H-D Curve Procedure

My process for creating an H-D curve.

Check Setup

Expose

Develop

Develop using normal procedure or adjusting times or temperatures depending on the goal the test. About one hour dry time is sufficient before moving on.

Measure

Compute and Plot

Go to hdcurve and select the two files and click “Compute and Plot”. The H-D curve will be plotted and the contrast curve will be calculated.

Notes

This is not an extremely scientific process but is sufficient for (1) learning, (2) getting a somewhat quantitative feel for how much affect various changes such as temperature, development time, developer, etc. have on contrast, curve shape, and relative speed, (3) confirming that my developing process is relatively consistent. There are probably things I’m doing “wrong” or that could be improved. Anyone reading this, please add comments with suggestions. My main purpose of documenting this is for my own reference for a repeatable process, but hopefully it is useful or at least interesting for others.

Determining Max Exposure

What follows is a description of how I determined the target max exposure and corresponding enlarger settings and then corrected my exposure measurements with the baseline assumption that the film I’m testing is exactly ISO 400. It also describes how I came up with the 7.9 lx reference measurement that I use for checking and adjusting my setup for subsequent tests. I know the assumption that the film is exactly ISO 400 is faulty but it is the best I can do lacking a truly calibrated sensitometer.

I used the definition of ISO speed from ISO 6 - 1974 to determine the illuminance of speed point H_m for ISO 400 speed film to be 0.002 lx*s or logH_m=-2.70 using S=0.8/H_m with S being the ISO speed of 400. This gave me a reference exposure for determining a target base exposure.

From looking at the H-D Curve from the T-Max 400 datasheet, I knew I wanted my minimum to be about 0.8 logH units lower than the H_m point so I chose a minimum target logH_min of -3.5.

My step wedge has 31 steps with a density range of 0.05 to 3.05 so the the logH_max (for part of film not covered by the step wedge) will be -3.5 + 3.05 = -0.45. Therefore H_max = 10^(-0.45) = 0.355 lx*s.

Also, I knew I wanted to target an exposure time of less than one second to avoid reciprocity failure so I chose 0.5s. That meant I needed 0.355/0.5 lx illuminance = 0.701 lx at the H_m speed point.

Based on the filter and attenuators I had, I determined with some trial and error that a -3 EV attenuator combined with the 80A filter at 29.5cm for 0.6s would result in 0.4407 lx*s or logH_max=-0.356. This was close enough to the target -0.45 logH_max from above.

I measured the illuminance without the attenuator to be in a range in which my lux meter was more sensitive and then subtracted off the delta logH of the attenuator. Also I had previously calculated the actual delta logH of the attenuator (and 80A filter) at higher illuminance levels (it wasn’t exactly -3 EV). In all of this, I’m making the assumption my lux meter was fairly linear even if the value of the reading was not exactly accurate.

I then put the enlarger in the final configuration with the exception of the -3 EV attenuator and set the lux meter on top of the proof printer glass and took a measurement. The only purpose of this measurement is that it is a convenient configuration for checking my setup and fine tuning the aperture to a known illuminance for future tests (this is where I got the 7.9 lx number).

This got me in the ballpark to ensure I captured the important part of the H-D curve, but I was under no illusion that this would be perfectly accurate. There are many factors such as the accuracy of my lux meter, the accuracy of my timer, the ramp up time of the enlarger light, etc. that could affect the actual exposure. I am however pretty confident that if I setup the enlarger the same way each time, I’ll get a consistent exposure within a reasonable margin of error for my purposes.

I then performed several tests at different exposure times and then interpolated the logH_m that would result in a 0.8 to 1.3 slope required by ISO 6 for determining ISO speed. Using this logH_m and the expected logH_m for ISO 400 film of -2.70, I determined a calibration offset of -0.173. Once again, an assumption was made that the test film was exactly ISO 400 speed film.

H-D Curve Online Plotting Tool

Screenshot of hdcurve tool At first, I just used excel to plot the H-D curves and then printed them out at the correct aspect ratio (X and Y axis scales the same) and used the graphical method of determining contrast index from the Kodak Sensitometry Workbook. However, just for fun, I had ChatGPT write me a single page HTML/Javascript app to do the same. The result is this tool Although I do software development as part of my day job, I have not closely reviewed the resulting code. This was more an experiment in vibe coding, but I did cross check the results with my curves and contrast indexes that were plotted and calculated more or less manually and they agree. Here is the ChatGPT session.

A couple of test files:

What I’ve Learned So Far

My initial experience with XTOL and T-Max 400 has been that it is low contrast when developed per the T-Max 400 datasheet. This was before doing any H-D curve testing. The H-D curves confirmed what I saw and showed that I needed significantly more development time to get reasonable contrast. With 7 minute development time at 21.5C which is 1 minute more than the 6 minutes nominal time at 21.5C (interpolated between 6.25 minutes for 21C and 5.75 minutes for 22C), I get a contrast index of about 0.49 whereas the datasheet shows I should get more like 0.56 even at 6 minutes. I found that I needed about 8.25 minutes to get the nominal 0.56 contrast index at 21.5C. Its possible my thermometer is inaccurate, but I crossed checked it with a new calibrated digital thermometer and it matched within 0.25C. Similarly, my agitation method could be different in some way, but I followed the T-Max 400 datasheet instructions and even attempted to be on the more vigorous end of the ranges they gave (7 inversions in 5 seconds every thirty seconds). So really, I’m at a loss to explain why my time is so much higher. It is interesting to note though that my findings agree with Tim’s recommended time here, actually he is recommending 8.5 minutes at 20C so he is recommending higher times than what I’m finding for nominal (he doesn’t talk about contrast index though so I presume it is just based on his experience). This photrio thread seems to come to similar conclusions as me, at least to the extent “conclusions” can come from a forum thread.

Perhaps the Kodak datasheet assumes using stainless steel tanks and spools which I understand result in a little more aggressive agitation. I find it hard to believe it makes that much difference though.

Also possibly it has something to do with my particular batch of XTOL or the age of it combined with how I have stored it. It was about seven weeks old when I performed these test and was stored in mostly airless containers but not completely. I’ll be interested to perform tests with freshly mixed developer.

The other thing I’ve noticed is that the H-D curves have an upswept shape where the slope increases at higher exposures. They probably do shoulder at some point, but I did not hit that high of density in my curves. This increasing slope area moves right as the development time increases. So at low development times, the contrast index understates how low the contrast is for the linear part of the curve. This combined with the low contrast from the Kodak published development times may be why T-Max 400 gets a reputation as “low contrast”. Once you get to moderate contrast, this part of the curve has moved right far enough that it no longer factors into the contrast index measurement or has a practical impact except maybe increased contrast in the lightest highlights which may or may not be perceptible.

Summary

I’ve enjoyed the nerd factor of testing film and getting a better feel for how my development process affects my negatives. I caution anyone reading this that this is my first experience with this type of testing and should not be interpreted as “the way”, but hopefully it will prove useful to someone. I will probably do further testing in the future to further refine my process and and results. Besides just pure intellectual curiosity, one of my primary goals of this testing continues to be to try to eliminating as many random variables from my photography process as possible. There are so many variables that cannot be controlled that I find it valuable to try to control or at least understand the ones that I can.

Equipment and Supplies