BTZS five sheet testing

[Paul Butzi]

One thing I find interesting is that BTZS uses five sheets to produce a time/contrast curve.

The method I settled on uses three sheets, with development times spaced widely around a 'best first guess' (so that I'm interpolating rather than extrapolating). I then fit a cubic polynomial to those points, and use the resulting curve. I suppose I really ought to fit a cubic spline, but I'm lazy and the cubic polynomial seems to produce results that are sufficient to my needs.

So I'm left wondering - if I were to use the 5 sheet method, would I see inflections in the time/contrast curve that I'm not exposing with my 3 sheet method?

Or are the curves pretty darn smooth, and using 5 sheets just reduces the need to generate a good initial guess and thus reduce the chance that I'll have to do more runs because my initial guess was not very good (this has, in fact, happened to me).

[clay harmon]

It depends on the film and developer combination. Some respond in a very linear fashion, which means your 3 sample appproach would be fine. Others do strange things like lose film speed only at SBR's greater than 10, for instance. IOW, in some cases, three samples would be plenty, in others, probably not...

[Michael Jones]

Paul:

Phil used to use 4 sheets and then went to 5 sheets. I suppose you could use 6 if you wanted and there is not reason you could not use 3 sheets. To an extent there is always interpolating and extrapolating to create the charts. The difference is the amount of data to draw from. The more data the smoother the curves and theoretically, the "cleaner" and more accurate the end results. The bottom line is that if you have a system that allows you to work, use it. Good luck and enjoy.

[Ron Marshall]

Paul, I did five sheets for one of my film-dev combinations. I then constructed curves two ways: using all five data points; using the two end points and the midpoint (as if using three sheets). The curves were not signifigantly different. Of course for some film-dev combinations this may not be the case.