Not looking to replace C-41 with any of this, ideally what I want is two things from two different formulations:
1) A large speed losing contrast increasing developer (would like the same amount of stops of range on the neg in a straight line but a steeper line.. ie reaching a higher dMax on the straight line), to only develop the fine grains and perhaps provide a solvent effect .
2) The extreme other end. As useful as possible for underexposed images.
Ignore the 'ideal' results, and think of the C-41 measured results as the ideal results, and compare those with the split-bath measured results, that shows hue shifting from the C-41 results.
I will see what the gamma for the red, green and blue parts are to balance the grey scale vs the C-41 charts.. that should also be telling. If the gamma/contrast of one part (say green image) is significantly lower than the others for example, you may have prints with either green shadow cast or magenta highlight cast depending on printing.
It does not last 'forever', last time I tried it a while ago, I found it stopped working after 6+ months, I don't know exactly when as I didn't use it very often (only a couple of rolls). I also mixed it up with tap water and no chelating agent.
The other issue is iodide and bromide build up (in both bathes) and weakening of bath A from some development occurring, and dissolved silver build up from my addition of thiocyanate.
Which may change developer characteristics with usage, my recommendation would be to replenish an arbitrary amount - it should theoretically self-balance to an equilibrium point.
The preferred Bath A I would like to use, is with CD-4 with enough alkali to neutralise the H2SO4, and then HAS to set the desired pH, maybe with a weak alkali buffer to get an extra amount of HAS for better keeping properties.
It is still quite economical, regardless of that, Flexicolor is also very economical, the cost is generally all in the bleach. One can be very economical in C-41 one-shot if they keep and replenish the bleach (and fixer too I suppose).

Adv Reply