand...
Yes, they do. But still erratic. The average is around 105,105,105 with extremes at 95,95,95 and (!) 160,160,160....

Could you please tell me where the numbers come from?

Thanks!
 
Karl Gnter Wnsch said:
ESfishdoc said:
I've never heard of using the 18% grey for proper exposure. It is used for white balance in post processing.
Click to expand...
Wrong, the 18% grey card is used for proper exposure.

regards
Karl Günter Wünsch
Click to expand...
Ding ding ding! We have a winner! :)

Karl is correct.

ESfishdoc, you obviously weren't around in the days of handheld meters and manual film cameras.
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top