Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computing > Digital Photography > custom white balance/exposure questions ?

Reply
Thread Tools

custom white balance/exposure questions ?

 
 
John McWilliams
Guest
Posts: n/a
 
      07-12-2006
http://www.velocityreviews.com/forums/(E-Mail Removed) wrote:
> Thank you for not dismissing my opinion out of hand. Recently I
> photographed a bright orange flower on a high key background using
> tungsten halogen. The first three exposures I adjusted only the
> aperture (because this was a very deep flower). Then I realized I had
> forgot to do a custom wb. So I did the custom wb then re-shot with the
> same settings. In each case i kept the information about a third stop
> from the right side of the histogram. After examining each image in
> post I had dramatic clipping on one channel all three precustom wb and
> no clipping after custom wb. Admittedly this was a test with a
> dominant color but I have had similar results with a turtle in tall
> grass and other subjects. Does this seem like a valid test? My guess
> is that the custom wb only has an influence on the way the histogram is
> displayed but your scenarios are definite possibilities. I would like
> to know what is going on.


I've always tried to have the WB set correctly when shooting RAW, and
your tests seem to indicate that there's a benefit to it beyond having
fewer adjustments to make in post.

I am pretty sure there are others here who have, or will, test this
rigourously, and look forward to more on this.

--
john mcwilliams
 
Reply With Quote
 
 
 
 
Bart van der Wolf
Guest
Posts: n/a
 
      07-12-2006

<(E-Mail Removed)> wrote in message
news:(E-Mail Removed) oups.com...
SNIP
> My guess is that the custom wb only has an influence on the way
> the histogram is displayed but your scenarios are definite
> possibilities. I would like to know what is going on.


It depends on the camera's internal histogram generation code. It
often is built from either Luminocity (= weighted R+G+B), or just the
Green channel response (= faster). By doing a Custom WB you will most
likely also change exposure to avoid clipping, and that exposure
change will be reflected in the histogram population.

Bart

 
Reply With Quote
 
 
 
 
Ben Brugman
Guest
Posts: n/a
 
      07-13-2006

<(E-Mail Removed)> schreef in bericht
news:(E-Mail Removed) oups.com...
> aperture (because this was a very deep flower). Then I realized I had
> forgot to do a custom wb. So I did the custom wb then re-shot with the
> same settings. In each case i kept the information about a third stop
> from the right side of the histogram. After examining each image in


How do you keep 1/3 of a stop from the right side of the histogram ?
How are the 'stops' positioned on the histogram ?

> post I had dramatic clipping on one channel all three precustom wb and
> no clipping after custom wb. Admittedly this was a test with a
> dominant color but I have had similar results with a turtle in tall


The test with a dominant color in tungsten ligh is very suetable for
examination of influence of the WB on raw. Why ? Because it is
extreme and changes are more visible.


Test 1.
Does the WB setting influence the exposure? For all set ups the
situation should be as constant as possible. So use your
'extreme' color scene. But have the frame fixed,
(tripod, fixed zoom distance, if possible prime lens).
Make testshots, with no alterations. See if the exposure
does reproduce.

Make testshots with WB alterations. Try to go from
one extreem to the other extreem. See if the exposere
alters more between the different WB settings than
between the same white balance settings.

Conclusions we hope to draw from test 1.
WB setting does influence Exposure. *)
WB setting does not influence Exposure. (<-- My original guesstimate) **)


Test 2.
Does the WB setting influence the Raw image? Same set up
as test 1. But with Manual setting of time and aperature so
that there is no exposure difference in the shots.
Now make the shots with different WB settings.

Conclusions we hope to draw from test 2.
WB setting does influence the RAW. *)
WB setting does not influence the RAW. (<-- My original guesstimate) ***)

*) One of the two is your indication.

**) To my knowledge (but am not sure) the exposure system of the D300
is not color orientated. The Canon 350D has a 35 field evaluation exposure,
there is no mention of color.
(The Nikon D70(s) does have a color orientated exposure system).

***) To my knowledge settings do not influence the RAW picture. Settings
are used for the embedded jpg and the settings are registered in the
'header' information. If the setting would influence the RAW, how should
third party RAW programs process a picture, without the knowledge of
what settings have allready been processed.

ben


 
Reply With Quote
 
acl
Guest
Posts: n/a
 
      07-13-2006

Ben,

I had a similar argument with someone. To convince him, I eventually
did take several raw shots with different WB settings (and the camera
on a tripod, with constant exposure) and converted with the same colour
temperature and tint in the converter. They were the same. So, at least
for my D200, WB does not affect the raw data (actually it doesn't
affect the raw files of most cameras, it seems, with the exception
perhaps of the D2X).

It is also trivial to check that the WB does not affect the exposure
either (even with lightmeters which take into account colour such as
the Nikons).

Cheers.

 
Reply With Quote
 
ronviers@gmail.com
Guest
Posts: n/a
 
      07-13-2006
Okay here are my test settings:
Before custom wb

Manuel exposure
Shutter speed = 1/6
Aperture = 22
Fixed 50
WB = auto
Focus = manual

Then I did a custom wb and reshot with these settings:

Manuel exposure
Shutter speed = 1/6
Aperture = 22
Fixed 50
WB = custom
Focus = manual

Camera was tripod mounted and lighting (tunsten halogen) was
unchanged. And like I said the three channed histogram was
dramatically different for each picture and showed a lot of clipping on
the red channel before custom wbing. Seems fair to me but I know how
easy it is to miss something and how hard it is to get control.
I used this test rather than the "keep the info 1/3 stops from the
right side of histogram" because it so clearly show the impact of
custom wbing on histogram and I think that is what needs do be
demonstrated to support my original position.



Ben Brugman wrote:
> <(E-Mail Removed)> schreef in bericht
> news:(E-Mail Removed) oups.com...
> > aperture (because this was a very deep flower). Then I realized I had
> > forgot to do a custom wb. So I did the custom wb then re-shot with the
> > same settings. In each case i kept the information about a third stop
> > from the right side of the histogram. After examining each image in

>
> How do you keep 1/3 of a stop from the right side of the histogram ?
> How are the 'stops' positioned on the histogram ?
>
> > post I had dramatic clipping on one channel all three precustom wb and
> > no clipping after custom wb. Admittedly this was a test with a
> > dominant color but I have had similar results with a turtle in tall

>
> The test with a dominant color in tungsten ligh is very suetable for
> examination of influence of the WB on raw. Why ? Because it is
> extreme and changes are more visible.
>
>
> Test 1.
> Does the WB setting influence the exposure? For all set ups the
> situation should be as constant as possible. So use your
> 'extreme' color scene. But have the frame fixed,
> (tripod, fixed zoom distance, if possible prime lens).
> Make testshots, with no alterations. See if the exposure
> does reproduce.
>
> Make testshots with WB alterations. Try to go from
> one extreem to the other extreem. See if the exposere
> alters more between the different WB settings than
> between the same white balance settings.
>
> Conclusions we hope to draw from test 1.
> WB setting does influence Exposure. *)
> WB setting does not influence Exposure. (<-- My original guesstimate) **)
>
>
> Test 2.
> Does the WB setting influence the Raw image? Same set up
> as test 1. But with Manual setting of time and aperature so
> that there is no exposure difference in the shots.
> Now make the shots with different WB settings.
>
> Conclusions we hope to draw from test 2.
> WB setting does influence the RAW. *)
> WB setting does not influence the RAW. (<-- My original guesstimate) ***)
>
> *) One of the two is your indication.
>
> **) To my knowledge (but am not sure) the exposure system of the D300
> is not color orientated. The Canon 350D has a 35 field evaluation exposure,
> there is no mention of color.
> (The Nikon D70(s) does have a color orientated exposure system).
>
> ***) To my knowledge settings do not influence the RAW picture. Settings
> are used for the embedded jpg and the settings are registered in the
> 'header' information. If the setting would influence the RAW, how should
> third party RAW programs process a picture, without the knowledge of
> what settings have allready been processed.
>
> ben


 
Reply With Quote
 
ronviers@gmail.com
Guest
Posts: n/a
 
      07-13-2006
What I was saying is that if effects the histogram. That is, the
histogram is more accurate using custom wb. I do not think it has an
impact on the raw data.
But now I would like to qualify my postition to - the histogram is more
accurate on the Canon 300d using custom wb - because that is the only
one I can test. The OP is using a 20d and it may not behave the same.
I would like to know.


acl wrote:
> Ben,
>
> I had a similar argument with someone. To convince him, I eventually
> did take several raw shots with different WB settings (and the camera
> on a tripod, with constant exposure) and converted with the same colour
> temperature and tint in the converter. They were the same. So, at least
> for my D200, WB does not affect the raw data (actually it doesn't
> affect the raw files of most cameras, it seems, with the exception
> perhaps of the D2X).
>
> It is also trivial to check that the WB does not affect the exposure
> either (even with lightmeters which take into account colour such as
> the Nikons).
>
> Cheers.


 
Reply With Quote
 
acl
Guest
Posts: n/a
 
      07-13-2006
(E-Mail Removed) wrote:
> What I was saying is that if effects the histogram. That is, the
> histogram is more accurate using custom wb. I do not think it has an
> impact on the raw data.
> But now I would like to qualify my postition to - the histogram is more
> accurate on the Canon 300d using custom wb - because that is the only
> one I can test. The OP is using a 20d and it may not behave the same.
> I would like to know.
>
>


Hello,
It depends on what you mean by "more accurate". If you correctly white
balance then the histogram should more closely correspond to the colour
distribution of the original scene. However, this is not what
necessarily what one wants for accurate exposure determination; rather,
it is the exposure for each channel in the raw data.

What I mean is this: Imagine you are shooting under light with little
blue and green (eg lamps). Now if you expose so that the red channel
nearly clips (ie in the raw data, the highest value for the red channel
is near its maximum). Obviously, the blue and green channels are
nowhere near clipping (in the raw data, I mean); their maximum values
will be well below the maximum possible.

When this raw data is converted into (say) a tiff file, and you choose
the correct white balance so that white objects are rendered white (and
not red), what happens is that the blue and green channels are
amplified more than the red channel (this is why images shot under such
light have more noise: one or two of the channels have to be amplified
more, and are basically underexposed). If you look at the histogram of
the resulting tiff file, it will be different than the histogram for
the raw data (had you been able to see such a thing), since each
channel has now been amplified by different amounts.

Anyway, the point I was trying to make is that the histogram you see is
made from a jpeg. To see this, try the following: Set the camera to
raw, and put it on a tripod. shoot something, then change various
settings for the jpegs, eg contrast, saturation, colour space etc;
shoot the scene again. Compare the histograms: They're different (if
you had separate histograms for R, G and B, it would have been even
more dramatic). Thus, the camera actually produces a jpeg file, using
with the settings you chose for jpeg files, and takes the histogram
from there. By the way, with a 3-channel histogram, it's very easy to
see the effect of the white balance.

Whether or not this is more accurate is a matter of definition (and
even after a clear definition of "accurate", it remains a matter of
taste). I'd prefer to have histograms for each channel of the raw data,
showing the distribution of values before any processing (it is
possible to approximate this in various Nikons by loading the proper
white balance setting, but the contrast, saturation etc settings still
do alter the histogram). I'd rather know when a channel is actually
clipping (rather than when the jpeg produced by the camera is actually
clipping, which is a different point).

Of course, if you make sure nothing clips in the histogram displayed by
the camera, then nothing clips in the raw data either.

OK, this rant is not getting clearer, I'd better stop here! I hope what
I wrote is comprehensible.

 
Reply With Quote
 
Ben Brugman
Guest
Posts: n/a
 
      07-13-2006
>
> OK, this rant is not getting clearer, I'd better stop here! I hope what
> I wrote is comprehensible.
>

No you did not.

>
> What I mean is this: Imagine you are shooting under light with little
> blue and green (eg lamps). Now if you expose so that the red channel
> nearly clips (ie in the raw data, the highest value for the red channel
> is near its maximum). Obviously, the blue and green channels are
> nowhere near clipping (in the raw data, I mean); their maximum values
> will be well below the maximum possible.


If using blue and green (eg lamps). And expose so that the red channel
nearly clips. Then the blue and green will be totaly overexposed.

>
> When this raw data is converted into (say) a tiff file, and you choose
> the correct white balance so that white objects are rendered white (and
> not red), what happens is that the blue and green channels are


If exposed correctly having only blue and green will never create a correct
wb for white objects. The blue and green do not need much amplification.
Red is totaly absend and should not be amplified. So there isn't a noise
problem except in the dark area's where also blue and green are absent.


> amplified more than the red channel (this is why images shot under such
> light have more noise: one or two of the channels have to be amplified
> more, and are basically underexposed). If you look at the histogram of
> the resulting tiff file, it will be different than the histogram for
> the raw data (had you been able to see such a thing), since each
> channel has now been amplified by different amounts.



ben



 
Reply With Quote
 
Ben Brugman
Guest
Posts: n/a
 
      07-13-2006

<(E-Mail Removed)> schreef in bericht
news:(E-Mail Removed) ps.com...
> Okay here are my test settings:
> Before custom wb
>
> Manuel exposure
> Shutter speed = 1/6
> Aperture = 22
> Fixed 50
> WB = auto
> Focus = manual
>
> Then I did a custom wb and reshot with these settings:
>
> Manuel exposure
> Shutter speed = 1/6
> Aperture = 22
> Fixed 50
> WB = custom
> Focus = manual
>
> Camera was tripod mounted and lighting (tunsten halogen) was
> unchanged. And like I said the three channed histogram was


Do you mean a three channel histogram. Does the D300 have
a three channel histogram ?

Or do you mean a histogram on your PC ?
I get the impression that you mean the histogram on the PC.
If a picture gets displayed on a PC some processing has to
be done, especially for raw. It does use a gamma curve,
it could use sharpening and it does use A white balance.
Having set a white balance on the 300 D a lot off software
will use that white balance to display the picture and also to
calculate the histogram (1 channel or three channel) from
the processed raw data.
So this method can not be used to judge if WB does influence
the raw information, because the channel data if from the
PROCESSED raw picture and this is definitly not the same.

My question how do you judge that you are 1/3 of a stop
of clipping ?

ben


> dramatically different for each picture and showed a lot of clipping on
> the red channel before custom wbing. Seems fair to me but I know how
> easy it is to miss something and how hard it is to get control.
> I used this test rather than the "keep the info 1/3 stops from the
> right side of histogram" because it so clearly show the impact of
> custom wbing on histogram and I think that is what needs do be
> demonstrated to support my original position.
>
>
>
> Ben Brugman wrote:
>> <(E-Mail Removed)> schreef in bericht
>> news:(E-Mail Removed) oups.com...
>> > aperture (because this was a very deep flower). Then I realized I had
>> > forgot to do a custom wb. So I did the custom wb then re-shot with the
>> > same settings. In each case i kept the information about a third stop
>> > from the right side of the histogram. After examining each image in

>>
>> How do you keep 1/3 of a stop from the right side of the histogram ?
>> How are the 'stops' positioned on the histogram ?
>>
>> > post I had dramatic clipping on one channel all three precustom wb and
>> > no clipping after custom wb. Admittedly this was a test with a
>> > dominant color but I have had similar results with a turtle in tall

>>
>> The test with a dominant color in tungsten ligh is very suetable for
>> examination of influence of the WB on raw. Why ? Because it is
>> extreme and changes are more visible.
>>
>>
>> Test 1.
>> Does the WB setting influence the exposure? For all set ups the
>> situation should be as constant as possible. So use your
>> 'extreme' color scene. But have the frame fixed,
>> (tripod, fixed zoom distance, if possible prime lens).
>> Make testshots, with no alterations. See if the exposure
>> does reproduce.
>>
>> Make testshots with WB alterations. Try to go from
>> one extreem to the other extreem. See if the exposere
>> alters more between the different WB settings than
>> between the same white balance settings.
>>
>> Conclusions we hope to draw from test 1.
>> WB setting does influence Exposure. *)
>> WB setting does not influence Exposure. (<-- My original guesstimate) **)
>>
>>
>> Test 2.
>> Does the WB setting influence the Raw image? Same set up
>> as test 1. But with Manual setting of time and aperature so
>> that there is no exposure difference in the shots.
>> Now make the shots with different WB settings.
>>
>> Conclusions we hope to draw from test 2.
>> WB setting does influence the RAW. *)
>> WB setting does not influence the RAW. (<-- My original guesstimate) ***)
>>
>> *) One of the two is your indication.
>>
>> **) To my knowledge (but am not sure) the exposure system of the D300
>> is not color orientated. The Canon 350D has a 35 field evaluation
>> exposure,
>> there is no mention of color.
>> (The Nikon D70(s) does have a color orientated exposure system).
>>
>> ***) To my knowledge settings do not influence the RAW picture. Settings
>> are used for the embedded jpg and the settings are registered in the
>> 'header' information. If the setting would influence the RAW, how should
>> third party RAW programs process a picture, without the knowledge of
>> what settings have allready been processed.
>>
>> ben

>



 
Reply With Quote
 
ronviers@gmail.com
Guest
Posts: n/a
 
      07-13-2006

> It depends on what you mean by "more accurate". If you correctly white
> balance then the histogram should more closely correspond to the colour
> distribution of the original scene. However, this is not what
> necessarily what one wants for accurate exposure determination; rather,
> it is the exposure for each channel in the raw data.


By more accurate I simply mean that you are less likely to get single
channel clipping if you do a custom wb.


> What I mean is this: Imagine you are shooting under light with little
> blue and green (eg lamps). Now if you expose so that the red channel
> nearly clips (ie in the raw data, the highest value for the red channel
> is near its maximum). Obviously, the blue and green channels are
> nowhere near clipping (in the raw data, I mean); their maximum values
> will be well below the maximum possible.


With a luminance only histogram the point is moot.

> Of course, if you make sure nothing clips in the histogram displayed by
> the camera, then nothing clips in the raw data either.


This is the exact opposite of what I was trying to say.

Brgds,
Ron

 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Any programs to trim white space/ remove all white space in HTML file? Ben C HTML 6 01-28-2007 11:41 PM
White font on white background???? Stubby Firefox 3 08-18-2006 04:25 PM
Custom White Balance: Gray Card or White Card? Steve Cutchen Digital Photography 31 10-28-2005 09:13 PM
Color.white vs. Color.WHITE Niels Dybdahl Java 3 10-06-2004 03:21 PM
Affects of a PL filter on white balance and white balance (D100) jeff liss Digital Photography 1 09-05-2003 02:07 PM



Advertisments