Testing and evaluating KODAK T-MAX P3200 and Ilford Delta 3200

[bernard_L]

I finally got around to developing KODAK T-MAX P3200 and Ilford Delta 3200 in XTOL.

<<SNIP>>

So does XTOL have magic? I would say, with T-MAX P3200, there's some magic to it. Yes.

Now I am really curious how seasoned XTOL will perform.

View attachment 322757 View attachment 322759

Confused by the graphs. They mention D-76 while the current topic seems to be Xtol. And the two curves (blue, red) are labeled with identical dev times: 16min. Makes it difficult to mesh curves and discussion.

I must be missing something, since the discussion goes on without anybody mentioning this...

 

[Anon Ymous]

Confused by the graphs. They mention D-76 while the current topic seems to be Xtol. And the two curves (blue, red) are labeled with identical dev times: 16min. Makes it difficult to mesh curves and discussion.

I must be missing something, since the discussion goes on without anybody mentioning this...

You are not alone...

 

[aparat]

Confused by the graphs. They mention D-76 while the current topic seems to be Xtol. And the two curves (blue, red) are labeled with identical dev times: 16min. Makes it difficult to mesh curves and discussion.

I must be missing something, since the discussion goes on without anybody mentioning this...

Yes, sorry about that. I will fix that. They were all developed in XTOL and are incorrectly labeled as D76. The times are all 16 minutes because all four curves represent films developed in the same tank. They show how the film responds to being exposed at EI 1600 and EI 3200, respectively.

 

[bernard_L]

They show how the film responds to being exposed at EI 1600 and EI 3200, respectively.

Do we expect other than a shift of 0.3 along the X-axis? That is whatallows one to perform sensitometry without needing to know (at least not accurately) beforehand the ISO rating of the film under test. With deviations giving some measure (not fully foolproof) of the global reproducibility of the process. And indeed T-MAX P3200 shows close to 0.3 X-offset between the two curves.

But not so for Delta 3200. At that point one of the more savvy forum members will point out that dumb me overlooked reciprocity failure. Didn't go back and proofread the 100+ previous posts to find (or not) the relevant data; I more or less remember that you (aparat) were careful to have an exposure time that would be safe from long (and short) times conducive to reciprocity failure. And if reciprocity failure would be significant then the product millilux.seconds would not be a proper measure of the exposure: illuminance and time would need to be separately specified: two independent variables. Not that I suggest actually doing that; parameter space is already uncomfortably large (as in: N-dimensional).

P.S. aparat, keep going with this nice work. Reliable data is rare here.

 

[Sirius Glass]

Yes, thank you for pointing that out. That's what I do in my tone reproduction function, which can simulate different camera exposure, vary curves by CI and LSLR, and align (or key) curves to either highlights or shadows. Here, I was going to switch to XTOL, anyway, so I thought I'd make the exposure more realistic to how people are using the film. Most of all, I was curious how the film would respond. After all, it's got "3200" on the box. A lot of people love XTOL for its suitability for pushing film and for rotary processing. Plus, people rave about its tonality. So far, I am impressed, but it's still early days.

I found that not only is XTOL a fine grain developer the tonality is much smoother and the developer is very forgiving. Mistakes in pouring too slowly or improperly or mistakes in timing do not cause the film to have problems. I recommend it to people who are just starting out for those reasons.

 

[Steven Lee]

Xtol is a good developer. But having seen Delta 3200 in Microphen I would never advise using Xtol for this film.

 

[aparat]

It did seemed a little strange to me.

I think you're right about plotting a simulated underexposure from existing data, i.e., without running another film test. As I mentioned, I do this as part of tone reproduction analysis, but, based on your advice, I might implement it in my curve family plotting function.

This is not exactly what you meant, I know, but I think it does provide a useful visualization of a one-stop underexposure, Here, the curve of KODAK T-MAX P3200, developed in D76, is showing a one-stop underexposure. The red curve is at ISO 1600 exposure, and the blue one at ISO 3200. The speed point densities are -0.20 and 0.10, respectively. Both curves are developed to the same Average Gradient (G) of 0.6, which could also be changed to simulate push processing.



However, I still see value in switching to XTOL and exposing the film for one or two stop less exposure than the previously established baseline. I think this represents a more realistic scenario for a lot of photographers. Plus, I discovered that XTOL produces more linear curves than D76. I will abstain from making any other claims until I have more data.

 

[Stephen Benskin]

I think you're right about plotting a simulated underexposure from existing data, i.e., without running another film test. As I mentioned, I do this as part of tone reproduction analysis, but, based on your advice, I might implement it in my curve family plotting function.

This is not exactly what you meant, I know, but I think it does provide a useful visualization of a one-stop underexposure, Here, the curve of KODAK T-MAX P3200, developed in D76, is showing a one-stop underexposure. The red curve is at ISO 1600 exposure, and the blue one at ISO 3200. The speed point densities are -0.20 and 0.10, respectively. Both curves are developed to the same Average Gradient (G) of 0.6, which could also be changed to simulate push processing.

View attachment 322852

However, I still see value in switching to XTOL and exposing the film for one or two stop less exposure than the previously established baseline. I think this represents a more realistic scenario for a lot of photographers. Plus, I discovered that XTOL produces more linear curves than D76. I will abstain from making any other claims until I have more data.

You admit the CI needs to be changed in order for the example to really illustrate "pushing for speed." What you have is one curve shifted Δ0.30 log-H to the right. You could have just changed the values on the X-axis with the same curve and no one would have known. If you were using actual log-H values the two curves would overlap. Their response is the same for the same exposure value. Drawn correctly, the blue curve would start further to the left.

What the example is actually communicating is the blue film has a film speed one stop slower than the red, not that the exposure was different.

The curve looks a little steep as your X & Y axis aren't proportioned the same. Not that it's wrong. It's potentially visually misleading if you consider a gradient of 1.0 is a 45 degree angle. Mostly nit-picking but important details to consider.

 

[aparat]

You admit the CI needs to be changed in order for the example to really illustrate "pushing for speed." What you have is one curve shifted Δ0.30 log-H to the right. You could have just changed the values on the X-axis with the same curve and no one would have known. If you were using actual log-H values the two curves would overlap. Their response is the same for the same exposure value. Drawn correctly, the blue curve would start further to the left.

What the example is actually communicating is the blue film has a film speed one stop slower than the red, not that the exposure was different.

The curve looks a little steep as your X & Y axis aren't proportioned the same. Not that it's wrong. It's potentially visually misleading if you consider a gradient of 1.0 is a 45 degree angle. Mostly nit-picking but important details to consider.

Thank you. That's very useful. I agree that details matter.

I actually only wanted to show the degree of underexposure by one stop, so people can see how the 0.1 over B+F speed points are 0.3 apart. I agree that the blue curve looks like it was just "moved" but, in fact, both exposure and density values were recalculated for each curve based on a model. The problem is that, visually, it's not that clear without a corresponding change of the Average Gradient. It would have been better to also show increased G, yes, I agree. Right now, I can only generate different slopes based on G, not CI, by the way. Maybe at some point, I will implement CI. I think G is accurate for a rough kind of analysis. I do compute CI for curve analysis, but not for curve synthesis.

Next time, when I present the XTOL data, I will show a progression of curves with different parameters changed. Thanks again for your input. I appreciate it very much.

 

[aparat]

I am still waiting for my XTOL to become seasoned. It's going to take some time. Meanwhile, I put together a simple demo of push-processing KODAK T-MAX P3200 in D76. The model predicts how underexposure and overdevelopment cause some small increase in film speed, with a simultaneous increase in contrast and a reduction in shadow detail. There's also an increase in base fog density, as the development time increases. I found that the P3200 has higher B+F density than most 35mm films I've used over the years, except for expired films, so that had to be included in the model. This is not real data so it probably won't be super accurate, but I think it conveys the general idea.

Total speed is predicted to increase by around 0.9 of a stop, whereas Ḡ increases from around "normal" by 0.32. It will be interesting to see if this bears out with real data.

 

[Paul Howell]

So how does Tmax developer or DDX compare to Extol? Other option for full ISO is DK50 stock or perhaps 1:1.

 

[aparat]

So how does Tmax developer or DDX compare to Extol? Other option for full ISO is DK50 stock or perhaps 1:1.

I am considering either TMAX Developer or DD-X, but, first, I am going to run it in replenished XTOL. I took what @Sirius Glass said that replenished XTOL "is where the rubber meets the road" seriously, but I also looked through the forum, and there seems to be a lot of consensus that XTOL-R is one of the best developers for push processing films, especially the modern emulsions, T-MAX and Delta.

Paul, I am not familiar with DK50, I think I have all the chemicals needed so, in theory, I could try it. What is it about DK50 that you like?

 

[aparat]

Sorry it's taken such a long time, but I had to be sure my XTOL-R was properly seasoned and stable. I ended up running the equivalent of twenty one 36-exposure 35mm films before it fully stabilized. I wonder if that's typical for you guys. I replenished at the rage of 70 ml of stock/replenisher for every 36-exposure film, testing with a step tablet, exposed in a calibrated sensitometer, along the way. It was relatively stable after about ten films, but, in my experience, it only fully stabilized after about twenty.

I ran a test of KODAK PROFESSIONAL T-MAX P3200 and Ilford Delta 3200, developed in XTOL-R, simulating a typical "pushing" scenario, i.e., exposing for about 5.2 stops less exposure than for a typical ISO 100 film, and developing for Kodak's recommended seventeen minutes in a Jobo (1510 tank, 150 ml of developer) at 20C. I ran the test four times, after which I decided to stop because I was getting virtually identical results, telling me that the process was stable. The differences among individual curves were within the margin of error of my densitometer. Here are images that show the overview of the curves obtained in the four trials:


So what are my conclusions? Probably nothing new to advanced users, but perhaps this will be useful to beginners:

• KODAK XTOL-R can be a stable, consistent developer.
• Compared to D76, XTOL-R requires more care and a specific replenishment routine to make sure it is stable.
• Compared to D76, XTOL-R is cleaner-working, at least with those two films.
• Compared to D76, XTOL-R produces significantly less "lumpy" curves with both films, but particularly with the P3200.
• Compared to D76, XTOL-R produces more linearity through the mid-tones and highlights (see the discussion of tonality here).
• Compared to D76, XTOL-R gets slightly more film speed (manifested in a slight gain in shadow detail).
• XTOL-R produces predictable, manageable contrast with Delta 3200 in the seventeen-minute development. I can see why people love this combination.
• XTOL-R produces high contrast with the P3200, so one needs to be careful to not blow out the highlights, though the slight highlight compression might help here.
• While pushing the films does not result in a magical gain in film speed, the two films seem engineered specifically to produce great results in this particular scenario. In my experience, Ilford Delta does that a little better, though, admittedly, the P3200 has inherently more speed.

I have a lot of data from these tests, but for the sake of brevity, I am going start by including one-page summaries. Please, let me know if you have any questions. I'll be happy to include whatever details you need.

 

[MattKing]

Did you start replenishing immediately, or did you wait until you had developed several films before you started replenishing?
The purpose of running the initial several films through the developer first, without replenishing, is to speed up the build-up of development byproducts.
Can you remind us what the size of your working solution container is?

 

[aparat]

Did you start replenishing immediately, or did you wait until you had developed several films before you started replenishing?
The purpose of running the initial several films through the developer first, without replenishing, is to speed up the build-up of development byproducts.
Can you remind us what the size of your working solution container is?

Sure. I made a five-liter batch, and decanted it into five one-liter bottles. One, being designated my working solution, the remaining four, the replenisher. I started by developing three 36-exposure films, exposed for Zone VIII, with a few inches devoted to the step tablet and DMAX. Then, I ran two more films with actual photographs, so varied densities, and poured the developer back. Then I started developing and replenishing. Perhaps my problem was using very small quantities of XTOL (150 ml to 250 ml), rather than 500 ml or even 1000 ml, which one would need with inversion agitation? That should not matter, I think. Maybe it's because rotary processing causes more oxidation? After about 10 films, I got good stability (CI within 0.1), but not good enough for a sensitive test, like the one I had to do here. Most of the film that I used for this initial run was Ultrafine Finesse 100, if that makes any difference. Thanks for your help with this!

 

[MattKing]

Thanks.
I don't have and almost never use a densitometer. Usually I find that about six rolls before replenishing is enough for "stable enough", but that is for practical use and actual photographs. My working solution volume is 2 quarts - because that is the size of the container I use.
By the way, for your purposes you should probably standardize on a single volume of developer used, because varying the quantity of developer used but keeping the volume of replenisher the same makes a tiny difference to the arithmetic!
I deal with that issue by always using the full volume - 1 litre - in my Paterson inversion tank. With replenishment, I don't have to worry about wasting developer!

 

[aparat]

Thanks.
I don't have and almost never use a densitometer. Usually I find that about six rolls before replenishing is enough for "stable enough", but that is for practical use and actual photographs. My working solution volume is 2 quarts - because that is the size of the container I use.
By the way, for your purposes you should probably standardize on a single volume of developer used, because varying the quantity of developer used but keeping the volume of replenisher the same makes a tiny difference to the arithmetic!
I deal with that issue by always using the full volume - 1 litre - in my Paterson inversion tank. With replenishment, I don't have to worry about wasting developer!

Matt,
Thank you. You're right, standardizing on a fixed quantity is a great idea. Old habits die hard, but with, XTOL-R, I don't need to worry about economy, as it is fixed to 70 ml per film. Another idea that needs to be added to the specific workflow for replenished developers.

For normal photography, I would never worry about CI drift of 0.1. In fact, regular D76 can, in my experience, drift beyond 0.1 over the span of its shelf life. But I needed to bring CI variability down to 0.01-0.02 because I am going to compare replenished XTOL to stock XTOL, basically repeating the test, to see if there's any difference there. Again, this is probably well-known stuff to experienced XTOL users, but it's new to me, so I want to find out for myself.

 

[Moose22]

For my workflow, -R seems to be the easiest xtol yet.

I have a wine bag-in-a-box next to my sink with fresh xtol in it, and a graduate that I've drawn sharpee lines on at 70, 140, and 210 ml. While the film is in the developer I fill the graduate to the correct level and pour it in the replenished/working bottle, then top it up from the developing tank when developing is done.

I also just fill the whole tank -- probably the same tank, actually. Patterson 4 does 2 of 120 or 3 of 135, takes a liter to do 2 rolls of 120.

SOOOO easy to be consistent.

I am using a 1.25 liter bottle for storage and I think I might move to a 2 liter because it has a flatter bottom, though. I also think the extra volume of working replenished xtol might make the whole system more stable, but my workflow was lifted 90% from Matt's. It's very low effort, very inexpensive.

 

[aparat]

@Moose22 and @MattKing I have looked around for those collapsible bags and holders. A lot of them come with a kind of spout. It makes it easy to pour the liquid out, but are these spouts air-tight? Also, I've found a couple of beautiful wood dispensers/holders, but they are pricey. Still, this wine bag idea is really cool.

 

[MattKing]

@Moose22 and @MattKing I have looked around for those collapsible bags and holders. A lot of them come with a kind of spout. It makes it easy to pour the liquid out, but are these spouts air-tight? Also, I've found a couple of beautiful wood dispensers/holders, but they are pricey. Still, this wine bag idea is really cool.

If a spout let air in, it would spoil the wine .
I have friends who have bought boxed wine, drunk the wine, and happily re-used (for developer) the cleaned out bag and spout, along with the fairly tatty looking cardboard box it came in.
Edit: the quality of the bags can vary a bit - I'm not sure I recommend re-use of the ones that held wine previously.

 

[Sirius Glass]

So the development times for XTOL-R at 68 degrees F [20 degrees C] for Ilford Delta 3200 shot at ISO 3200 and Kodak P3200 shot at ISO 3200 in a Jobo processor is ... <<Insert Drum Roll here please>>??????

 

[aparat]

So the development times for XTOL-R at 68 degrees F [20 degrees C] for Ilford Delta 3200 shot at ISO 3200 and Kodak P3200 shot at ISO 3200 in a Jobo processor is ... <<Insert Drum Roll here please>>??????

In my experience, Kodak's recommended 17 minutes, and Ilford's recommended 18 minutes (for XTOL 1+1) should be very good starting points.

 

[Stephen Benskin]

Thought I'd do an illustration of "pushing for speed" with TMZ.

A standard luminance range of 2.20 with a one stop flare factor exposed normally for the film's EI and printed to match the NDR and LER.

Rating the film at a higher EI than its ISO / tested EI is the exact same thing as underexposure. The second example uses the same film and paper but with a -1 1/2 stop reduction in exposure. The film curve doesn't change. The change is where the exposure falls on the curve. It shifts down into the toe compressing the shadow and darker values consequently reducing the effective NDR. Of course a paper with an LER can be utilized to match the lower NDR, but that is a different example.

In order for the NDR to reach the NDR of the normally exposed negative, the film must be processed to a higher average gradient. Shadow density will increase somewhat with a slight increase in film speed. The darker values remain mostly compressed while the mid-tone and higher values are expanded.

This example uses an average luminance range but not all scenes have the same full range. Scene's with little to no shadows will not experience the same compression. Creatively, lack of detail in the darker values may be aesthetically unimportant in some cases, or a necessary compromise, or it can be a desired effect or mood. Because of the subjective nature of photography, it's impossible to judge the effectiveness of the technique. What isn't happening is changing the film speed to any real degree.

 

[Sirius Glass]

In my experience, Kodak's recommended 17 minutes, and Ilford's recommended 18 minutes (for XTOL 1+1) should be very good starting points.

Thank you

 

[aparat]

What do you guys think of this kind of dispenser for wine bags? It's $100 in the US. Worth it for XTOL? I am afraid the cheap cardboard boxes will get wet and rot.
Ebay link.