Understanding Split Grade Printing

[Pieter12]

Could you explain how it saves paper and chemicals, please?

For me, it takes 2 test-strip prints to find the final exposure combination of the 00 and 5 filters. I can usually judge from those test prints how much to dodge or burn with each filter too, giving me close to a final print with the 3rd sheet of paper.

 

[albada]

For me, it takes 2 test-strip prints to find the final exposure combination of the 00 and 5 filters. I can usually judge from those test prints how much to dodge or burn with each filter too, giving me close to a final print with the 3rd sheet of paper.

Doesn't adding exposure through the #5 filter darken the highlights determined by the first test-strip?
The first posting in this thread shows that adding #5 light darkens highlights, and vice versa.

 

[Pieter12]

Doesn't adding exposure through the #5 filter darken the highlights determined by the first test-strip?
The first posting in this thread shows that adding #5 light darkens highlights, and vice versa.

No. Unless the 5 exposure is extreme (making the whole print dark) it will not affect the highlights at all. And the 5 test strop is made on the same sheet, after exposing the paper with the 00 filter time determined by the first test so any affect it may have will be seen at that point.

 

[Sirius Glass]

"Best print" at lowest contrast (max magenta) will have too-dark highlights and too-light shadows. Or did I misunderstand?

What I said is correct. It works and I learned it at a workshop. Try it in the darkroom. Magenta does not change the high lights, it works on the high contrast areas. The highlights and sky are effected by the yellow.

 

[Frank53]

Doesn't adding exposure through the #5 filter darken the highlights determined by the first test-strip?
The first posting in this thread shows that adding #5 light darkens highlights, and vice versa.

No, that’s one of the advantages of split grade. You can i.e. burn in the sky and not worry too much about masking the foreground.

 

[Pieter12]

No, that’s one of the advantages of split grade. You can i.e. burn in the sky and not worry too much about masking the foreground.

...depends on the tones in the foreground.

 

[snusmumriken]

So are you guys making a 2-way test of the entire image? Printing with straight grades, I make a test strip (about 1/10 of a sheet) across a highlight area - usually at G2 - to determine exposure. I then do another strip at that exposure, from which I judge the appropriate contrast grade. The next thing is a full-sheet print, which is the one I keep in about 50% of cases, unless a lot of burning/dodging is required. I’m reasonably experienced but very far from being an expert printer. It doesn’t sound to me as though doing split grade printing by default is appreciably more economical - or is it?

 

[Pieter12]

So are you guys making a 2-way test of the entire image? Printing with straight grades, I make a test strip (about 1/10 of a sheet) across a highlight area - usually at G2 - to determine exposure. I then do another strip at that exposure, from which I judge the appropriate contrast grade. The next thing is a full-sheet print, which is the one I keep in about 50% of cases, unless a lot of burning/dodging is required. I’m reasonably experienced but very far from being an expert printer. It doesn’t sound to me as though doing split grade printing by default is appreciably more economical - or is it?

You can just do a narrow test strip if you want to. I prefer to use an 8x10 sheet so I can see what is happening over the entire image. I also use an 8x10 when I print 11x14, as it covers most of the printable area. You can also just test a critical area, moving the paper through a window cut in a black board instead of moving the the masking board over the paper.

I have seen some people (well, one person) make an 8x10 test sheet with both filters, masking horizontally for one of the filters, then vertically for the other. It gives you a checkerboard of exposure combinations. I find it a bit confusing to interpret, mainly because I don't have much practice doing it that way--I only tried once.

 

[Frank53]

...depends on the tones in the foreground.

Of course.

 

[Pieter12]

No, that’s one of the advantages of split grade. You can i.e. burn in the sky and not worry too much about masking the foreground.

Now that I think about it, if you are burning the entire image, you really should just add to the exposure for that filter.

 

[Donald Qualls]

an 8x10 test sheet with both filters, masking horizontally for one of the filters, then vertically for the other.

This is how I've done it, mostly. The main drawback is the same as making a traveling mask test strip: you're testing different exposures/contrast in different parts of the image. May or may not cause problems.

 

[Sirius Glass]

So are you guys making a 2-way test of the entire image? Printing with straight grades, I make a test strip (about 1/10 of a sheet) across a highlight area - usually at G2 - to determine exposure. I then do another strip at that exposure, from which I judge the appropriate contrast grade. The next thing is a full-sheet print, which is the one I keep in about 50% of cases, unless a lot of burning/dodging is required. I’m reasonably experienced but very far from being an expert printer. It doesn’t sound to me as though doing split grade printing by default is appreciably more economical - or is it?

You can just do a narrow test strip if you want to. I prefer to use an 8x10 sheet so I can see what is happening over the entire image. I also use an 8x10 when I print 11x14, as it covers most of the printable area. You can also just test a critical area, moving the paper through a window cut in a black board instead of moving the the masking board over the paper.

I have seen some people (well, one person) make an 8x10 test sheet with both filters, masking horizontally for one of the filters, then vertically for the other. It gives you a checkerboard of exposure combinations. I find it a bit confusing to interpret, mainly because I don't have much practice doing it that way--I only tried once.

Sheet 1: I do one full sheet 8"x10" with magenta [5] with time varying columns to get that exposure.
Sheet 2: 8"x10" with the selected expose. Sometime I may need to do a second one.
Sheet 3: 8"x10" I expose magenta [5] with the time for the selected magenta [5] exposure and then do time varying columns with yellow [00].
Sheet 4: 8"x10" I expose magenta [5] with the time for the selected magenta [5] exposure and then expose yellow [00] with the time for the selected yellow [00] exposure. Sometime I may need to do a second one.

So I end up doing 4 to 6 sheet of paper to get the print. If necessary I will work on burn and dodge magenta [5] and then yellow [00].

 

[cliveh]

I would suggest that getting the correct print exposure time is more important, before employing any contrast control.

 

[Pieter12]

Different contrast grades (4 & 5) require different times. The whole point of split-grade printing is to not be bothered by determining contrast grade, but to concentrate on a great final print. And the simplicity of selective burning and dodging with just the 00 and 5 filters.

 

[cliveh]

Different contrast grades (4 & 5) require different times. The whole point of split-grade printing is to not be bothered by determining contrast grade, but to concentrate on a great final print. And the simplicity of selective burning and dodging with just the 00 and 5 filters.

But what if you are using a colour head enlarger, or a condenser, or diffuser with multigrade paper?

 

[Pieter12]

But what if you are using a colour head enlarger, or a condenser, or diffuser with multigrade paper?

I don't think that changes anything. Ilford MGFB paper needs 1.5-2X more exposure for #4 and 5 filters.

 

[ic-racer]

This graph is clever. I spent 30 minutes studying it, and mentally picturing how it would be used to change grade without changing density. Nice.
A few questions:

What paper is this for?​

What density did you use for the measurements?​

After obtaining one equi-density set of points (i.e., one colored line), did you scale them to produce the other colored lines?​

Again, this graph is helpful. You might want to post instructions on how to use it, because I suspect that many people will be overwhelmed by the many lines and immediately give up.

I lined up all the step wedges and picked a value somewhere in the middle. For example if a step wedge contact print had 8 gray bands, I lined up #4 with the "index" test strip and any shift was the exposure compensation for that contrast test.

The final chart answers two questions. 1) How to make the print lighter or darker without changing the contrast and 2) How to change the contrast without making the print lighter or darker. Mostly used in the final stages of tweaking the image. Sometimes even looking at the print the next day in different light etc. and needing a tiny change. The chart is not used to get the base exposure and contrast.

 

[Steve Goldstein]

One thing I haven't seen mentioned in this thread, though I could certainly have missed it, has to do with green/blue, as opposed to yellow/magenta, split printing. Not all blue filters are created equal, and it seems that most of them (Rosco, at least) transmit some green. And they (again, Rosco) tend not to have very high transmission in the blue, requiring long exposure times. On the plus side, a 20"x24" Rosco sheet costs just a few bucks and makes quite a few above-the-negative filters, so I don't have to baby them quite as I would with Ilford 00 and 5 filters.

There's one magical filter in the Rosco catalog, though, the P1394 glass filter that's a special-order item through a Rosco dealer. It has surprisingly high blue transmission, something like 90%, and an extremely sharp band edge that transmits no green. The 6"x6" P1394 I got a couple of years ago was around $100. This was the filter that finally enabled me to get a satisfying print from an extraordinarily flat (even after intensification) negative.

I don't always split print, but when do it's usually with the R389 Green and R68 Blue. I've tried an R384 Blue in place of the R68 and haven't found much difference other than significantly longer blue exposure times with the R384. But every once in a while the P1394 goes into the enlarger, and it makes all the difference, especially if there's a lot of dodging or burning, where its increased brightness really helps my tired old eyes.

 

[ic-racer]

In line with what Steve G brought up, I wonder if any of the hard core split grade users have sourced two nice dichroic filters?

Any way, I found it easy to test the filters to find favorites, all one needs is an uncalibrated step wedge. I never tried any of the nice glass ones, but I bought a bunch of the inexpensive ones and found that picking the "best" Roscoe based on their published transmission data sometimes resulted in a filter that was too dark to be useful.

 

[albada]

I lined up all the step wedges and picked a value somewhere in the middle. For example if a step wedge contact print had 8 gray bands, I lined up #4 with the "index" test strip and any shift was the exposure compensation for that contrast test.

The final chart answers two questions. 1) How to make the print lighter or darker without changing the contrast and 2) How to change the contrast without making the print lighter or darker. Mostly used in the final stages of tweaking the image. Sometimes even looking at the print the next day in different light etc. and needing a tiny change. The chart is not used to get the base exposure and contrast.

I'm still wondering what paper your graph is for.

If I print with that paper, I could use your graph.

 

[ic-racer]

I'm still wondering what paper your graph is for.

If I print with that paper, I could use your graph.

Sorry, that is for Forte that is not made any more and I ran out of ten years ago. Try the graph and see with your paper. Free for anyone to copy and tweak.
Though, I think the more important thing would be which filters are being used. I don't recall which Rosco numbers I used, off hand...ok I found my post about it from 2008 on LF forum, cut and pasted here:

Lines of common density
Here are my graphs I made up so that I can print the way I want to print. These graphs let me change contrast without messing up my middle values.

The Y axis is exposure with the Blue #80 Roscoe filter doubled on itself. The numbers represent seconds.

The X axis is the exposure with the Green #389 Rosco filter (single).

Each colored line represents a series of theoretical exposures where the middle gray stays constant. (The Yellow line from 21 blue to 42 green is actual data (10 datapoints), the other curves are all calculated assuming no failure of receprocity)

The radial lines are lines of constant contrast. Of course any intermediate value is possible, the lines pictured are just the ones I happened to test. The numbers by the lines represent the density range in log units. It roughly goes from grade 5 to grade 00.

The graph is simple to use. From any point in the graph, to make a contrast change without changing the middle grays, just move paralell to one of the colored lines to your new contrast grade. The graph will then show you the new times for green and blue.

 

[ic-racer]

Something I wrote in 2008 from tests on filters at the time maybe still be of interest:

Results
MAX MAGENTA: 0.60 log = 60 ISO = Grade 4
Rosco BLUE: 0.75 log = 75 ISO = Grade 4
MAX CYAN: 0.90 log = 90 ISO = Grade 3
NO FILTER: 1.05 log = 105 ISO = Grade 2.5?
MAX Y & C (GREEN): 2.1 log = 210 ISO = Grade 000
MAX YELLOW: 2.1 log = 210 ISO = Grade 000
Rosco GREEN: 2.1 log = 210 ISO = Grade 000


Interpretations:
...
2) The contrast from Magenta or Blue is similar, or perhaps identical within experimental error.
3) The contrast from Yellow or Green is identical.

 

[ic-racer]

More of my test results from 2008:

#91 = Rosco Primary Green
#80 = Rosco Primary Blue
#389 = Rosco Chroma Green


Log values:
#91 = 2.10
#389 = 1.95
#80 = 0.60
Bare = 0.90


Relative speed to 'middle gray' (the step in the middle, between the black and white; not some standard gray value)

Bare = +1 stop
#80 = 0 by my definition
#389 = -1.5 stop
#91 = -3.5 stop

So, I knew #91 is really slow, but now I have a number for it.

 

[Pieter12]

One thing I haven't seen mentioned in this thread, though I could certainly have missed it, has to do with green/blue, as opposed to yellow/magenta, split printing. Not all blue filters are created equal, and it seems that most of them (Rosco, at least) transmit some green. And they (again, Rosco) tend not to have very high transmission in the blue, requiring long exposure times. On the plus side, a 20"x24" Rosco sheet costs just a few bucks and makes quite a few above-the-negative filters, so I don't have to baby them quite as I would with Ilford 00 and 5 filters.

There's one magical filter in the Rosco catalog, though, the P1394 glass filter that's a special-order item through a Rosco dealer. It has surprisingly high blue transmission, something like 90%, and an extremely sharp band edge that transmits no green. The 6"x6" P1394 I got a couple of years ago was around $100. This was the filter that finally enabled me to get a satisfying print from an extraordinarily flat (even after intensification) negative.

I don't always split print, but when do it's usually with the R389 Green and R68 Blue. I've tried an R384 Blue in place of the R68 and haven't found much difference other than significantly longer blue exposure times with the R384. But every once in a while the P1394 goes into the enlarger, and it makes all the difference, especially if there's a lot of dodging or burning, where its increased brightness really helps my tired old eyes.

The disadvantage of using green/blue is twofold: sourcing the proper filters (I have learned not to trust Roscoe, they are great for gelling lights in the studio, but not necessarily that well-made, long-lasting or accurate) and the difficulty in seeing the projected image through those filters--something pretty essential for dodging and burning. I purchased 12"x12" Ilford 00 and 5 gel filters and cut them down to 6"x6" so I am set for at least a few years.

 

[albada]

ic-racer:
Thanks for all the info. Now I know exactly how your graphs were created. I'm surprised that neither dichroic nor Rosco filters were able to reach grade 5, indicating that both let a significant amount of green sneak through. Or perhaps that paper couldn't reach grade 5. However, the exposure range in the Forte graph you posted got down to 0.45, which I think is grade 5. Did that use a different filter?

Here's an idea: Make a graph that keeps highlights unchanged (instead of midtones). Then you could:

1. Guess the grade, and find correct exposure for highlights the conventional way.
2. Estimate final exposure-range (implies grade) based on the test-print and last test-strip.
3. Using the graph, move to that exposure-range, which will preserve highlights and make shadows correct.

Many people use this approach of highlights-then-shadows, but without the assistance of your graph. Have you tried such a highlight graph?