# color temp a defacto filter?

Discussion in 'Digital Photography' started by drs@canby.com, Feb 20, 2005.

1. ### Guest

This isn't limited to digital photography. How appropriate is it to
consider color temperature a filter? And to what extent is the missing
or added portions of the color spectrum significant? Even with proper
white balance are data lost (or added?) at 3000 K when compared to
8000 K?

, Feb 20, 2005

2. ### Joseph MeehanGuest

wrote:
> This isn't limited to digital photography. How appropriate is it to
> consider color temperature a filter? And to what extent is the missing
> or added portions of the color spectrum significant? Even with proper
> white balance are data lost (or added?) at 3000 K when compared to
> 8000 K?

The real answer is this. Try it yourself and see if it is significant
to you in the work you do. Other than that you can get thee other answers.
Someone else's evaluation of the issue based on their equipment, shooting
conditions and personal judgment, a technically correct answer with lots of
numbers which are meaningless in the real world or a pure guess.

--
Joseph Meehan

26 + 6 = 1 It's Irish Math

Joseph Meehan, Feb 20, 2005

3. ### C J CampbellGuest

<> wrote in message
news:...
> This isn't limited to digital photography. How appropriate is it to
> consider color temperature a filter? And to what extent is the missing
> or added portions of the color spectrum significant? Even with proper
> white balance are data lost (or added?) at 3000 K when compared to
> 8000 K?

White balance is not at all like a filter. A filter subtracts portions of
the spectrum before the light reaches your lens. White balance is an
interpretation of that spectrum. The camera is a machine. It records
whatever photons strike the imaging material. It has no idea what color
those photons are. Whenever you take a picture, all the light is recorded.
Nothing is subtracted in order to adjust white balance.

C J Campbell, Feb 21, 2005
4. ### Guest

On Sun, 20 Feb 2005 16:06:36 -0800, "C J Campbell"
<> wrote:

>
><> wrote in message
>news:...
>> This isn't limited to digital photography. How appropriate is it to
>> consider color temperature a filter? And to what extent is the missing
>> or added portions of the color spectrum significant? Even with proper
>> white balance are data lost (or added?) at 3000 K when compared to
>> 8000 K?

>
>White balance is not at all like a filter. A filter subtracts portions of
>the spectrum before the light reaches your lens. White balance is an
>interpretation of that spectrum. The camera is a machine. It records
>whatever photons strike the imaging material. It has no idea what color
>those photons are. Whenever you take a picture, all the light is recorded.
>Nothing is subtracted in order to adjust white balance.
>

I'm assuming that different color temperatures are related to
differences in photon spectra that strike the sensor. Correct? And
doesn't that imply that some temps would not have the range of photons
that others would have? What I'm trying to find out is if the image
has more "color" at certain temperatures. I'm tempted to say photon
saturation but I doubt that's the right term. In other words, can you
extract the same range of color from a properly exposed photo taken at
3000 K as one taken at 8000 K?

, Feb 21, 2005

good question. i think you can, though i can't say for certain... as
long as the quantity of light is the same, the quality (tungsten vs.
daylight) of the light shouldn't affect the extractable range, barring
any inherent equipment prefrences...i think.

i have a question. why can i easily distinguish between different
notes in music but go bananas trying to distinguish RGB levels for any
given color? is it just me, or do many have problems with color
correcting?

googlegroups2sucks, Feb 21, 2005
6. ### C J CampbellGuest

<> wrote in message
news...
> >

> I'm assuming that different color temperatures are related to
> differences in photon spectra that strike the sensor. Correct? And
> doesn't that imply that some temps would not have the range of photons
> that others would have?

Well, suppose all your light comes from the red spectrum. No amount of
correction is going to bring back the blue and green colors. You cannot
record something that is not there. If you walk into a grocery store lit
with fluorescent lighting then you are not going to be able to see
everything that you would have with tungsten lighting. If elements of the
spectrum are missing in the lighting, neither your eyes nor a camera can
restore them.

I found it very difficult, myself, adjusting to your world's yellow sun.
Everything looks so much different than it does under the blue-white sun of
my planet, even with the special filters in my third eyelids.

C J Campbell, Feb 21, 2005
7. ### OwamangaGuest

On Mon, 21 Feb 2005 05:38:13 -0800, "C J Campbell"
<> wrote:

>
><> wrote in message
>news...
>> >

>> I'm assuming that different color temperatures are related to
>> differences in photon spectra that strike the sensor. Correct? And
>> doesn't that imply that some temps would not have the range of photons
>> that others would have?

>
>Well, suppose all your light comes from the red spectrum. No amount of
>correction is going to bring back the blue and green colors. You cannot
>record something that is not there. If you walk into a grocery store lit
>with fluorescent lighting then you are not going to be able to see
>everything that you would have with tungsten lighting. If elements of the
>spectrum are missing in the lighting, neither your eyes nor a camera can
>restore them.
>
>I found it very difficult, myself, adjusting to your world's yellow sun.
>Everything looks so much different than it does under the blue-white sun of
>my planet, even with the special filters in my third eyelids.

You should have noticed on your way here that our Sun is actually
white. Only when seen through our atmosphere does it appear yellow, we
needed to scatter out some of the blue to make the sky that color.

Presuming your sun is white too, yet it appears blueish from the
planet surface, I'm guessing your home planet's daylight skies are a
reddy-yellowish color.

BTW, no matter what the English tell you, just because you have a
Scottish name doesn't *actually* mean you come from another planet -
just that it seems that way to some.

--
Owamanga!

Owamanga, Feb 21, 2005
8. ### LarryGuest

In article <>,
says...
> good question. i think you can, though i can't say for certain... as
> long as the quantity of light is the same, the quality (tungsten vs.
> daylight) of the light shouldn't affect the extractable range, barring
> any inherent equipment prefrences...i think.
>
> i have a question. why can i easily distinguish between different
> notes in music but go bananas trying to distinguish RGB levels for any
> given color? is it just me, or do many have problems with color
> correcting?
>
>

Your EARS are MUCH more discriminating than your eyes.

Think of this... the picture on your TV is flickering at a rate of 30 times a
second (US NTSC) or 25 times a second (PAL) but you manage to watch it
anyhow.. If the music from your cd player did the same thing you would go
NUTS looking for the source of the hum/noise.

Color correcting/color timeing IS NOT AS EASY AS YOU WOULD THINK!

--
Larry Lynch
Mystic, Ct.

Larry, Feb 21, 2005
9. ### paulGuest

> good question. i think you can, though i can't say for certain... as
> long as the quantity of light is the same, the quality (tungsten vs.
> daylight) of the light shouldn't affect the extractable range, barring
> any inherent equipment prefrences...i think.
>
> i have a question. why can i easily distinguish between different
> notes in music but go bananas trying to distinguish RGB levels for any
> given color? is it just me, or do many have problems with color
> correcting?

I think I read that the human eye has built in white balance correction
so we naturally try to make things look normal. Certainly we adjust for
exposure.

paul, Feb 21, 2005
10. ### LarryGuest

In article <>, says...
> > good question. i think you can, though i can't say for certain... as
> > long as the quantity of light is the same, the quality (tungsten vs.
> > daylight) of the light shouldn't affect the extractable range, barring
> > any inherent equipment prefrences...i think.
> >
> > i have a question. why can i easily distinguish between different
> > notes in music but go bananas trying to distinguish RGB levels for any
> > given color? is it just me, or do many have problems with color
> > correcting?

>
>
> I think I read that the human eye has built in white balance correction
> so we naturally try to make things look normal. Certainly we adjust for
> exposure.
>

We do have "built in" white balance..

Go for a walk in a leafy forest with a friend who is wearing a white shirt.
Almost instantaneously the shirt will appear white although most of the
available light is coming through a green filter.

Without any "white balance" the shirt will photograph greenish, but you wont
notice it while you are walking through the forest.

This is only one example and perhaps not the best example, but it is one that
is EASY to try for most people.

--
Larry Lynch
Mystic, Ct.

Larry, Feb 21, 2005
11. ### Chris BrownGuest

In article <>,
Larry <> wrote:
>
>We do have "built in" white balance..
>
>Go for a walk in a leafy forest with a friend who is wearing a white shirt.
>Almost instantaneously the shirt will appear white although most of the
>available light is coming through a green filter.
>
>Without any "white balance" the shirt will photograph greenish, but you wont
>notice it while you are walking through the forest.

It's possible to fool your builtin colour balance. A couple of years ago, I
was hiking in Zion Canyon in the southern US, when I looked at my T-shirt,
which should have been white, and noticed that it was green.

This seemed strange, and then I looked at my hands, and they were green too.
So I looked back at my companions, and their skin was also green. In fact,
*everything* was tinged green.

When I mentioned this, some of them had also noticed that the world had
adopted a green tint. It seems that the narrow canyon, with its sheer red
sandstone walls towering above us had changed the colour of the ambient
daylight from blue to red, but because the walls were *so* high in that part
of the canyon, we weren't getting much of a reference from familliar things
such as the sky, and our colour balance had wandered off.

After we got into a more open area, everything went back to its usual colour
within a few minutes.

Chris Brown, Feb 21, 2005
12. ### Guest

In message <>,
wrote:

>This isn't limited to digital photography. How appropriate is it to
>consider color temperature a filter? And to what extent is the missing
>or added portions of the color spectrum significant? Even with proper
>white balance are data lost (or added?) at 3000 K when compared to
>8000 K?

Filters don't miss and add portions of the color spectrum, they merely
cut different frequency levels by different amounts. A red filter
doesn't make a blue laser beam red; it just makes it very dark.

Color filters are most useful in Digital only when they equalize the
maximum level in each channel preventing any of the channels from having
a low signal-to-noise ratio and heavy posterization.

The natural color balance of digital cameras, unlike most films, is
*nowhere* near white. Most RGB digitals are most sensitive in the green
channel (not because of twice as many green pixels in a Bayer matrix,
but on a per-pixel level), and almost all are less sensitive in the red
channel in anywhere from about a half stop to a stop plus. Blue is
usually about the same as green to about a half stop less sensitive.

In general, the natural color balance is in the green to cyan range. A
filter with a hue 180 degrees away from the natural balance of the
camera, equally saturated, is necessary to use all the dynamic range
with a white highlight in white light.

The normal "color temperature" filters are about deviations in raw light
sources, and do not balance green vs purple, so do not adequately
compensate a 3-dimensional color cube, or the natural color of the
digital to white. Putting an 80A filter on a digital under incandescent
light may allow you to shoot as if it were daylight and use the daylight
color balance on the camera, but in the RAW data, it will make the blue
stronger, relative to red and green, but it will make red stronger, too,
making it stronger than the green. An 80B lets less red through, and is
more centered on the blue, but it still leaves the blue channels about a
stop weaker than the red and green on my Canon 20D. It would be two
stops without the filter. The same filter, twice as thick, would
equalize the channels very well in incandescent light.

Why are there no filters adjusted for digital cameras? Because on
mental lethargy, probably. Digital cameras are treated with film
mentality.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><

, Feb 21, 2005
13. ### Guest

In message <>,
"C J Campbell" <> wrote:

>White balance is not at all like a filter. A filter subtracts portions of
>the spectrum before the light reaches your lens. White balance is an
>interpretation of that spectrum. The camera is a machine. It records
>whatever photons strike the imaging material. It has no idea what color
>those photons are. Whenever you take a picture, all the light is recorded.
>Nothing is subtracted in order to adjust white balance.

Most converters use the green channel as a reference, and scale the red
and green to achieve a color temperature. "Subtraction" is not a good
word to use here, but diminutive scaling of RAW data does in fact occur
for certain color temperatures.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><

, Feb 21, 2005
14. ### Guest

In message <>,
wrote:

>I'm assuming that different color temperatures are related to
>differences in photon spectra that strike the sensor. Correct? And
>doesn't that imply that some temps would not have the range of photons
>that others would have? What I'm trying to find out is if the image
>has more "color" at certain temperatures. I'm tempted to say photon
>saturation but I doubt that's the right term. In other words, can you
>extract the same range of color from a properly exposed photo taken at
>3000 K as one taken at 8000 K?

I guess my original reply had little to do with your question ...

There will be differences with different temperature light sources,
because that will change the balance of the frequencies under the
umbrella of each of the three color filter's effective ranges. How big
they are, I don't know. One way to find out is to use different color
temperature light sources and test their effect on the same subject,
under controlled lighting.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><

, Feb 21, 2005
15. ### C J CampbellGuest

"Owamanga" <> wrote in message
news:...
> >
> >I found it very difficult, myself, adjusting to your world's yellow sun.
> >Everything looks so much different than it does under the blue-white sun

of
> >my planet, even with the special filters in my third eyelids.

>
> You should have noticed on your way here that our Sun is actually
> white. Only when seen through our atmosphere does it appear yellow, we
> needed to scatter out some of the blue to make the sky that color.
>
> Presuming your sun is white too, yet it appears blueish from the
> planet surface, I'm guessing your home planet's daylight skies are a
> reddy-yellowish color.
>
> BTW, no matter what the English tell you, just because you have a
> Scottish name doesn't *actually* mean you come from another planet -
> just that it seems that way to some.
>
> --
> Owamanga!

Oh. Actually, it was an Irishman that told me that.

C J Campbell, Feb 21, 2005

i feel better knowing i'm not the only one. excellent example, btw.

googlegroups2sucks, Feb 22, 2005
17. ### Guest

On Mon, 21 Feb 2005 17:17:31 GMT, wrote:

>Filters don't miss and add portions of the color spectrum, they merely
>cut different frequency levels by different amounts. A red filter
>doesn't make a blue laser beam red; it just makes it very dark.
>

I didn't know that. I thought they limited or expanded the spectrum.

>Color filters are most useful in Digital only when they equalize the
>maximum level in each channel preventing any of the channels from having
>a low signal-to-noise ratio and heavy posterization.

Suppose the same lens is used on different cameras with different
sensors designs. Might one have more "need" for a color filter than
another?

And I'm still confused about real-world color temperature. That
"magical" light of the early morning or late afternoon--is the
spectrum range the same as at high noon? What I'm trying to understand
has to do with post-processing. Even though I may not find the light
as appealing to my eye at noon, if that captures a more robust
spectral range, then doesn't it follow that I have more to work with
in post-processing than I would with a photo taken at the more
appealing times of day? And I'm over my head here, it may be just the
opposite. Or neither.

, Feb 22, 2005
18. ### Guest

In message <>,
wrote:

>On Mon, 21 Feb 2005 17:17:31 GMT, wrote:
>
>>Filters don't miss and add portions of the color spectrum, they merely
>>cut different frequency levels by different amounts. A red filter
>>doesn't make a blue laser beam red; it just makes it very dark.
>>

>I didn't know that. I thought they limited or expanded the spectrum.
>
>>Color filters are most useful in Digital only when they equalize the
>>maximum level in each channel preventing any of the channels from having
>>a low signal-to-noise ratio and heavy posterization.

>
>Suppose the same lens is used on different cameras with different
>sensors designs. Might one have more "need" for a color filter than
>another?

Sure, but it depends on what you mean by "need". You can adjust purely
in software; I am talking about best-case scenarios, which most people
aren't doing, anyway. Different cameras have different sensors and
color filters, and therefore have different overall sensitivity ratios
between the channels.

>And I'm still confused about real-world color temperature. That
>"magical" light of the early morning or late afternoon--is the
>spectrum range the same as at high noon?

The ratios of different colors change. When the sun is close to the
horizon, more of the blue end of the spectrum is lost to diffusion in
the atmosphere, but the warmer colors cut through better.

>What I'm trying to understand
>has to do with post-processing. Even though I may not find the light
>as appealing to my eye at noon, if that captures a more robust
>spectral range, then doesn't it follow that I have more to work with
>in post-processing than I would with a photo taken at the more
>appealing times of day?

You could alter the noon colors to look like dusk/dawn colors, but the
difference in lighting between the two is not just color. At noon, the
light may be a virtual point source of light high in the sky; when the
sun is near the horizon, shadows are longer and a little softer.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><

, Feb 22, 2005
19. ### OwamangaGuest

On Tue, 22 Feb 2005 08:18:15 -0800, wrote:

>On Mon, 21 Feb 2005 17:17:31 GMT, wrote:
>
>>Filters don't miss and add portions of the color spectrum, they merely
>>cut different frequency levels by different amounts. A red filter
>>doesn't make a blue laser beam red; it just makes it very dark.
>>

>I didn't know that. I thought they limited or expanded the spectrum.

No, as far as I know the only thing that behaves that way is a
fluorescent surface. It's capable of emitting a different frequency of
light from the one that hit it. Normal surfaces only bounce some of
the original spectrum that hit them. As for fluorescent filters (real
fluorescing filters, *not* filters for shooting under florescent
light), I don't know if they 1) exist, 2) can actually shift
frequencies but generally this wouldn't be very useful.

>>Color filters are most useful in Digital only when they equalize the
>>maximum level in each channel preventing any of the channels from having
>>a low signal-to-noise ratio and heavy posterization.

>
>Suppose the same lens is used on different cameras with different
>sensors designs. Might one have more "need" for a color filter than
>another?

I doubt it.

>And I'm still confused about real-world color temperature. That
>"magical" light of the early morning or late afternoon--is the
>spectrum range the same as at high noon?

No, it's much warmer at daybreak or late afternoon than at noon. The
sun continues to emit white light all day and all night, but as the
angle of the sunlight hitting the atmosphere gets more acute, more and
more blue from the light gets scattered. Also, dirt in the atmosphere
helps add to the reddening and there is probably some refraction going
on too.

>What I'm trying to understand
>has to do with post-processing. Even though I may not find the light
>as appealing to my eye at noon, if that captures a more robust
>spectral range, then doesn't it follow that I have more to work with
>in post-processing than I would with a photo taken at the more
>appealing times of day?

Firstly, the color temp of the light is only one of the aspects of the
image that make it appealing. Shadows change (become longer) and
subjects can become backlit (which is rare at midday, unless the
subject is above you). Plus, the difference between the background (a
nice warm shadow-rich lighting) and a properly lit subject with
fill-flash would be lost if you tried to do this at noon.

Secondly, your exposure can only capture a fixed latitude, why waste
some of that capturing light that you'll end up throwing away in post
processing. (in other words, your exposure will be under-exposed as
far as the colors you want to keep in the final image) It's much
better to capture the light in as close to the spectrum as you'll be
wanting to print later.

These may sound contradictory, but overall, I'm arguing you shouldn't
use a filter at noon to color the light, capture the scene at the
right time of the day instead. It'll look much better.

>And I'm over my head here, it may be just the
>opposite. Or neither.

Yes, a bit of everything.

That all said, I can't think of a situation where I'd want to put a
color-casting filter on a digital camera. Even shooting B&W where a
strong color filter is often used in traditional film photography,
I'll capture a full-spectrum color image and do the filtration in
software. White balance for color photos follows the same rule - deal
with this when importing the raw file.

--
Owamanga!

Owamanga, Feb 22, 2005