Sunny 16 rule?

Roland Karlsson wrote:
Alan Browne wrote in news:WfsYc.15432
:


In any case, if R,G,B are each in the range 110 - 130 for a
properly exposed shot of a grey card, then it is likely the
metering of the camera is correct.



Yepp - but that does not tell you that the ISO
number is correct. Maybe it is ISO 50 instead
of (claimed) ISO 100 and the time is actually
1/50 and not the (claimed) 1/100. That would
give the same exposure.


Let us be clear: if the incident meter says, eg, 1/100 f/5.6
for ISO 100, the three variable for exposure have been isolated.
Set those three on the camera, shoot the grey card.

From there the grey card should behave as discussed. From the
web I get close to R=G=B=115 or R=G=B=128 (and suspect the later
is correct), so if the TIFF file gives R=G=B=128(ish) for the
incident setting set onto the camera, then the camera exposure
system is bahaving as it should per the ISO film spec.

Cheers,
Alan

--
-- rec.photo.equipment.35mm user resource:
-- http://www.aliasimages.com/rpe35mmur.htm
-- e-meil: there's no such thing as a FreeLunch.--

Roland Karlsson wrote:
Alan Browne wrote in news:WfsYc.15432
:


In any case, if R,G,B are each in the range 110 - 130 for a
properly exposed shot of a grey card, then it is likely the
metering of the camera is correct.



Yepp - but that does not tell you that the ISO
number is correct. Maybe it is ISO 50 instead
of (claimed) ISO 100 and the time is actually
1/50 and not the (claimed) 1/100. That would
give the same exposure.


Let us be clear: if the incident meter says, eg, 1/100 f/5.6
for ISO 100, the three variable for exposure have been isolated.
Set those three on the camera, shoot the grey card.

From there the grey card should behave as discussed. From the
web I get close to R=G=B=115 or R=G=B=128 (and suspect the later
is correct), so if the TIFF file gives R=G=B=128(ish) for the
incident setting set onto the camera, then the camera exposure
system is bahaving as it should per the ISO film spec.

Cheers,
Alan

--
-- rec.photo.equipment.35mm user resource:
-- http://www.aliasimages.com/rpe35mmur.htm
-- e-meil: there's no such thing as a FreeLunch.--

Roland Karlsson wrote:
John McWilliams wrote in news:bBHYc.86598$Fg5.55978
@attbi_s53:


Interesting thread, but under what circumstances - today - would one
want or need the sunny 16? Back when meters were atrocious or it was
hugely expensive to buy a good one, understood. Back when one might have
forgotten one's light meter and wanted an approximation, understood.


It is useful - to understand the nature of exposure.

Agreed. Useful, but hardly a required tool.


This gives you a deeper understanding. The world consists
of reflecting objects, from black to white. Nothing is
really blacker than say 1% and nothing is whiter than 100%.
So - you can use exactly the same exposure for all subjects,
totally independent of the subjects actual reflectance.

The problem arises when you have shadows and reflecting
objects of course.


If it helps one understand, but it hardly seems a favorable subsitute
for paying attention to what you're metering *for* in a sunny scene.
There are several -maybe dozens- of ways to arrive at the best exposure
for what *you* want to capture in a single shot. And the review feature
and histogram at least show if you're in the ball park.

--
John McWilliams

Roland Karlsson wrote:
John McWilliams wrote in news:bBHYc.86598$Fg5.55978
@attbi_s53:


Interesting thread, but under what circumstances - today - would one
want or need the sunny 16? Back when meters were atrocious or it was
hugely expensive to buy a good one, understood. Back when one might have
forgotten one's light meter and wanted an approximation, understood.


It is useful - to understand the nature of exposure.

Agreed. Useful, but hardly a required tool.


This gives you a deeper understanding. The world consists
of reflecting objects, from black to white. Nothing is
really blacker than say 1% and nothing is whiter than 100%.
So - you can use exactly the same exposure for all subjects,
totally independent of the subjects actual reflectance.

The problem arises when you have shadows and reflecting
objects of course.


If it helps one understand, but it hardly seems a favorable subsitute
for paying attention to what you're metering *for* in a sunny scene.
There are several -maybe dozens- of ways to arrive at the best exposure
for what *you* want to capture in a single shot. And the review feature
and histogram at least show if you're in the ball park.

--
John McWilliams

Roland Karlsson wrote:



Depends on what you define by correct. If you compare the
Sunny F16 to an incident meeter you will be surprised how
near it is on a clear day. An incident light meeter assumes
that the subject contains some almost white parts.

An incident meter cares not a whit for the subject and assumes
absolutely nothing. It measures only light falling towards the
subject, not reflected light from the subject (in camera or spot
meter). [Note that for this simple measurement, the dome is
pointed at the camera lens from about the position of the subject
.... so the subject has no effect on an _incident_ reading.]


If you use a spot meter and take a photo of a dark tree trunk,
then it contains no white parts and it can be exposed much
more than the Sunny F16 rule.

Better put: The dark tree trunk falls about 2 stops below 18%
grey. The spot meter expects it to be grey and so it opens up to
make it grey. If the tree were all white, the reverse would be
the case and underexposure would result.

But if I meter the light falling on a scene with an _incident_
meter, I will get the correct exposure regardless of the
color/reflectance of the subjects.

Poor man's incident meter: Place a clean, white (dry) styrofoam
cofee cup over the lens. Stand where the subject is. Point it
at where you will take the photograph from. Depress halfway...
read the f/ and S... it is the correct incident metered light for
the shot. (This I've found to be accurate to w/i 1/3 of a stop
by comparison to an incident light meter).

Cheers,
Alan.


--
-- rec.photo.equipment.35mm user resource:
-- http://www.aliasimages.com/rpe35mmur.htm
-- e-meil: there's no such thing as a FreeLunch.--

Roland Karlsson wrote:



Depends on what you define by correct. If you compare the
Sunny F16 to an incident meeter you will be surprised how
near it is on a clear day. An incident light meeter assumes
that the subject contains some almost white parts.

An incident meter cares not a whit for the subject and assumes
absolutely nothing. It measures only light falling towards the
subject, not reflected light from the subject (in camera or spot
meter). [Note that for this simple measurement, the dome is
pointed at the camera lens from about the position of the subject
.... so the subject has no effect on an _incident_ reading.]


If you use a spot meter and take a photo of a dark tree trunk,
then it contains no white parts and it can be exposed much
more than the Sunny F16 rule.

Better put: The dark tree trunk falls about 2 stops below 18%
grey. The spot meter expects it to be grey and so it opens up to
make it grey. If the tree were all white, the reverse would be
the case and underexposure would result.

But if I meter the light falling on a scene with an _incident_
meter, I will get the correct exposure regardless of the
color/reflectance of the subjects.

Poor man's incident meter: Place a clean, white (dry) styrofoam
cofee cup over the lens. Stand where the subject is. Point it
at where you will take the photograph from. Depress halfway...
read the f/ and S... it is the correct incident metered light for
the shot. (This I've found to be accurate to w/i 1/3 of a stop
by comparison to an incident light meter).

Cheers,
Alan.


--
-- rec.photo.equipment.35mm user resource:
-- http://www.aliasimages.com/rpe35mmur.htm
-- e-meil: there's no such thing as a FreeLunch.--

On 30 Aug 2004 15:46:09 -0700, wrote:

jpc wrote:

the noise levels in the highlights can be significantly different
than the noise in the shadows.

Why wouldn't they?

That's the interesting part. Why do I see variations in noise that
often go in the wrong direction from what I expect from theory.

I do some very low light photography and when I want to check noise on
a camera independently of the numbers posted on dpreview, I download
the Dave Box image from the imaging-resource web site. Then I run a
line profile on the step guage part of the image. I don't remember
the details on all the camera's I looked at but with the last one--the
oly 8080--there was a lot less noise at the bright end of the step
guage, where you don't need it, and more noise at the dark end where
you do.

I have no control over the how the Dave Box images are taken, but
from what I read about the proceedures on the website the images
seemed to have been taken in a consistant manner over the years. So
I've been assuming the differences are caused by the noise reduction
routines in the cameras' firmware.



There is a read-out noise from the electronics,
and there is the photon shot noise from the act of exposing the sensor
to light. The former is constant, the latter's variance scales
linearly with exposu

noise = sqrt(read_out + exposure*signal)

is a reasonable approximation. Indeed, if you measure noise levels
vs. signal (linear), you can fit reasonably straight lines to the
square (variance) of the noise as a function of exposure.


Are we saying the same thing? To the first appoximation the nosie
floor of the sensor is constant. To this we add the photo noise which
varies a the square of the number of photo electrons captured in the
sensor well. So if I have a noise floor of 10 photoelectons and a
sensor well holding 10000 electrons I should see a noise level of
about 110 photoelectrons. At the other extreme, if I run a short
exposure dark frame so I can ignore thermal noise I should only see
the noise floor of 10 photoelectrons.

In other word when I run a line profile on a step guage, I should see
more noise at the bright end and less noise at the dark end rather
than vice versa.

So if anyone has an explanation for this, enlighten me. I must confess
I'm confused as to what is going on

jpc

On 30 Aug 2004 15:46:09 -0700, wrote:

jpc wrote:

the noise levels in the highlights can be significantly different
than the noise in the shadows.

Why wouldn't they?

That's the interesting part. Why do I see variations in noise that
often go in the wrong direction from what I expect from theory.

I do some very low light photography and when I want to check noise on
a camera independently of the numbers posted on dpreview, I download
the Dave Box image from the imaging-resource web site. Then I run a
line profile on the step guage part of the image. I don't remember
the details on all the camera's I looked at but with the last one--the
oly 8080--there was a lot less noise at the bright end of the step
guage, where you don't need it, and more noise at the dark end where
you do.

I have no control over the how the Dave Box images are taken, but
from what I read about the proceedures on the website the images
seemed to have been taken in a consistant manner over the years. So
I've been assuming the differences are caused by the noise reduction
routines in the cameras' firmware.



There is a read-out noise from the electronics,
and there is the photon shot noise from the act of exposing the sensor
to light. The former is constant, the latter's variance scales
linearly with exposu

noise = sqrt(read_out + exposure*signal)

is a reasonable approximation. Indeed, if you measure noise levels
vs. signal (linear), you can fit reasonably straight lines to the
square (variance) of the noise as a function of exposure.


Are we saying the same thing? To the first appoximation the nosie
floor of the sensor is constant. To this we add the photo noise which
varies a the square of the number of photo electrons captured in the
sensor well. So if I have a noise floor of 10 photoelectons and a
sensor well holding 10000 electrons I should see a noise level of
about 110 photoelectrons. At the other extreme, if I run a short
exposure dark frame so I can ignore thermal noise I should only see
the noise floor of 10 photoelectrons.

In other word when I run a line profile on a step guage, I should see
more noise at the bright end and less noise at the dark end rather
than vice versa.

So if anyone has an explanation for this, enlighten me. I must confess
I'm confused as to what is going on

jpc

"Alan Browne" wrote:
Roland Karlsson wrote:
Depends on what you define by correct. If you compare the
Sunny F16 to an incident meeter you will be surprised how
near it is on a clear day. An incident light meeter assumes
that the subject contains some almost white parts.

An incident meter cares not a whit for the subject and assumes
absolutely nothing. It measures only light falling towards the
subject, not reflected light from the subject (in camera or spot
meter). [Note that for this simple measurement, the dome is
pointed at the camera lens from about the position of the subject
... so the subject has no effect on an _incident_ reading.]


If you use a spot meter and take a photo of a dark tree trunk,
then it contains no white parts and it can be exposed much
more than the Sunny F16 rule.

Better put: The dark tree trunk falls about 2 stops below 18%
grey. The spot meter expects it to be grey and so it opens up to
make it grey.

That's not how a spot meter is used. Unless one hasn't a clue.

If the tree were all white, the reverse would be
the case and underexposure would result.

The spot meter gives a correct reading of the brightness of the subject. The
photographer decides what zone to place that subject on and adjusts the EV
accordingly. It requires thinking about how you want your subject exposed,
it's not automagic.

You _can_ use an in-camera spotmeter in automagic mode. Dial in the required
EV compensation for the subject, and then lock exposure on the subject,
recompose, and shoot. Perfect exposure every time regardless of backlighting
or light sources in the frame that would completely mess up averaging,
matrix, evaluative.

But if I meter the light falling on a scene with an _incident_
meter, I will get the correct exposure regardless of the
color/reflectance of the subjects.

No, you'll get the correct exposure if you are shooting catalog shots in a
studio. If you want your dawn or dusk shots to look like dawn or dusk, you
can't do it with an incident meter. If you want your high-noon shots to look
like high noon, you can't do it with an incident meter.

David J. Littleboy
Tokyo, Japan

"Alan Browne" wrote:
Roland Karlsson wrote:
Depends on what you define by correct. If you compare the
Sunny F16 to an incident meeter you will be surprised how
near it is on a clear day. An incident light meeter assumes
that the subject contains some almost white parts.

An incident meter cares not a whit for the subject and assumes
absolutely nothing. It measures only light falling towards the
subject, not reflected light from the subject (in camera or spot
meter). [Note that for this simple measurement, the dome is
pointed at the camera lens from about the position of the subject
... so the subject has no effect on an _incident_ reading.]


If you use a spot meter and take a photo of a dark tree trunk,
then it contains no white parts and it can be exposed much
more than the Sunny F16 rule.

Better put: The dark tree trunk falls about 2 stops below 18%
grey. The spot meter expects it to be grey and so it opens up to
make it grey.

That's not how a spot meter is used. Unless one hasn't a clue.

If the tree were all white, the reverse would be
the case and underexposure would result.

The spot meter gives a correct reading of the brightness of the subject. The
photographer decides what zone to place that subject on and adjusts the EV
accordingly. It requires thinking about how you want your subject exposed,
it's not automagic.

You _can_ use an in-camera spotmeter in automagic mode. Dial in the required
EV compensation for the subject, and then lock exposure on the subject,
recompose, and shoot. Perfect exposure every time regardless of backlighting
or light sources in the frame that would completely mess up averaging,
matrix, evaluative.

But if I meter the light falling on a scene with an _incident_
meter, I will get the correct exposure regardless of the
color/reflectance of the subjects.

No, you'll get the correct exposure if you are shooting catalog shots in a
studio. If you want your dawn or dusk shots to look like dawn or dusk, you
can't do it with an incident meter. If you want your high-noon shots to look
like high noon, you can't do it with an incident meter.

David J. Littleboy
Tokyo, Japan