Optical Density vs. Transmission

[Alan Smithee]

OK I'm still flogging this horse Bart. Is there a way (a formula) for me to
convert optical density (of film) to transmission (photoshop)? Or, how do I
use that last formula to do this. Is there a way to edit in "density mode"
in photoshop or is that a bad idea? My intended application is for use on
silver and "alt" process where density of my output (neg) is critically
important. Thx.

I think it should follow a table like below...

Optical Density Transmission (%)
0 100
0.3 50
0.6 25
0.9 12.5
....
etc.

 

[Alan Smithee]

OK I'm still flogging this horse Bart. Is there a way (a formula) for
me to convert optical density (of film) to transmission (photoshop)?
Or, how do I use that last formula to do this. Is there a way to edit
in "density mode" in photoshop or is that a bad idea? My intended
application is for use on silver and "alt" process where density of
my output (neg) is critically important. Thx.

I think it should follow a table like below...

Optical Density Transmission (%)
0 100
0.3 50
0.6 25
0.9 12.5
...
etc.

How about:

T = 1/10d, where T=tranmission and where d= Optical Density

 

[Kennedy McEwen]

How about:

T = 1/10d, where T=tranmission and where d= Optical Density

No, you were right the first time, with the table and the formulae that
Bart already provided:
Opacity = 1/Transmission
Density = log10(opacity)
Hence Density = -log10(Transmission)

Thus, as per your original table above, if you have transmission of 50%
then the density is -log10(50/100) = 0.3 and if the transmission is
12.5% the density is -log10(12.5/100) = 0.9, just as in your table
above.

Conversely, if you have film of density 2.5, then the transmission can
be determined from 10^(-Density), ie. 10^(-2.5), which is 0.32%. Thus,
if your scanned image has been scaled such that 100% transmission gives
data of 255 on the linear 8-bit scale, that film will just barely be
represented by 1. On a 16-bit scale it would be around 207.

Of course, the actual result will be somewhat different because scanners
are not particularly accurate densitometers.

 

[Alan Smithee]

No, you were right the first time, with the table and the formulae
that Bart already provided:
Opacity = 1/Transmission
Density = log10(opacity)
Hence Density = -log10(Transmission)

Thus, as per your original table above, if you have transmission of
50% then the density is -log10(50/100) = 0.3 and if the transmission
is
12.5% the density is -log10(12.5/100) = 0.9, just as in your table
above.

Conversely, if you have film of density 2.5, then the transmission can
be determined from 10^(-Density), ie. 10^(-2.5), which is 0.32%.
Thus, if your scanned image has been scaled such that 100%
transmission gives data of 255 on the linear 8-bit scale, that film
will just barely be represented by 1. On a 16-bit scale it would be
around 207.

Of course, the actual result will be somewhat different because
scanners are not particularly accurate densitometers.

First of all. Thank you. What I'm trying to come to terms with is trying to
find a way to tweak, fool, tune --whatever -- photoshop so that it (and I)
can natively work to output a negative instead of paper positive. Photoshop
seems to understand the RGB table (0-255) and darkness/lightness values of
0-100 per cent well enough (is there a name for the 0 to 100 percent table?
"Blackness scale" perhaps?)
I'm not sure where I'm stuck at the moment. Is it because the DLog scale
is used for film and the "Linear" scale is being used by everything else?
How do people who "output to film" using film recorders work with Photoshop?
I'm guessing these devices are calibrated to work with whatever output media
they are using and have look up tables of some sort to calculate how to
recreate the image from Photoshop on film. Thx.

 

[CSM1]

First of all. Thank you. What I'm trying to come to terms with is trying
to
find a way to tweak, fool, tune --whatever -- photoshop so that it (and I)
can natively work to output a negative instead of paper positive.
Photoshop
seems to understand the RGB table (0-255) and darkness/lightness values of
0-100 per cent well enough (is there a name for the 0 to 100 percent
table?
"Blackness scale" perhaps?)
I'm not sure where I'm stuck at the moment. Is it because the DLog scale
is used for film and the "Linear" scale is being used by everything else?
How do people who "output to film" using film recorders work with
Photoshop?
I'm guessing these devices are calibrated to work with whatever output
media
they are using and have look up tables of some sort to calculate how to
recreate the image from Photoshop on film. Thx.

I think that the film recorder has the software the takes a positive image
and creates either a Slide or a negative depending on what you want the
final film to be.

You use a standard digital image that looks good and has the color profile
embedded.
The film record then works from that image.

So unless you are writing the software for a film recorder, you do not have
to worry about how it is done.

 

[Kennedy McEwen]

First of all. Thank you. What I'm trying to come to terms with is trying to
find a way to tweak, fool, tune --whatever -- photoshop so that it (and I)
can natively work to output a negative instead of paper positive. Photoshop
seems to understand the RGB table (0-255) and darkness/lightness values of
0-100 per cent well enough (is there a name for the 0 to 100 percent table?
"Blackness scale" perhaps?)

Generally just called "levels".

I'm not sure where I'm stuck at the moment. Is it because the DLog scale
is used for film and the "Linear" scale is being used by everything else?

I wouldn't think so - there is a direct relationship between linear
transmission and density so, providing that you are clear in what you
are talking about, and generally the context helps a lot, there is
little confusion. Density tends to range from 0 to about 4 or so.
Transmission ranges from 100% down to 0.1% or so, hence little
opportunity for confusion.

How do people who "output to film" using film recorders work with Photoshop?

The same way as they output to printers - usually the output device is
calibrated with a defined profile and this is called up by the Photoshop
colour management. There shouldn't be any need for more involvement on
your part, unless you are trying to actually determine the profile
itself.

Generally speaking, whilst I am sure everyone here wants to help, this
group has more experience with input from film and other media, so you
might not get all of your answers.

Do people *still* output to film? ;-)
OK - its rhetorical I know. My cat scans went onto film so the doctor
could make his diagnosis in a traditional manner rather than needing the
high contrast display for the raw data.

 

[Alan Smithee]

Generally just called "levels".

At 50 per cent black the black dots are equal to the number of white dots on
the page. I think "levels" is more a term made popular by Photoshop. Graphic
artists have been using the term screen when calling for a certain amount of
a particular color.

I wouldn't think so - there is a direct relationship between linear
transmission and density so, providing that you are clear in what you
are talking about, and generally the context helps a lot, there is
little confusion. Density tends to range from 0 to about 4 or so.
Transmission ranges from 100% down to 0.1% or so, hence little
opportunity for confusion.


The same way as they output to printers - usually the output device is
calibrated with a defined profile and this is called up by the
Photoshop colour management. There shouldn't be any need for more
involvement on your part, unless you are trying to actually determine
the profile itself.

Generally speaking, whilst I am sure everyone here wants to help, this
group has more experience with input from film and other media, so you
might not get all of your answers.

Do people *still* output to film? ;-)

Not film, that was ten years ago. But the popular equivalent today for those
of us stuck somewhere between the digital world and the darkroom is the
digital negative produced on pictorico OHP or HGWF. Same dilemma as creating
a negative however. My $200 Epson R300 printer is capable of printing at
5760 DPI which is essentially continous tone. Outputting on an 8.5X11 sheet
of OHP you can produce one honkin' big negative opening up the world of Alt
Processes (Cyanotype, Van Dyke, Pt/Pd. etc.) The problem lies in that
tranparency material absorbs the ink much differently than paper. What I'm
doing now I think is printing using a linear driver (paper) to print to a
non-linear medium (transparency). I'm needing software which is slightly
more sophisticated the the Photoshop Curves tool I now use. Curves is a good
starting point but the controls are a bit touchy-feely.

 

[Kennedy McEwen]

At 50 per cent black the black dots are equal to the number of white dots on
the page.

Not necessarily - it depends on the dot gain. On most media if each dot
on the image is alternately ink and not ink then the image will be much
darker than 50% grey, due to the dot gain alone.

I think "levels" is more a term made popular by Photoshop.

Possibly popularised by it, but it has been used in digital
representations of analogue circuits almost since the inception of the
ADC, interchangeably with "count".

Graphic
artists have been using the term screen when calling for a certain amount of
a particular color.

Again, I think this relates more to the half-tone structure rather than
a particular colour or the density of a particular ink. The screen
defines the resolution of the halftone dots and the angle between the
axes for dots of each primary colour.

Not film, that was ten years ago. But the popular equivalent today for those
of us stuck somewhere between the digital world and the darkroom is the
digital negative produced on pictorico OHP or HGWF. Same dilemma as creating
a negative however. My $200 Epson R300 printer is capable of printing at
5760 DPI which is essentially continous tone. Outputting on an 8.5X11 sheet
of OHP you can produce one honkin' big negative opening up the world of Alt
Processes (Cyanotype, Van Dyke, Pt/Pd. etc.) The problem lies in that
tranparency material absorbs the ink much differently than paper. What I'm
doing now I think is printing using a linear driver (paper) to print to a
non-linear medium (transparency). I'm needing software which is slightly
more sophisticated the the Photoshop Curves tool I now use. Curves is a good
starting point but the controls are a bit touchy-feely.

No - what you need is a printer profile for the media you are using! The
system colour management should be able to cope with this - it is
exactly what it is there for.

There should be a default profile for your Epson R300 printer that is
suitable for the transparency material. Make sure that you are printing
to the printer's colour space (eg. by selecting Printer Color Management
in Print/Preview) and Photoshop should do all of the required
conversions transparently (no pun intended) if you select the
appropriate media in your printer driver.

Generally these default profiles are fairly reasonable for the
combination of Epson transparencies and Epson inks, if that medium is
supported by the printer. However, more accurate results can probably
be obtained using a profiling tool and it is certainly required if you
use non-OEM consumables. I personally have never found the need for
them, but a lot of professionals and quite a few amateurs swear by them.