filters used with digital ?

Paul Mitchum wrote:
> stuseven <stuseven.TakeThisOut@hotmail.com> wrote:
>
>>+ With my old Canon EF... a 35mm film camera... I used screw on filters,
>>with a color-temperature meter, for nice color correct slides. With a
>>digital camera, and photo-processing software, filters are "no longer
>>necessary", as just about any kind of color correction can be done on the
>>computer... however... the following question occurs to me_
>>
>> + would color-correcting via filters mounted in front of lens_ improve
>>definition in the image, having now removed any non-neutral color
>>pollution ? While I've had tremendous success with digital processing,
>>sans filters, I'd suspect anything which could be done to remove
>>extraneous effects could only strengthen the printable image. Anyone who
>>uses filters with their digital camera familiar with this issue ?
>
> If all you're doing is color correcting, then there's no good reason to
> shoot digital with a filter. Unless, of course, you have artistic
> reasons, and/or you want to minimize post-processing.

Actually, there is a good reason: maximize signal-to-noise ratio.
Color correction in camera can boost signal and that boosts the
noise in that channel, e.g. in shade mode, red is amplified,
along with noise.

A color correction filter cuts color(s) so you get better
signal over the whole visible spectrum. This may not be as
important on large pixel cameras as much as small pixel
cameras that suffer from lower S/N.

Roger
 >> Stay informed about: filters used with digital ? 

stuseven wrote:
> +
> With my old Canon EF... a 35mm film camera... I used screw on
> filters, with a color-temperature meter, for nice color correct
> slides. With a digital camera, and photo-processing software,
> filters are "no longer necessary", as just about any kind of color
> correction can be done on the computer... however... the following
> question occurs
> to me_
>
> + would color-correcting via filters mounted in front of lens_
> improve
> definition in the image, having now removed any non-neutral color
> pollution ?
> While I've had tremendous success with digital processing, sans
> filters,
> I'd suspect anything which could be done to remove extraneous effects
> could only strengthen the printable image. Anyone who uses filters
> with
> their digital camera familiar with this issue ?
>
> Thanks

This answer is best answered by trial and error. It depends on the
particulars of the camera and even the file type. For example you will get
more room to make adjustments using RAW images than JEP.

Some filters can not reasonably be simulated post exposure such as
polarizing filters.

--
Joseph Meehan

Dia 's Muire duit
 >> Stay informed about: filters used with digital ? 

Roger N. Clark (change username to rnclark) <username DeleteThis @qwest.net> wrote:

> Paul Mitchum wrote:
> > stuseven <stuseven DeleteThis @hotmail.com> wrote:
> >
> >>+ With my old Canon EF... a 35mm film camera... I used screw on filters,
> >>with a color-temperature meter, for nice color correct slides. With a
> >>digital camera, and photo-processing software, filters are "no longer
> >>necessary", as just about any kind of color correction can be done on the
> >>computer... however... the following question occurs to me_
> >>
> >> + would color-correcting via filters mounted in front of lens_ improve
> >>definition in the image, having now removed any non-neutral color
> >>pollution ? While I've had tremendous success with digital processing,
> >>sans filters, I'd suspect anything which could be done to remove
> >>extraneous effects could only strengthen the printable image. Anyone who
> >>uses filters with their digital camera familiar with this issue ?
> >
> > If all you're doing is color correcting, then there's no good reason to
> > shoot digital with a filter. Unless, of course, you have artistic
> > reasons, and/or you want to minimize post-processing.
>
> Actually, there is a good reason: maximize signal-to-noise ratio.
> Color correction in camera can boost signal and that boosts the
> noise in that channel, e.g. in shade mode, red is amplified,
> along with noise.

I should have pointed out that I meant while shooting RAW. But even
then, unless you're shooting under green-gel stage lights and want to
boost red by, like, three stops or something, it's not an issue.

> A color correction filter cuts color(s) so you get better signal over the
> whole visible spectrum. This may not be as important on large pixel
> cameras as much as small pixel cameras that suffer from lower S/N.

OK, so changing white balance in a digital camera is equivalent to
changing the response curves of the red, green, and blue sensors, such
that the low end of one curve might end up being boosted more than those
of others. However, in any practical application, all this will do is
alter the average tint of the noise, not increase it.

And, coincidentally, this is exactly what using a color correction
filter will do, as well.
 >> Stay informed about: filters used with digital ? 

Paul Mitchum wrote:

> Roger N. Clark (change username to rnclark) <username.DeleteThis@qwest.net> wrote:
>>Actually, there is a good reason: maximize signal-to-noise ratio.
>>Color correction in camera can boost signal and that boosts the
>>noise in that channel, e.g. in shade mode, red is amplified,
>>along with noise.
>
> I should have pointed out that I meant while shooting RAW. But even
> then, unless you're shooting under green-gel stage lights and want to
> boost red by, like, three stops or something, it's not an issue.
>
>>A color correction filter cuts color(s) so you get better signal over the
>>whole visible spectrum. This may not be as important on large pixel
>>cameras as much as small pixel cameras that suffer from lower S/N.
>
> OK, so changing white balance in a digital camera is equivalent to
> changing the response curves of the red, green, and blue sensors, such
> that the low end of one curve might end up being boosted more than those
> of others.

That is a gain change of the amplifier for that channel, not a change
in response curve. The data out of the sensor are linear.

> However, in any practical application, all this will do is
> alter the average tint of the noise, not increase it.

No. Changing gain changes noise in proportion to the gain.

> And, coincidentally, this is exactly what using a color correction
> filter will do, as well.

No, a color correction filter absorbs light, not boosts gain.
The photographer applies a filter factor to increase the exposure
time, so the number of photons remains high.

Roger
 >> Stay informed about: filters used with digital ? 

On Tue, 09 Jan 2007 18:42:01 -0800, Roger N. Clark (change username
to rnclark) wrote:

>> If all you're doing is color correcting, then there's no good reason to
>> shoot digital with a filter. Unless, of course, you have artistic
>> reasons, and/or you want to minimize post-processing.
>
> Actually, there is a good reason: maximize signal-to-noise ratio.
> Color correction in camera can boost signal and that boosts the
> noise in that channel, e.g. in shade mode, red is amplified,
> along with noise.
>
> A color correction filter cuts color(s) so you get better
> signal over the whole visible spectrum. This may not be as
> important on large pixel cameras as much as small pixel
> cameras that suffer from lower S/N.

Ah, yes, it's important to stress "suffer". At least your bias is
consistent. With the amount of red channel (shades of the 1950's)
amplification provided by a "shade" mode, would the additional noise
be noticeable in most prints, or is it most useful as a pixel
peeper's talking point? It would seem that for additional noise to
be really noticeable, a WB correction more extreme than a "shade"
mode would have to be applied. And any comparably opaque external
color filter would probably also result in increased noise if used
in the less brightly illuminated conditions referred to as "shade".
 >> Stay informed about: filters used with digital ? 

ASAAR wrote:

> On Tue, 09 Jan 2007 18:42:01 -0800, Roger N. Clark (change username
> to rnclark) wrote:
>
>>>If all you're doing is color correcting, then there's no good reason to
>>>shoot digital with a filter. Unless, of course, you have artistic
>>>reasons, and/or you want to minimize post-processing.
>>
>>Actually, there is a good reason: maximize signal-to-noise ratio.
>>Color correction in camera can boost signal and that boosts the
>>noise in that channel, e.g. in shade mode, red is amplified,
>>along with noise.
>>
>>A color correction filter cuts color(s) so you get better
>>signal over the whole visible spectrum. This may not be as
>>important on large pixel cameras as much as small pixel
>>cameras that suffer from lower S/N.
>
>
> Ah, yes, it's important to stress "suffer". At least your bias is
> consistent. With the amount of red channel (shades of the 1950's)
> amplification provided by a "shade" mode, would the additional noise
> be noticeable in most prints, or is it most useful as a pixel
> peeper's talking point? It would seem that for additional noise to
> be really noticeable, a WB correction more extreme than a "shade"
> mode would have to be applied. And any comparably opaque external
> color filter would probably also result in increased noise if used
> in the less brightly illuminated conditions referred to as "shade".

Like I said, whether or not this is important is dependent
on the camera, its noise, and the light conditions.
For example, here in the western US, skies can be very
blue with a factor of 10 or so lower intensity in red
than blue compared to normal sunlight.
So this would be like working the red channel at
ISO 1000 and the blue at ISO 100. If your
camera does well at ISO 1000, it may not be an issue.
Of course this assumes you can increase the exposure
time to compensate for the absorption by the filter,
as was normally done with film cameras.

I had a situation where I was copying color negatives
with a 1D Mark II. The film base orange cast was strong enough
that when I corrected the images, the blue channel was very
noisy. A blue color correction filter helped.
While this is an extreme case, I was using a high end camera
and it was still an issue.

Roger
http://www.clarkvision.com
 >> Stay informed about: filters used with digital ? 

On Tue, 09 Jan 2007 20:02:25 -0800, Roger N. Clark (change username
to rnclark) wrote:

>> It would seem that for additional noise to
>> be really noticeable, a WB correction more extreme than a "shade"
>> mode would have to be applied. And any comparably opaque external
>> color filter would probably also result in increased noise if used
>> in the less brightly illuminated conditions referred to as "shade".
>
> Like I said, whether or not this is important is dependent
> on the camera, its noise, and the light conditions.
> For example, here in the western US, skies can be very
> blue with a factor of 10 or so lower intensity in red
> than blue compared to normal sunlight.

But now you're not talking about conditions where a standard
"shade" WB option wouldn't be suitable. As I said, unless the use
of a more extreme WB or filter is needed, noise shouldn't be much of
a concern.


> I had a situation where I was copying color negatives
> with a 1D Mark II. The film base orange cast was strong enough
> that when I corrected the images, the blue channel was very
> noisy. A blue color correction filter helped.
> While this is an extreme case, I was using a high end camera
> and it was still an issue.

I don't doubt that it was an issue, but now you're backtracking
and raising a different concern, one which I don't disagree with.
But I highly doubt that you'd have used a "shade" WB for copying
your color negatives, and it was with the use of a "shade" WB that
you claimed that noise might be an issue. I don't think so.
 >> Stay informed about: filters used with digital ? 

Roger N. Clark wrote :
> Actually, there is a good reason: maximize signal-to-noise ratio.
> Color correction in camera can boost signal and that boosts the
> noise in that channel, e.g. in shade mode, red is amplified,
> along with noise.

On the other hand, using a less-than-perfect color filter in front of
the lens (and we don't live in a perfect world, do we?) may add some
unwanted flare and loss of contrast, which also affect SNR (but in a
different way, and that depends on the lighting conditions).

My 2 (euro's) cents : with others, I'd say that for "normal" ie
moderate color cast corrections, it is best to rely on the raw
processing. Of course this is not the case for strong color filters (ie
filters eating more than one stop light, eg?).
 >> Stay informed about: filters used with digital ? 

acl <achilleaslazarides.TakeThisOut@yahoo.co.uk> wrote:

> Paul Mitchum wrote:
> > Roger N. Clark (change username to rnclark) <username.TakeThisOut@qwest.net> wrote:
[..]
> >>A color correction filter cuts color(s) so you get better signal over
> >>the whole visible spectrum. This may not be as important on large pixel
> >>cameras as much as small pixel cameras that suffer from lower S/N.
> >
> > OK, so changing white balance in a digital camera is equivalent to
> > changing the response curves of the red, green, and blue sensors, such
> > that the low end of one curve might end up being boosted more than those
> > of others. However, in any practical application, all this will do is
> > alter the average tint of the noise, not increase it.
>
> Did you actually try it? It most certainly increases noise. Photograph
> something at your camera's highest ISO and the lowest colour temperature
> it offers (or do a manual white balance with warm lights, eg in my living
> room I need around 3000K to white balance). Convert from raw with as
> little noise reduction as possible (use raw to avoid as much NR as
> possible), and look at the blue channel. If the ISO is high enough and the
> temperature low enough, [..]

Note that I said 'practical applications.'

--
I have been thinking that I would make a proposition to my Republican
friends... that if they will stop telling lies about the Democrats, we
will stop telling the truth about them. -- Adlai E. Stevenson
 >> Stay informed about: filters used with digital ? 

Paul Mitchum wrote:
> acl <achilleaslazarides.DeleteThis@yahoo.co.uk> wrote:
>
> > Paul Mitchum wrote:
> > > Roger N. Clark (change username to rnclark) <username.DeleteThis@qwest.net> wrote:
> [..]
> > >>A color correction filter cuts color(s) so you get better signal over
> > >>the whole visible spectrum. This may not be as important on large pixel
> > >>cameras as much as small pixel cameras that suffer from lower S/N.
> > >
> > > OK, so changing white balance in a digital camera is equivalent to
> > > changing the response curves of the red, green, and blue sensors, such
> > > that the low end of one curve might end up being boosted more than those
> > > of others. However, in any practical application, all this will do is
> > > alter the average tint of the noise, not increase it.
> >
> > Did you actually try it? It most certainly increases noise. Photograph
> > something at your camera's highest ISO and the lowest colour temperature
> > it offers (or do a manual white balance with warm lights, eg in my living
> > room I need around 3000K to white balance). Convert from raw with as
> > little noise reduction as possible (use raw to avoid as much NR as
> > possible), and look at the blue channel. If the ISO is high enough and the
> > temperature low enough, [..]
>
> Note that I said 'practical applications.'

Well, in my living room, I need ISO 1600 and 3000K, as I have said. And
I have often ran into far more extreme lighting (street lamp light
where I live). Didn't you see it? Or are you one of those people that
never admit to being wrong? I'd bet the second.

Have it your way: you remain wrong no matter how many one-liners you
post.
 >> Stay informed about: filters used with digital ? 

Joseph Meehan wrote:

> This answer is best answered by trial and error. It depends on the
> particulars of the camera and even the file type. For example you will get
> more room to make adjustments using RAW images than JEP.
>
> Some filters can not reasonably be simulated post exposure such as
> polarizing filters.

Polarizers, definitely... ND and grad-ND filters are also handy for things not
easily duplicated in post-processing... for example, to control
very-high-contrast situations, or if you need to seriously cut the light down
for something (say, a long exposure with a wide aperture on a bright scene).
 >> Stay informed about: filters used with digital ? 

Paul Mitchum wrote:
> stuseven <stuseven.DeleteThis@hotmail.com> wrote:
>
> What sort of 'extraneous effects' are you wishing to remove?

thanks for your answer Paul M._ that may have been a poor choice
of terms... in particular, what I had referred to in "extraneous
effect"
was the "mushiness" seen in prints from negatives, in areas which
had not been color-corrected.
My original post was asking if a similar effect applied to digital,
which the next poster seemed to address specifically
 >> Stay informed about: filters used with digital ? 

+ ah Roger, yes... this is more what I was looking for with
my original question... it just makes sense that feeding a digital
sensor "purer color" will result in [stronger signal recording].

Thanks to you, and all others, for posting their excellent remarks !

Roger N. Clark (change username to rnclark) wrote:

> Actually, there is a good reason: maximize signal-to-noise ratio.
> Color correction in camera can boost signal and that boosts the
> noise in that channel, e.g. in shade mode, red is amplified,
> along with noise.
>
> A color correction filter cuts color(s) so you get better
> signal over the whole visible spectrum. This may not be as
> important on large pixel cameras as much as small pixel
> cameras that suffer from lower S/N.
>
> Roger
 >> Stay informed about: filters used with digital ?