Ok, so I had no need to complicate that formula above because I just noticed that those dev times in fact follow the simple geometric series t(n+1)=t(n)xSQRT(2). Which doubles every 2nd point, just like f-stops do !!
Anyway my initial exponential curve fit was pretty darn close and it still seems I will need to go out to about 32min dev time.
Did my step wedge suffer from ambient light reflections during film testing ?
Well folks, I think I've got a problem. I've completed all the transmission density measurements of the 5 rolls of HP5plus 120 film developed in XTOL 1+2. You can see photos of my setup here.
I've entered the raw data from my (frequently calibrated densitometer) into the spreadsheet supplied by Ralph over 5 months ago (and re-attached in xls format with my data) and I have two main concerns. Firstly though is a screen shot giving a comparison of my Family of Curves LHS to the default data in the spreadsheet (RHS) which was from Ralph's testing.
My first concern is that the lowest densities I get (in the toe) are way above 0.05, and none of them overlap in the toe area. According to my interpretation of the instructions posted by Ralph and WBM, they need to go below 0.17 so to determine the effective film speed (Zone 1.5=0.17). The fb+f density ranges from about 0.1 to 0.15, so I think this means the film was fixed adequately.
So why do I have too much transmission density in the toe ?
The other concern I have is that the highest density measured is only 1.5, whereas Ralph's was 2.1 It would seem I'm missing a good chunk of dynamic range here.
Should I expect a greater transmission density than 1.5 for a relative log exposure of 3.0 ?
The possible explanation for the added toe density might be reflections from ambient light as I did this outside. I did shield one side of the path from camera lens to the transmission step wedge using opaque card to hopefully eliminate reflections. The side I shielded was the side the sun was on. I know Lee warned me about reflections, but is this truly the side effect, or is this response typical of my film+dev combo ? I did use a lens hood to reduce flare.
PS Ralph. I hope you are recovering OK. Take it easy if that's possible.
Last edited by PeterB; 07-09-2011 at 10:11 AM. Click to view previous post history.
Reason: additional info
There looks to be two things going on with your test. First, you have under exposed the step tablet. That's what is causing the large flat line. You have yet to overcome the inertia of the film. Second, you really need to contact the step tablet. It eliminates many variables and makes it easier to track down problems. In this case, the problem looks like flare. The flat line portion should equal the Fb+b as there hasn't been enough exposure to produce density differences. The density above Fb+f has to come from non-image producing exposure. A light leak is one possible answer, but based on the description of the test, the large white background is at fault. Contacting eliminates flare. There's a reason why things are done a certain way. Don't try to reinvent the wheel.
Stephen's right, there has to be tons of flare in your setup.
The difference between tungsten and daylight on film has got to be far less than the difference between your graph and ideal. You may be working too hard to rule out a variable that is significant, but not as gross.
At the very least, build a dark chamber for the area in front of the camera and mask out the step tablets with black construction paper to minimize flare. But I would second Stephen's recommendation to make your tests with the step wedges in contact with film.
Stephen and Bill, thank you for your suggestions. I totally forgot about the flare from the large white card I used. I was only considering flare caused directly from the sun !! I won't make that mistake again.
If I contact expose the step wedge with my film it will be fiddly as I'm using 120 roll film not sheet film. Also won't that eliminate any inherent contrast reduction caused by my Mamiya's 80mm lens? I assumed one of the points of this test was to incorporate one's equipment in the light path.
What do you think about the idea of using a light box and taping the step's border with black to eliminate flare?
My original reason not to use it was its unknown colour temperature, but as Bill points out that is minor compared to the other problems I introduced.
When I expose the step wedge again I will increase the exposure. I had followed closely all the instructions but at least now I have a baseline to help me work out how much more exposure I need.
Sponsored Ad. (Subscribers to APUG have the option to remove this ad.)
It'll be better but you'll still have flare. The clear steps will light up the whole works giving the denser steps some illumination.
Testing the camera system including flare was often taught in Zone System workshops. There is great value to be had learning the effects of exposure and development in a short time. I have a hunch you may have just learned that lesson.
The practice of camera system tests is often discouraged these days because it includes too many variables and you won't know what the results mean.
Film tests in contact with test strips will get you the response of film to light and development. Even if you do not have an accurate sensitometer, with a consistent exposure you'll get good graph shapes. The graph that matches the ASA gradient will give you the benchmark for speed, you can look down at 0.1 density and work backwards (trusting the manufacturer's box speed).
This will give you numbers to put in Ralph's spreadsheet for your film and development technique and times.
You can test for flare separately from the test for film speed. The book Beyond the Zone System describes a black box. Basically a cardboard box painted black inside with a hole in front. A coffee can painted black is another one. The idea is you shoot this in your typical scene and theoretically it should be perfectly clear on film because it is pure black. But flare will give it some density.
You kind of already did a flare test.
Masking off any stray light from a light box will improve the results, but basically you are testing for too many factors. If you want to test the film, test only the film. If you want to test the lens, test only the lens. If you have too many variables, it's near impossible to attribute the influence of any particular variable.
Originally Posted by PeterB
There isn't a single flare factor for a given lens. 80% of flare comes from the subject and it varies depending on the luminance range and tonal distribution. A small black area with a light surround will have greater flare than visa-versa even though they have identical luminance ranges. So it isn't really beneficial to do a film test using a camera if testing for flare is your goal. The results will only reflect that specific situation and it will limit the usefulness of applying the data to other situations.
A key reason to contract the step tablet is to eliminate flare from the mix so that you can ascertain the proper film characteristics. Flare is then factored in at the evaluation stage. For instance, the flare factor for a statistically average scene is around one stop to 1 1/3 stop with modern lenses. If you are attempting to determine the film density at a seven stop scene luminance range you simply take the seven stops and subtract flare to give you the illuminance range to use when evaluating the film curve. For one stop of flare, the film effectively has a six stop illuminance range for a seven stop luminance range.
Conceptually, you can think of testing as having two different kinds of exposure. There's the sensitometric exposure that produces the characteristic curve and the photographic or camera exposure that juxtaposes the luminance values of the subject on top of the film curve and is what we use to evaluate the results for a given set of conditions. Film speed and camera exposure of EI should also be viewed in a similar manner.
I've attached an two examples to help illustrate this concept. One example has a one stop flare factor and the other a 1 1/3 stop flare factor. Notice how the subject's values are changed in the camera (Quadrant 1) and are placed on top of a fixed film curve (Quadrant 2). The straight line in Quadrant 1 represents a zero flare model and you will notice how the exposure would fall further into the film's toe without the presence of flare. The resulting negative density ranges from the two examples will be different, but in neither case does the film curve change.
Last edited by Stephen Benskin; 07-10-2011 at 01:59 AM. Click to view previous post history.
I actually did a similar test once using the Davis black box. I cut out four different squares to surround the hole in the black box and made four identical exposures. The squares were white, black, gray, and a mixture of all three. I've attached the results.
Thanks again Bill and Stephen. That dreaded flare rears its head everywhere ! Before I read your last posts I was thinking to myself, how on earth can I hope to properly expose and develop a scene with lots of light areas in it (either in or just outside the image) if that is going to contribute to contrast reduction through flare. Clearly as you point out I need to test for that separately with my (non modern, (single coated ?) lens) as well as make assumptions/adjustments based on how my scene's brightness levels are weighted/distributed.
I will do a bit of reading up on doing the contact method, I think the main thing I need to determine now is how much exposure to give it and by what method (as I won't have a shutter to accurately time anything. The exposure must be at least 5-10 seconds if I use my enlarger timer. I was thinking of using that timer to switch a the light from a distant globe in the same room perhaps (my enlarger light would be way to bright for that many seconds worth of exposure).
If you don't want to contend with reciprocity failure, you need to keep the exposure under a second.
BTW, a good way to calculate aim CI is:
desired negative density range / (log subject luminance range - flare)
Example: 1.05 / (2.2 - .40) = 0.58
Also, flare tends to increase as the scene's luminance range increases and decreases as the scene's luminance range decreases. It averages a 1/3 stop increase or decrease in flare per one stop change in luminance range.