Shooting Black & White with Digital

I'm interested in experimenting with Black and White prints. I have a
Canon 10D which (unlike the 20D) will take only in colour, so I need to
convert afterwards.

I'm familiar with the use of Channel Mixers in Photoshop to convert the
RGB channels to grey in varying proportions. What I'm interested in
finding out is, how does a typical black and white film record the
luminosity of the various colour channels it is receiving?

ie, if you take a standard Ilford B&W film, is it capturing the
luminosity of the image across all 3 RGB channels equally (analogous to
desaturating a colour image in Photoshop), or is it insensitive to some
parts of the spectrum (in which case what parts), or is it actually
sensitive to any parts of the spectrum (eg towards the infra-red) which
normal colour film (or a CCD sensor) is not?

The reason I ask is, I'd like to understand how best to model the
behaviour of a genuine B&W film, rather than just mess around with
Channel Mixers until I get something I like.

Thanks for any comments on this.

David

David French wrote:

I'm interested in experimenting with Black and White prints. I have a
Canon 10D which (unlike the 20D) will take only in colour, so I need to
convert afterwards.

I'm familiar with the use of Channel Mixers in Photoshop to convert the
RGB channels to grey in varying proportions. What I'm interested in
finding out is, how does a typical black and white film record the
luminosity of the various colour channels it is receiving?

ie, if you take a standard Ilford B&W film, is it capturing the
luminosity of the image across all 3 RGB channels equally (analogous to
desaturating a colour image in Photoshop), or is it insensitive to some
parts of the spectrum (in which case what parts), or is it actually
sensitive to any parts of the spectrum (eg towards the infra-red) which
normal colour film (or a CCD sensor) is not?

The reason I ask is, I'd like to understand how best to model the
behaviour of a genuine B&W film, rather than just mess around with
Channel Mixers until I get something I like.

Thanks for any comments on this.

David

Google for emulsion response curves. All the info you need.

Colin

David French wrote:

I'm familiar with the use of Channel Mixers in Photoshop to convert the
RGB channels to grey in varying proportions. What I'm interested in
finding out is, how does a typical black and white film record the
luminosity of the various colour channels it is receiving?

Most film datasheets contain spectral sensitivity graphs,
but the different manufacturers use different methods.
Ilford displays wedge-spectrograms under 2850K light,
while Kodak displays spectral sensitivity based on the
amount of energy needed to produce a specific density.
The results are difficult to compare, but it is safe to say
that the spectral response of most Ilford films is not
wildly unlike Tri-X.

The wedge spectrograph method used by Ilford produces rather
different looking results because 2850K tungsten light is rich
in red and deficient in blue and the optical elements
in the set-up absorb ultra-violet resulting in an apparent
sharp drop-off in the UV which is not inherent in the films.

See Kodak graphs:

Tri-X Pan 400:
http://www.kodak.com/global/en/profe...009_0506ac.gif

Plus X Pan:
http://www.kodak.com/global/en/profe...009_0431ac.gif

T-Max 100
http://www.kodak.com/global/en/profe...002_0547ac.gif

The curve for Tri-x is fairly typical of the classic panchromatic
films. The UV response almost certainly remains fairly strong
out to 230nm where the gelatine in the film absorbs UV, even
though Kodak only shows it down to 300nm. The response of
the film is stronger in the blue/violet region than it is in the
yellow/green or orange/red. Human vision has its strongest
response to the yellow green part of the spectrum, so typical
pan films show blues and violets as much lighter than they
appear to the human vision, and orange to medium-red as
somewhat lighter than the eye sees them.

The curve for Plus-x shows that it has a UV filter layer built in,
which is a fairly modern feature in a black and white film, but
it also has stronger response to the violet-end than Tri-X..
The Tmax films have a more even spectral response than the
older style films. but still show extra sensitivity at the
violet end of the spectrum.

Almost all modern panchromatic films (except for those
designated as extended-red) have a deep red response that
drops like a stone at around 650nm. The reason for this
is that the combination of extended-red film and tungsten
light (which is rich in the longest wavelengths of red) can
give some rather unnatural looking effects.

Many photographers use a medium yellow (Wratten 8 or K2)
filter in daylight to correct for the excess response to blue
and violet. The yellow-green (Wratten 11 or X1) is also
popular: it corrects both the blue-violet end and the red
end of the spectrum.

The reason I ask is, I'd like to understand how best to model the
behaviour of a genuine B&W film, rather than just mess around with
Channel Mixers until I get something I like.

I think that mucking about with mixing the channels, perhaps
even adjusting the contrast on each channel is probably your
best strategy. There is no real way of duplicating the spectral
response with the three colour channels, but there is nothing
particularly "correct" about the way panchromatic films render
colours. Panchromatic film response is based on the intersection
of what is technically possible with good overall sensitivity
and with what appears pleasing.

With b&w conversion from digital, you only have to worry about
what is pleasing.

Peter.
---

David French wrote:

I'm familiar with the use of Channel Mixers in Photoshop to convert the
RGB channels to grey in varying proportions. What I'm interested in
finding out is, how does a typical black and white film record the
luminosity of the various colour channels it is receiving?

Most film datasheets contain spectral sensitivity graphs,
but the different manufacturers use different methods.
Ilford displays wedge-spectrograms under 2850K light,
while Kodak displays spectral sensitivity based on the
amount of energy needed to produce a specific density.
The results are difficult to compare, but it is safe to say
that the spectral response of most Ilford films is not
wildly unlike Tri-X.

The wedge spectrograph method used by Ilford produces rather
different looking results because 2850K tungsten light is rich
in red and deficient in blue and the optical elements
in the set-up absorb ultra-violet resulting in an apparent
sharp drop-off in the UV which is not inherent in the films.

See Kodak graphs:

Tri-X Pan 400:
http://www.kodak.com/global/en/profe...009_0506ac.gif

Plus X Pan:
http://www.kodak.com/global/en/profe...009_0431ac.gif

T-Max 100
http://www.kodak.com/global/en/profe...002_0547ac.gif

The curve for Tri-x is fairly typical of the classic panchromatic
films. The UV response almost certainly remains fairly strong
out to 230nm where the gelatine in the film absorbs UV, even
though Kodak only shows it down to 300nm. The response of
the film is stronger in the blue/violet region than it is in the
yellow/green or orange/red. Human vision has its strongest
response to the yellow green part of the spectrum, so typical
pan films show blues and violets as much lighter than they
appear to the human vision, and orange to medium-red as
somewhat lighter than the eye sees them.

The curve for Plus-x shows that it has a UV filter layer built in,
which is a fairly modern feature in a black and white film, but
it also has stronger response to the violet-end than Tri-X..
The Tmax films have a more even spectral response than the
older style films. but still show extra sensitivity at the
violet end of the spectrum.

Almost all modern panchromatic films (except for those
designated as extended-red) have a deep red response that
drops like a stone at around 650nm. The reason for this
is that the combination of extended-red film and tungsten
light (which is rich in the longest wavelengths of red) can
give some rather unnatural looking effects.

Many photographers use a medium yellow (Wratten 8 or K2)
filter in daylight to correct for the excess response to blue
and violet. The yellow-green (Wratten 11 or X1) is also
popular: it corrects both the blue-violet end and the red
end of the spectrum.

The reason I ask is, I'd like to understand how best to model the
behaviour of a genuine B&W film, rather than just mess around with
Channel Mixers until I get something I like.

I think that mucking about with mixing the channels, perhaps
even adjusting the contrast on each channel is probably your
best strategy. There is no real way of duplicating the spectral
response with the three colour channels, but there is nothing
particularly "correct" about the way panchromatic films render
colours. Panchromatic film response is based on the intersection
of what is technically possible with good overall sensitivity
and with what appears pleasing.

With b&w conversion from digital, you only have to worry about
what is pleasing.

Peter.
---

Even though my camera does have a black and white mode, I prefer to
take all shots in color mode, and greyscale them myself in software.
That way, I can make some hue corrections before greyscaling, giving
differences in the resulting black and white image.

You can, as you say, study the responses others have suggested a source
for, but eventually you will want to adjust away anyway, since what you
may envision as what the scene should look like will vary with light
color temp and whatever responses either your camera or software
default to.

Even though my camera does have a black and white mode, I prefer to
take all shots in color mode, and greyscale them myself in software.
That way, I can make some hue corrections before greyscaling, giving
differences in the resulting black and white image.

You can, as you say, study the responses others have suggested a source
for, but eventually you will want to adjust away anyway, since what you
may envision as what the scene should look like will vary with light
color temp and whatever responses either your camera or software
default to.

I don't think I can add much to this very informative response, or to
your conclusions. However I will support them with the additional
observation that, in addition to variations in films and lens filters,
there is also variation in printing papers, enlarging filters, and
printing techniques that affects how we perceive black and white
photos. So pursuing the "authentic" black and white photo look can
lead us in almost any direction we like.

The original poster's desire to achieve objectivity "rather than just
mess around with Channel Mixers until I get something I like" therefore
seems doomed to subjectivity in more ways than one.

Alan

I don't think I can add much to this very informative response, or to
your conclusions. However I will support them with the additional
observation that, in addition to variations in films and lens filters,
there is also variation in printing papers, enlarging filters, and
printing techniques that affects how we perceive black and white
photos. So pursuing the "authentic" black and white photo look can
lead us in almost any direction we like.

The original poster's desire to achieve objectivity "rather than just
mess around with Channel Mixers until I get something I like" therefore
seems doomed to subjectivity in more ways than one.

Alan

David French wrote:
I'm interested in experimenting with Black and White prints. I have a
Canon 10D which (unlike the 20D) will take only in colour, so I need to
convert afterwards.

I'm familiar with the use of Channel Mixers in Photoshop to convert the
RGB channels to grey in varying proportions. What I'm interested in
finding out is, how does a typical black and white film record the
luminosity of the various colour channels it is receiving?

A photographic emulsion made up of only silver halides in a carrier has little or no response to the longer wavelengths of
light. Adding certain dyes to the emulsion sensitizes the emulsion to the other wavelengths. There is more on htis at
http://www.cheresources.com/photochem.shtml,, particularly in the section headed "Color Sensitizing".

ie, if you take a standard Ilford B&W film, is it capturing the
luminosity of the image across all 3 RGB channels equally (analogous to
desaturating a colour image in Photoshop), or is it insensitive to some
parts of the spectrum (in which case what parts), or is it actually
sensitive to any parts of the spectrum (eg towards the infra-red) which
normal colour film (or a CCD sensor) is not?

It depends. Infrared emulsions are sensitive to wavelengths longer than the human eye can see. Panchromatic emulsions aim
for a uniform response across the range of colors.

The reason I ask is, I'd like to understand how best to model the
behaviour of a genuine B&W film, rather than just mess around with
Channel Mixers until I get something I like.

The behavior of each B&W film is tailored for a purpose. I think the same approach is appropriate for digital editing of images.

Thanks for any comments on this.

David

David French wrote:
I'm interested in experimenting with Black and White prints. I have a
Canon 10D which (unlike the 20D) will take only in colour, so I need to
convert afterwards.

I'm familiar with the use of Channel Mixers in Photoshop to convert the
RGB channels to grey in varying proportions. What I'm interested in
finding out is, how does a typical black and white film record the
luminosity of the various colour channels it is receiving?

A photographic emulsion made up of only silver halides in a carrier has little or no response to the longer wavelengths of
light. Adding certain dyes to the emulsion sensitizes the emulsion to the other wavelengths. There is more on htis at
http://www.cheresources.com/photochem.shtml,, particularly in the section headed "Color Sensitizing".

ie, if you take a standard Ilford B&W film, is it capturing the
luminosity of the image across all 3 RGB channels equally (analogous to
desaturating a colour image in Photoshop), or is it insensitive to some
parts of the spectrum (in which case what parts), or is it actually
sensitive to any parts of the spectrum (eg towards the infra-red) which
normal colour film (or a CCD sensor) is not?

It depends. Infrared emulsions are sensitive to wavelengths longer than the human eye can see. Panchromatic emulsions aim
for a uniform response across the range of colors.

The reason I ask is, I'd like to understand how best to model the
behaviour of a genuine B&W film, rather than just mess around with
Channel Mixers until I get something I like.

The behavior of each B&W film is tailored for a purpose. I think the same approach is appropriate for digital editing of images.

Thanks for any comments on this.

David