Massive Spectral Density Comparison of Films

Another project from the depths of the lah-bor-a-tree on the bayou.....

From the intro: "I’ve started getting intrigued with “Spectral Density Curves” for films. Is this topic the missing “secret” for film behavior? Is it the background for the H&D curves that we spend so much time looking at and deriving - inadequately informed - conclusions? Is this the criteria why some films seem so wonderful and others suck? "

A look at the 20 35mm films from all of the major companies.

Thanks and a tip of the hat to D F Cardwell, whose comments got me thinking. And wondering.

I also think spectral sensitivity is a pretty important factor that most people ignore. It's the main reason I don't care for T-max films--their response is too linear, like B&W video.

David A. Goldfarb said:

I also think spectral sensitivity is a pretty important factor that most people ignore. It's the main reason I don't care for T-max films--their response is too linear, like B&W video.

Click to expand...

What we all need to remember is that the same film and a different developer combination can generate a dramatic and vastly different density / exposure relationship from each other.

T Max 400 in Xtol or Pyrocat at regular dilutions and processing is as you said, a straight 45 degree line.

Dilute Pyrocat with T Max 400 utilizing stand or semi stand process generates an extended long toe and an emphasized S shoulder to such a degree that you would swear that this could not be the same emulsion. I am sure that there are many other examples that fit this model.

As a result I am convinced that photographers that really know their materials can make quality photographs with virtually any film and developer combination. Looking at a graph is "interesting" but prints are the only thing that really matter.

Interesting thoughts, and yes, the spectral sensitivity curves are certainly something we don't spend enough time looking at. I try to raise this issue whenever we discuss filter factors and tonal contrast / separation.

P.S. I'll just add, for what it's worth, that after some initial forays into UV photography, I think panatomic x and type 55 are some of the most UV-sensitive films still in use. I suspect that their sensitivity remains high well down into the 300s. I'll be doing some tests at ~325 in the coming weeks.

Michael, you're talking about the H&D curve, which shows the relation between exposure and density, and I'm talking about the spectral sensitivity curve, which shows the film's sensitivity to different colors of light. Developer choice doesn't change the spectral sensitivity of the film.

Spectral sensitizers can be development and fog inhibitors, desenstizers to blue light, and have a number of other properties. It is the total combination of emulsion, sensitizer and other addenda that make up both the wavelength that a film is sensitive to and the characteristics of that film.

A case in point is a sensitizer that can be a green sensitzer for one emulsion and a red sensitizer for another. It is based on the emulsion, not the dye. A second dye can be a red sensitizer that leaves residual blue sensitivity, and another can repress blue sensitivity. And a third dye can be a super sensitizer giving huge boosts to an emulsion speed but imparting no special spectral sensitivity.

Nothing substantive can be derived from the data you refer to other than the exact sensitivity of the given film itself.

Much data on this is found in the chapter of Mees and James to which Paul Gilman contributed.

PE

I agree with David and I think what is seldom appreciated is that:

- spectral (wavelength) sensitivity can be a major player in tonal separation and contrast

- stated filter factors are quite arbitrary around the edges of the sensitivity spectrum; you have to convolve the lighting spectrum (~colour temp) with the film sensitivity spectrum to get reliable numbers

- no matter what the manufacturer's stated toe and knee look like, you don't know from the H&D curves what actual highlight and shadow curves you're going to get unless you again consider the actual lighting spectrum used for your shot and the film sensitivity spectrum. A pan film can deliver an image that is tonally contrasty or it can be blah flat... with the same developer and the same dilution and exposure time and everything, just by shooting through a colour filter. Extreme example: rollei R3 unfiltered, Rollei R3 red filtered. Please ignore the clouds in the sky, I am not talking so much about the dehazing effect of a red filter, but rather how the wood and grass tones are rendered, which has nothing to do with haze or Rayleigh scattering.

David I don't know if this is what you meant, and I don't mean to put words in your mouth, but when you said tmax's response is too linear for your taste, I took that to mean too linear in overall wavelength sensitivity, which is an issue I see in these R3 test shots. My feeling is that tones simply don't separate well at all unless a filter is deployed.

If a filter selects out that wavelength range that coincides with a big change in the derivative of the wavelength sensitivity curve, then, voila, more tonal contrast. I suppose this is because small changes in the tone/colour across the subject give rise to large changes in actual density in the neg. Requires more thought.

I seem to have more success generating tonal interest and good separation with traditional-grained films, and the tones clump up for me with tmax and the deltas. Maybe it's just me :s

That's exactly what I was after

Photo Engineer said:

Nothing substantive can be derived from the data you refer to other than the exact sensitivity of the given film itself.

PE

Click to expand...

The details of how it is done, while I admire the alchemy, aren't important. But I do enjoy hearing your thoughts, always.

Paul

Awesome. I'm grateful for folks who see more clearly than I,
and write more coherently !

(Many of whom see to be posted ahead of me !)

Here is a visual aid borrowed from York Univ.


And below, here is the late lamented APX 100 data.

df cardwell said:

Paul

Awesome. I'm grateful for folks who see more clearly than I,
and write more coherently !

(Many of whom see to be posted ahead of me !)

Here is a visual aid borrowed from York Univ.

And below, here is the late lamented APX 100 data.

Click to expand...

Thanks for the color graph. There is also one in the UV filter link I supplied, although he has the colors reversed from convention.

Wow, that APX is quite the roller coaster, even more than Kodak's general curve.

When comparing spectral sensitivity it is important how they were derived. There are at least two "standard" ways to plot the graphs. Based on the same data they look completely different.
Compare the spectrae at page 3 for these two datasheets (german and english version of the same film, Ilford Delta 100): http://web.archive.org/web/19980610161333/www.ilford.com/html/german/pdf/100_400_Delta_G.pdf and http://web.archive.org/web/19980610175025/www.ilford.com/html/us_english/pdf/100_Delta.pdf

Nice summary. Thank you for the effort.

The Tri-X look is indeed due to the blue sensitivity mixed with its inherently low contrast, and I have heard that before. I call the TriX look a bit hazy, low in contrast, darker than normal skin tones, darker than normal plants...in short, slightly odd tonal relationships! I have always preferred HP5 for 35mm, and have recently switched in large format at well. I used to almost always filter Tri-X with a yellow-green filter...

More detail would be nice, like being able to see the actual graphs instead of the writer's description. Raw information is much more educational to me than a running commentary trying to achieve the difficult task of explaining a graph that is not there.

One thing irked me. I want to hear a more detailed explanation of why it is a fact that UV filters are necessary with the more UV-sensitive films...or was that just an opinion?

Opinion. It would also depend on where you are shooting, of course. Not so much here at sea level.

ath said:

When comparing spectral sensitivity it is important how they were derived. There are at least two "standard" ways to plot the graphs. Based on the same data they look completely different.
Compare the spectrae at page 3 for these two datasheets (german and english version of the same film, Ilford Delta 100): http://web.archive.org/web/19980610161333/www.ilford.com/html/german/pdf/100_400_Delta_G.pdf and http://web.archive.org/web/19980610175025/www.ilford.com/html/us_english/pdf/100_Delta.pdf

Click to expand...

Interesting. Other than I just stuck with English language data sheets, I don't know what to say. Why would Ilford use and publish two different methods?

Paul Verizzo said:

Another project from the depths of the lah-bor-a-tree on the bayou.

Click to expand...

Paul - I'm kind of disappointed - I was getting excited as I thought you had tested each of these films yourself, instead of collecting a bunch of links to somewhat related graphs of spectral sensitivity...

I think your idea does hold some merit, but it looks like more work needs to be done to make a good comparison.

Did you notice that not all the spectral graphs were made with the same color temp of light source? That could explain some of the "weird" behavior from the Foma. The Foma was tested at 2580K, while Fuji was at 5400K, Ilford was also at 2850K, and who knows what temp Kodak used - I can't see where they've stated it... Also, Foma did not show any scale on the y-axis. Perhaps they used a linear scale when the other 3 show that they are using a log scale. A linear scale would exaggerate small differences in response while the log scale would compress large differences.

If you really want to take this to the next step (without going out and buying/making a recording spectrograph), you could get some software called Graphula (I think it is freeware) which lets you take a graph like you have from these companies, and then trace the plot on them, and it will generate a set of x,y data from it. Then, you can use info for blackbody radiation at 2850K and 5400K and then be able to convert the graphs that you've presented into graphs with the same color temp.

Now that would be cool to see!

Paul Verizzo said:

Opinion. It would also depend on where you are shooting, of course. Not so much here at sea level.

Click to expand...

My point is that this is entirely up to the shooter, and how he or she wants the film to respond to UV. It can't be made as a definite blanket statement of fact.

Something like, "in environmental circumstances in which it is desired to reduce blue sensitivity of Tri-X and other Kodak films, a UV filter can be employed. These situations might include..."

Thanks for some great observations!

Kirk Keyes said:

Paul - I'm kind of disappointed - I was getting excited as I thought you had tested each of these films yourself, instead of collecting a bunch of links to somewhat related graphs of spectral sensitivity...

I think your idea does hold some merit, but it looks like more work needs to be done to make a good comparison.

Did you notice that not all the spectral graphs were made with the same color temp of light source? That could explain some of the "weird" behavior from the Foma. The Foma was tested at 2580K, while Fuji was at 5400K, Ilford was also at 2850K, and who knows what temp Kodak used - I can't see where they've stated it... Also, Foma did not show any scale on the y-axis. Perhaps they used a linear scale when the other 3 show that they are using a log scale. A linear scale would exaggerate small differences in response while the log scale would compress large differences.

If you really want to take this to the next step (without going out and buying/making a recording spectrograph), you could get some software called Graphula (I think it is freeware) which lets you take a graph like you have from these companies, and then trace the plot on them, and it will generate a set of x,y data from it. Then, you can use info for blackbody radiation at 2850K and 5400K and then be able to convert the graphs that you've presented into graphs with the same color temp.

Now that would be cool to see!

Click to expand...

Working backwards, no, I'm done. I did notice a few of the points you mention, and mentioned the lack of Foma X axis info.

As to color temperature, I'm sort of scratching my head on that. How could you get strong UV/violet measurements from a 2800 K source? If you had some theoretical perfectly flat response film, you would still have a curve showing the light source, wouldn't you?

And panchro films are really geared to daylight, some even having a different tungsten rating.

I'd also like to hear more thoughts on the Foma curve shape. How intentional? Why?

Paul Verizzo said:

As to color temperature, I'm sort of scratching my head on that. How could you get strong UV/violet measurements from a 2800 K source? If you had some theoretical perfectly flat response film, you would still have a curve showing the light source, wouldn't you?

Click to expand...

Are you saying the Kodak data for Tri-X was made with a 2850K light source? How did you determine that? It does not say what color light source was used...

I'm sure Fuji used 5400K since they make color film, I bet they use the same instrumentation to test their B&W films. Ilford, only makes B&W film, and 2850 is a fine temp for that, as long as you realize you it will show an increased response to red and a decreased response to blue (and little response to UV). Foma, again, B&W only, if I remember right. Kodak, perhaps, since they make color film, is using the same equipment to test B&W film as they use for color. That would mean they tested it at around 5400K or 5500K. That would give more blue and UV light in the test, which could be shown in the graph as a "greater" response to blue and UV.

THat's why I suggested that the data needs to be converted to one color temp for all graphs (and one scale for all graphs too), otherwise they comparisons really can't be make at the level that you are trying to do. At least they can't be made with much accuracy...

Foma curve shape is probably an artifact of the way the data was graphed combined with the 2850K light source.

Theoretically flat response - at what color temp? You will get different responses at different temps. I guess shooting RAW in digital and then doing a color balance for individual lighting conditions will act as a flat spectral response. But I don't think you can make film to behave that way.

Warm lighting will raise the red response, and cooler lighting will raise the blue response...

I am also interested to see the tests *all* done with daylight equivalent temp (somewhere in the 5000s, depending on location and season), photoflood equivalent temp (3200-3400K), and household lamp equivalent temp.

I would imagine that if we were to contact the manufacturers asking what color temps light they used to determine the spectral sensitivity of their b/w films, they would tell us.

As far as I can tell, everyone but Kodak listed the temp. Perhaps PE knows the practice there?

Am I reading correctly that the spectral sensitivity that is reported is a convolution of the lamp spectrum and the actual sensitivity of the film? And these spectra are not corrected for the lamp spectrum?

I am a bit surprised that these spectra aren't lamp-corrected. At least in spectroscopy, the standard practice is to take a spectrum, and then to correct it for the lamp spectrum and other instrumental factors and background. In other words, the spectrum we report corrects for uneven weighting of the illumination source as a function of wavelength- no lamp is perfect. This lamp-correction procedure is easy enough if the illumination spectrum is smooth and broad, as it must be if the colour temp is typically defined.

keithwms said:

Am I reading correctly that the spectral sensitivity that is reported is a convolution of the lamp spectrum and the actual sensitivity of the film? And these spectra are not corrected for the lamp spectrum?

Click to expand...

Where the temp is listed, we have to conclude the test was done at that temp. So, yes, it is a convolution of the lamp and the film.

Keithh - When one corrects for the lamp, would that be called an "equal energy" spectrum?

Wedge Spectrogram

Here is a wedge spectrogram. The visible spectrum in nanometers is the x axis and the Log E values in 0.3 increments is the y axis with increasing speed moving upwards and increasing exposure downwards.

This is a green sensitive material. The peak in the red region is a harmonic of UV light, as this was done with a hand coating on a spectrosensitiometer with no UV filter.

As you can see, the visual slope varies as a function of wavelength in this case. If it does (It does not sometimes and does other times depending on dye and emulsion), the singe line curve in the example above and shown on Kodak's site are not revealing. You have to construct a curve at each wavelength from the entire exposure such as shown here in the data I have included.

The line drawing of spectral sensitivity is just an approximation unless you know the full curve.

PE

PE, when you showed us one of those in your workshop here in New York, I finally understood why a safelight works.

The best prevention in B&W against blue and UV is to use yellow filter (result of red and green primary color), if you desire is to achieve the best contrast on your film.

IMO, it's not always the case.