Proper Blind Exposure with a Gray Card

Started Nov 12, 2012 | Discussions
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
Proper Blind Exposure with a Gray Card
Nov 12, 2012

I would be interested in your opinion with regards to the following concepts on metering ‘blind’ with a gray card:

  1. that ‘blind’ exposure is defined in ths context as the exposure to be chosen with the sole help of a spot meter, without knowledge of the scene, time or place – it is the input quantity to the camera sensing medium
  2. that ‘proper’ blind exposure is defined in this context as the exposure that allows a camera to best record on its sensing medium a subset of the relative distribution of photons reflected from the scene; the full set is the relative distribution of photons that would be perceived by the human visual system in the same setting as the camera
  3. that brightness is an output medium dependent quantity typically not related linearly to exposure
  4. that for a given ISO the job of calibrated spot meters is to provide indications of shutter speed and aperture to achieve proper blind exposure
  5. that spot meters do not know or care what average scene reflectance or reference card is in front of them
  6. that in order to achieve proper blind exposure with most DSLRs it is best to use a standard Middle Gray card (that is one with 18% reflectance), which is what their spot meters are calibrated to
  7. that when metering off of and capturing such a target a DSLR will typically record digital raw values about 0.127 times the full scale reading of the sensor, subject to many provisos
  8. that if instead of a Middle Gray card a 12% reflectance card (or alternatively an 18% card with a ½ stop positive Exposure Correction) is used to set Exposure, the meter is fooled into raising Exposure by half a stop, reducing a DSLR’s designed highlight hard ceiling accordingly
  9. that if exposure is locked as per this last case, a Middle Gray tone will be recorded by a DSLR at raw values around 0.18 times the full scale of the sensor, so that a capture of an 18% Gray Card will record raw values of about 18% of Full Scale
  10. that other than a warm and fuzzy feeling there is no reason why those two percentages should be the same
  11. that we should all write an email to someone with whom I share a last name telling him that his blog entry on this subject is confusing a lot of peopleand does not stand up to his otherwise excellent standards

To help with the discussion I include this image:



Thanks for your input,
Jack

aardvark7
Veteran MemberPosts: 3,110
Like?
Equal confusion, or not
In reply to Jack Hogan, Nov 12, 2012

To a lay man, both your post and the blog entry would seem equally confusing (although some may say the technical aspect of your post has the edge!).

However, the answer to both lies in one simple fact that neither of you has proved unequivocally: the specification used by camera manufacturers for calibrating the internal exposure meter.

It would seem that your namesake believes, based on his observations and conversations, that it is the equivalent of a 12% grey card, whereas you understand it to be 18%. Until one is shown to be definitively true the argument is somewhat moot.

For my part, I appreciate his opinion about the influence of the print industy as I know only too well how the terminology and methodology from there has migrated into other fields. I am, for my sins, also a embroidery designer/puncher and I'm sick of customers asking for threads in Pantone x or y. The fact that Pantone is only ever printed ink on white paper seems to elude them, even when I try to educate. It is now so ingrained in people that believe Pantone colours are simply colours and can be applied to any field. They can't, even though they can give a very rough frame of reference.

-- hide signature --

2012 : My new year's resolution -
To be decided...

Reply   Reply with quote   Complain
nikkorwatcher
Contributing MemberPosts: 684
Like?
Re: Proper Blind Exposure with a Gray Card
In reply to Jack Hogan, Nov 12, 2012

Film responds non-linearly to high and low exposure values- this I knew.

Some Canon Ds habitually underexpose and need 1/3-1/2 stop correction - this I knew.

When using some Nikons I often need to underexpose 1/3 stop - this I knew.

As for whether some manufacturers are using 12% or 18% grey cards- I don't think this will fool reviewers. They will do ISO tests after correcting exposure. Or was there some other point, like bashing Thom?

Reply   Reply with quote   Complain
steephill
Veteran MemberPosts: 9,095
Like?
Pointless obsession
In reply to Jack Hogan, Nov 12, 2012

This 12% vs 18% is a very old argument indeed, ever since Ansel Adams was a lad in fact. And it is utterly redundant now that we have direct access to histograms in camera as well as a whole host of colour management tools from camera to print. There is no longer any need to worship at the feet of your chosen guru, just pick up the tools and do it for yourself.

-- hide signature --

Steve
www.pbase.com/steephill

Reply   Reply with quote   Complain
Deleted1929
Forum ProPosts: 13,050
Like?
Re: Pointless obsession
In reply to steephill, Nov 12, 2012

I think the very idea of exposing blind without reference to the scene is utterly pointless, like driving without reference to the road conditions - you're bound to come a cropper sooner rather than later.

-- hide signature --

StephenG

Reply   Reply with quote   Complain
gl2k
Veteran MemberPosts: 3,232
Like?
there is no such thing as proper exposure
In reply to Jack Hogan, Nov 12, 2012

It all depends on what you intend to do with the photo. In most cases (especially outdoors) the dynamic range of the scene by far exceeds the DR of the camera. So you have to decide what to capture at best quality.

On the other hand in a highly controlled environment like a photo studio using an exposure meter works great and gives perfect results for skin tones.

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
Re: Equal confusion, or not
In reply to aardvark7, Nov 12, 2012

aardvark7 wrote:  For my part, I appreciate his opinion about the influence of the print industy

I understand that Middle Gray in the printing industry is 18%, but I would be interested to hear your first hand thoughts.

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
Re: Proper Blind Exposure with a Gray Card
In reply to nikkorwatcher, Nov 12, 2012

Thanks for your comments. I normally really like the bloke, his other stuff is excellent.

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
steephill, sjgcit, gl2k
In reply to Jack Hogan, Nov 12, 2012

I agree with you.  But indulge me: taken the definition of 'proper blind' exposure for garnted, what are your thoughts on the other concepts?

Reply   Reply with quote   Complain
GMack
Senior MemberPosts: 2,928
Like?
Re: Proper Blind Exposure with a Gray Card
In reply to Jack Hogan, Nov 12, 2012

I've gone back to the old incident light-meter for exposure (Minolta IVF I think?).  Personally, I've thought about going to a "Red card with texture" since that is the one I generally blow off the histogram (Funny why the red pixel is always the one that seems reported going bad in the sensor too. It is very bright when it goes bad.).  I don't know the reflectance percentage yet for the red fabric.  I suspect "color" is also influencing the camera meters to some extent.  Seems to affect the AF tuning as well since my red flash assist light seems to cause mine to go off in AF tuning.

Fwiw, the old Kyoritsu calibration machines the techs used in the industry to calibrate meters used to be a warm orange light (3,200-3,800K).  Spectron in Colorado came along with theirs which used a blue filter (5,500-6,000K) and it became a topic then.  Spectron won out with their daylight balanced filter over the warm Kyoritsu machines and Hass'y favored them as did Leica.  No doubt some of the old CDS cells (and some were filtered with greenish/blue gels) favored one over the other.  The Japanese were never fond of the Colorado "daylight-balanced" design, even though they adopted it, and stayed with the re-worked Kyoritsu machine and I think Spectron is long gone now too.

No cards were ever used on those devices.  All they did was to take some slide film and use their reference camera to determine the brightness of the calibration window with a known shutter speed and a T-stoped reference lens. "Everything" was then matched to that window with a densitometer to set the correct red channel reading off the film.  A diaphragm was used to control the brightness of the light out of the window, with about a 5 stop range if I recall.  Color negative film always required the owner to compensate for by overexposing about 1 stop to hit the red channel in the densitometer off the gray card placed in the scene.  Pros knew to set their ASA/ISO to half what Kodak claimed it was for negative film, but exact for slide film.  The math behind the ASA/ISO never seemed to follow what happened in real life.  Kodak's engineers always claimed their ASA/ISO values were mathematically correct, but they didn't work out that way in real life.  Funny how their control strips exposure could not be duplicated using the "theoretical exposure values" vs. some general over-exposure. Ansel Adams (Who I met so that tells you my age!) and Kodak went around and around on it and even he over-exposed his film at times as he didn't trust their claimed values either.  Some of that distrust led to the Zone System.

Mack

Reply   Reply with quote   Complain
JDinOregon
New MemberPosts: 8
Like?
Re: steephill, sjgcit, gl2k
In reply to Jack Hogan, Nov 12, 2012

By "Blind Exposure" I take the situation as similar to doing a portrait without the benifit of an incident light meter. The photographer does not want to guess at the reflectance of the subjects skin (although I remember something like one stop brighter than middle gray as typical for caucasions). So instead an 18% gray card is used as a standard.

The meter constant for a Seconic Incident meter is 250 and the constant for Nikon and Canon reflected meters is 12.5. With exposure defined by

EV= log2(L*s/k) = log2(E*s/c)

This implies reflectance from a uniform perfect diffuser is 15.7%. From

L=E*R/pi

I think this means that the gray card will indicate an exposure about 0.2 stops off that of the incident meter.  Not so much for the reason that 18% is the wrong reflectance, but rather that an average scene is not a flat plane. (L not equal to E*R/pi).

It would be interesting to know what the model of an average scene was and also how that is interpreted by matrix metering in the different camera brands.

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
Thanks Mack
In reply to GMack, Nov 12, 2012

That's very helpful, and I understand that there were many variables in the old days in part because of the very different shapes of characteristic curves of many types of film available.  Today though DSLR sensors characteristic curves all look like parallel straight lines because modern sensors are linear and therefore all pretty well vary 1:1 with Exposure.  The only difference in this context is where their saturation exposure falls (Hsat - see my image above).

So the only variable left to determine is this: if I spot meter and take the picture of a neutral card where along the scale of raw values, from zero to maximum, should the camera record it?

Please note that the question of card reflectance hasn't come up.

The standards (ISO 2240, 2720, 2721, 12232:2006 and ANSI/NAPM IT3.302-1994) give the camera manufacturers lots of latitude to decide, but include a recommendation for DSLR sensors: recording the spot metered neutral card at 1/7.8 of the maximum raw value.  That's 12.8% of full scale.

Please note that the question of card reflectance hasn't come up.

Cheers,
Jack

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
Re: steephill, sjgcit, gl2k
In reply to JDinOregon, Nov 12, 2012

JDinOregon wrote:

By "Blind Exposure" I take the situation as similar to doing a portrait without the benifit of an incident light meter. The photographer does not want to guess at the reflectance of the subjects skin (although I remember something like one stop brighter than middle gray as typical for caucasions). So instead an 18% gray card is used as a standard.

Yes!

The meter constant for a Seconic Incident meter is 250 and the constant for Nikon and Canon reflected meters is 12.5. With exposure defined by

EV= log2(L*s/k) = log2(E*s/c)

This implies reflectance from a uniform perfect diffuser is 15.7%. From

L=E*R/pi

I think this means that the gray card will indicate an exposure about 0.2 stops off that of the incident meter. Not so much for the reason that 18% is the wrong reflectance, but rather that an average scene is not a flat plane. (L not equal to E*R/pi).

Hmm, interesting. But aside from different readings from different types of meters, aren't calibrated reflected light meters fully characterized by the constant K? And wouldn't this mean that if one were presented with a neutral card, irrespective of reflectance, it would produce values for aperture and shutter speed that when dialled into a DSLR at the same ISO should result in the following Exposure on the sensor

Hm = pi/4 * K/S

which with CaNikon's K=12.5 we could rewrite as about

Hm = 10 / S lux-seconds

which comfortingly is the same equation found in ISO 2240 and 2721 for the automatic controls of exposure in electronic imaging systems?

And that if the neutral card presented above were Middle Gray (18%, as calibrated) then the rest of the scene would be blind exposed 'properly'?

We know Hsat, because DxO handily measures it for us, so that leaves us with the pesky question of S ...

Jack

PS

It would be interesting to know what the model of an average scene was and also how that is interpreted by matrix metering in the different camera brands.

The following is not about matrix metering (a different beast altogether) but it may give you an idea:

Minolta IV users manual, page 3: " Standard reference subjects may have a reflection factor somewhere between 12% and 26%. Minolta bases its readings on a reference subject with an 18% reflection factor."

Sekonic Studio Deluxe L-398, page 2: " it is designed around a standard reflectivity of 18%, which has been derived through measurements of various subjects throughout the seasons of the year"

Reply   Reply with quote   Complain
panos_m
Senior MemberPosts: 1,212
Like?
Re: Proper Blind Exposure with a Gray Card
In reply to Jack Hogan, Nov 12, 2012

Hmm! For a camera meter to be calibrated at 18% it must read 2.47 stops lower than saturation for the green channel. I meter my D3 to 3.3 stops and my GF1 to 3 stops. So they are not calibrated to 18% :).

Jack Hogan wrote:

I would be interested in your opinion with regards to the following concepts on metering ‘blind’ with a gray card:

  1. that ‘blind’ exposure is defined in ths context as the exposure to be chosen with the sole help of a spot meter, without knowledge of the scene, time or place – it is the input quantity to the camera sensing medium
  2. that ‘proper’ blind exposure is defined in this context as the exposure that allows a camera to best record on its sensing medium a subset of the relative distribution of photons reflected from the scene; the full set is the relative distribution of photons that would be perceived by the human visual system in the same setting as the camera
  3. that brightness is an output medium dependent quantity typically not related linearly to exposure
  4. that for a given ISO the job of calibrated spot meters is to provide indications of shutter speed and aperture to achieve proper blind exposure
  5. that spot meters do not know or care what average scene reflectance or reference card is in front of them
  6. that in order to achieve proper blind exposure with most DSLRs it is best to use a standard Middle Gray card (that is one with 18% reflectance), which is what their spot meters are calibrated to
  7. that when metering off of and capturing such a target a DSLR will typically record digital raw values about 0.127 times the full scale reading of the sensor, subject to many provisos
  8. that if instead of a Middle Gray card a 12% reflectance card (or alternatively an 18% card with a ½ stop positive Exposure Correction) is used to set Exposure, the meter is fooled into raising Exposure by half a stop, reducing a DSLR’s designed highlight hard ceiling accordingly
  9. that if exposure is locked as per this last case, a Middle Gray tone will be recorded by a DSLR at raw values around 0.18 times the full scale of the sensor, so that a capture of an 18% Gray Card will record raw values of about 18% of Full Scale
  10. that other than a warm and fuzzy feeling there is no reason why those two percentages should be the same
  11. that we should all write an email to someone with whom I share a last name telling him that his blog entry on this subject is confusing a lot of peopleand does not stand up to his otherwise excellent standards

To help with the discussion I include this image:



Thanks for your input,
Jack

-- hide signature --

Panagiotis

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
The Crux of the Matter
In reply to panos_m, Nov 12, 2012

panos_m wrote:

Hmm! For a camera meter to be calibrated at 18% it must read 2.47 stops lower than saturation for the green channel. I meter my D3 to 3.3 stops and my GF1 to 3 stops. So they are not calibrated to 18%

Thank you Panos! And can you tell us what the reference to which a meter is calibrated has to do with what raw value a Middle Gray tone gets recorded at?

Calibration of a reflected light meter is all about making sure that when the meter thinks it is seeing 100% diffuse light it actually is. It doesn't matter what the standard used to ensure that that is indeed the case. Let's say for the sake of argument that camera meters are calibrated to a 75% reflectance standard. That means that to verify that the meter is reading correctly at the factory they present it with a 75% reflectance standard and it thinks that's 75%. So that when we present it in the field with a middle gray card, it reads it accurately at 18%.

Now you spot meter and take a picture with default settings of your D3 and GF1 of a sheet of paper reflecting 60% of the available light. They record it at -3.3 and 3 stops of saturation, that is at about raw values 1663 and 512, or 10.2% and 12.5% of full scale respectively.

What do 75%, 60% and 10.2% or 12.5% have to do with each other? And would you say in this case that you would need to use a 12% gray card to get the correct exposure?

Jack

PS Nein.  With your cameras you would need a Middle Gray card to get 'proper' 'blind' exposure (pull it Thom!)

Reply   Reply with quote   Complain
panos_m
Senior MemberPosts: 1,212
Like?
Re: The Crux of the Matter
In reply to Jack Hogan, Nov 12, 2012

I believe that is very simple. No matter what reflectance you put in front of it (10%, 60%, 75%, 100%) it will always read it the same that is it will always suggest the exposure that will produce the same "middle gray" raw value (the same distance in stops from saturation). What else do you need to know to make an exposure decision and place the tonal range of the scene where you want?

Jack Hogan wrote:

panos_m wrote:

Hmm! For a camera meter to be calibrated at 18% it must read 2.47 stops lower than saturation for the green channel. I meter my D3 to 3.3 stops and my GF1 to 3 stops. So they are not calibrated to 18%

Thank you Panos! And can you tell us what the reference to which a meter is calibrated has to do with what raw value a Middle Gray tone gets recorded at?

Calibration of a reflected light meter is all about making sure that when the meter thinks it is seeing 100% diffuse light it actually is. It doesn't matter what the standard used to ensure that that is indeed the case. Let's say for the sake of argument that camera meters are calibrated to a 75% reflectance standard. That means that to verify that the meter is reading correctly at the factory they present it with a 75% reflectance standard and it thinks that's 75%. So that when we present it in the field with a middle gray card, it reads it accurately at 18%.

Now you spot meter and take a picture with default settings of your D3 and GF1 of a sheet of paper reflecting 60% of the available light. They record it at -3.3 and 3 stops of saturation, that is at about raw values 1663 and 512, or 10.2% and 12.5% of full scale respectively.

What do 75%, 60% and 10.2% or 12.5% have to do with each other? And would you say in this case that you would need to use a 12% gray card to get the correct exposure?

Jack

PS Nein. With your cameras you would need a Middle Gray card to get 'proper' 'blind' exposure (pull it Thom!)

-- hide signature --

Panagiotis

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
For Clarity
In reply to Jack Hogan, Nov 12, 2012

Jack Hogan wrote:  So that when we present it in the field with a middle gray card, it reads it accurately at 18%.

For clarity, the sentence above should read:

So that when we present it in the field with a middle gray card, it reads it as 18% and gives us accurate values for Aperture and Shutter Speed in this light at the given ISO.

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
Re: The Crux of the Matter
In reply to panos_m, Nov 12, 2012

panos_m wrote:

I believe that is very simple. No matter what reflectance you put in front of it (10%, 60%, 75%, 100%) it will always read it the same that is it will always suggest the exposure that will produce the same "middle gray" raw value (the same distance in stops from saturation).

Agreed, sort of.  A calibrated meter will always produce the same exposure

Hm = 10 / S lux-seconds

What else do you need to know to make an exposure decision and place the tonal range of the scene where you want?

What percent reflectance card  would you need so that when spot metered it will give you 'proper' 'blind' exposure?

Reply   Reply with quote   Complain
panos_m
Senior MemberPosts: 1,212
Like?
Re: The Crux of the Matter
In reply to Jack Hogan, Nov 12, 2012

Jack Hogan wrote:

panos_m wrote:

I believe that is very simple. No matter what reflectance you put in front of it (10%, 60%, 75%, 100%) it will always read it the same that is it will always suggest the exposure that will produce the same "middle gray" raw value (the same distance in stops from saturation).

Agreed, sort of. A calibrated meter will always produce the same exposure

Hm = 10 / S lux-seconds

What else do you need to know to make an exposure decision and place the tonal range of the scene where you want?

What percent reflectance card would you need so that when spot metered it will give you 'proper' 'blind' exposure?

If "proper blind" exposure is that I obey the spot meter reading with the goal to reproduce the gray of the reflectance card then I need a card with the same reflectance the particular camera spot meter reads. That is different for my two cameras and obviously not 18%.

If I can deviate from the meter suggestion I can use a 18% card but I will have to add exposure in order to render it (the card) at 18%.

-- hide signature --

Panagiotis

Reply   Reply with quote   Complain
Jack Hogan
Senior MemberPosts: 3,823Gear list
Like?
Re: The Crux of the Matter
In reply to panos_m, Nov 12, 2012

panos_m wrote:

If "proper blind" exposure is that I obey the spot meter reading with the goal to reproduce the gray of the reflectance card then I need a card with the same reflectance the particular camera spot meter reads. That is different for my two cameras and obviously not 18%.

Let's assume that your camera manufacturers, knowing their sensors well, chose a good compromise for where a mid-tone should be recorded in the range of possible raw values (10.2% of full scale for the D3 and 12.5% for the GF1).

No matter what reflectance card is placed in front of their meters, each of them will always record them as the designed mid-tone.  So wouldn't you want to meter off a middle gray card to ensure that it be recorded as a mid-tone, therefore ensuring that the rest of the scene is also consequently  'properly' exposed?

And why on earth would anybody (anybody!) want to use a 12% card in this context?

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads