Is pre-exposing negs a useful tool or a complete waste of time?

[Stephen Benskin]

Agree. In the end the result is the same. My only preference all along would have been to present the effects of non-image forming light in quadrant 2, or on its own, rather than in quadrant 1. Just a personal preference I guess (the way it is presented on p. 160 in Henry's book, for example).

Having a separate quadrant each factor would be ideal so that each influence can be identified and measured. Unfortunately I'm restricted to the four quad.

 

[Xmas]

Well I understood post #44 or thought I did.

A preflash is not supposed to show (in practice it might lift the base fog level a little dependent upon film), all it is supposed to do is alter the toe of the H&D softening the contrast and increasing the ISO but not to detectably alter the shoulder (the degree of change in contrast and ISO would depend upon the film).

I used to use it for duping slides, way long ago and it made more acceptable dupes. compared with omitting a preflash.

 

[markbarendt]

all it is supposed to do is alter the toe of the H&D softening the contrast

It does soften the contrast, but...

and increasing the ISO but not to detectably alter the shoulder (the degree of change in contrast and ISO would depend upon the film).

The ISO rating of a film doesn't change. Your "EI" might.

Any film and it's response to light remains constant as long as development remains constant. The film curve, doesn't change just because we use two exposures instead of one.

What changes is where subjects land on the curve, placement.

Mathematically, the only way to get the curve to change shape, is to change the scaling of the graph.

 

[Rudeofus]

Any film and it's response to light remains constant as long as development remains constant. The film curve, doesn't change just because we use two exposures instead of one.

I have no idea where you have this from, and why this is still actively posted here in this thread. The characteristic curve is a special case, which just happens to cover common photographic practice. Expose at low temperatures, or with very weak or strong light levels, and your characteristic curve goes down the drain. The whole idea behind preflashing is that the exposure-density curve is not even approximately linear in the lower toe region, and that you can not interpret the combination of these two exposures like you would in a linear case. Athiril could not post such images if preflashing just shifted the curve.

BTW even if the characteristic curve was a universal law of physics, fully continuous and whatnot, a preflash would not translate into a linear shift, neither sideways nor upwards. Remember, it's a logarithmic scale in both axes.

 

[DREW WILEY]

The whole point of flashing is that it is such a small amount of exposure itself that it has only an insignificant effect higher up the curve than
where it is targeted. But down there it does by definition add density and lift the toe of the film, hence diminishes the pace of shadow and
textural separation. Whether this is good or bad depends on what you are trying to achieve, the specific film, and the lighting ratio. But the risk is ending up with mud down there if you overdo it.

 

[markbarendt]

The whole point of flashing is that it is such a small amount of exposure itself that it has only an insignificant effect higher up the curve than
where it is targeted. But down there it does by definition add density and lift the toe of the film, hence diminishes the pace of shadow and
textural separation. Whether this is good or bad depends on what you are trying to achieve, the specific film, and the lighting ratio. But the risk is ending up with mud down there if you overdo it.

What I'm getting at and what Stephen's charts show is that subject placement changes, the shape of the curve does not.

Surely there is compression, yes the relative distance between tones changes vertically, but that's not because the shape of the curve changes; it's because the extra exposure has pushed the tones furthar right on the normal curve.

 

[Xmas]

Could you then explain why my preflashed duping slide film was a lot less contrasty than non preflashed duping film.

Ditto for the graded paper I preflashed for softer highlights?

 

[RobC]

What I'm getting at and what Stephen's charts show is that subject placement changes, the shape of the curve does not.

Surely there is compression, yes the relative distance between tones changes vertically, but that's not because the shape of the curve changes; it's because the extra exposure has pushed the tones furthar right on the normal curve.

There is a disconnect between making a test chart/curve and making an actual print.

If you photographed a reflection step wedge without a pre-flash and then photographed it with a pre-flash and then charted two two results, you would get a different shape curve for each in chart. what has remained fixed is the development. But we already know that. So what you are saying is kind of correct in one way but wrong in another because the whole point of pre-flashing is to alter the distribution of the tones on the toe of the curve which does change the resulting curve.
The confusion is that you are talking are talking about a film speed/contrast curve whereas everyone else (mostly) is talking about a tone dstribution curve. The two are subtely different in concept. So I guess both sides of the argument are correct.

 

[markbarendt]

If you photographed a reflection step wedge without a pre-flash and then photographed it with a pre-flash and then charted two two results, you would get a different shape curve for each in chart.

No just different placement. This is exactly what Shephen's chart shows in quadrant 2.

To get a different curve you have to change the assumptions of the math involved. Change the Log scale or factor out the fog or some such thing.

The basic film curve doesn't change.

the whole point of pre-flashing is to alter the distribution of the tones on the toe of the curve which does change the resulting curve.

I would rephrase that:

The whole point of pre-flashing is to improve the shadow detail in the print without affecting the highlights.

In Stephen's chart it's obvious that the whole toe of the film curve gets buried with flare and pre-flash.

Pre flash pushes everything usable rightward on the curve, all the way onto the straight line.

Quadrant 1 shows the curve I think your are suggesting.

You can superimpose one over the other to illustrate your point but the scales used to define each of those curves, are different.

 

[markbarendt]

BTW even if the characteristic curve was a universal law of physics, fully continuous and whatnot, a preflash would not translate into a linear shift, neither sideways nor upwards. Remember, it's a logarithmic scale in both axes.

Exactly.

 

[Xmas]

There is a disconnect between making a test chart/curve and making an actual print.

If you photographed a reflection step wedge without a pre-flash and then photographed it with a pre-flash and then charted two two results, you would get a different shape curve for each in chart. what has remained fixed is the development. But we already know that. So what you are saying is kind of correct in one way but wrong in another because the whole point of pre-flashing is to alter the distribution of the tones on the toe of the curve which does change the resulting curve.
The confusion is that you are talking are talking about a film speed/contrast curve whereas everyone else (mostly) is talking about a tone dstribution curve. The two are subtely different in concept. So I guess both sides of the argument are correct.

HiRobC

When I printed on graded paper with a preflash I was able to get softer highlights and areas of non fogged paper where I had burnt the film on the same sheet.

It was a high key print.

It took some effort with preflash exposure and print exposure test strips.

I was not moving up the H&D to get reduced contrast that would have left me with Qtips damp paper and farmers reducer.

This was some dacades ago so my memory could be at fault.

I assumed the papers curve had changed from the preflash.

Noel

 

[RobC]

No just different placement. This is exactly what Shephen's chart shows in quadrant 2.

To get a different curve you have to change the assumptions of the math involved. Change the Log scale or factor out the fog or some such thing.

The basic film curve doesn't change.



I would rephrase that:



In Stephen's chart it's obvious that the whole toe of the film curve gets buried with flare and pre-flash.

Pre flash pushes everything usable rightward on the curve, all the way onto the straight line.

Quadrant 1 shows the curve I think your are suggesting.

You can superimpose one over the other to illustrate your point but the scales used to define each of those curves, are different.

That's one way of looking at it. The other is that film or paper doesn't have a curve. The curve is a mathematical construct after the event which means there is no pre-existing curve to place anything on. The curve construct after the event just shows what happened and in the two scenarios I gave above, the first would produce a different curve than the second.

Your concept is that there is a pre-existing curve which youn are predicting. But if you change the parameters to the curve then it should be obvious that the curve will change. What I don't think you have grasped is that a single exposure, and the basis of your curve prediction, follows the exponential laws of exposure. By adding a pre-flash (double exposure) you break those laws so that the rules for constructing your predicted curve change. Result = new curve.

 

[markbarendt]

That's one way of looking at it. The other is that film or paper doesn't have a curve. The curve is a mathematical construct after the event which means there is no pre-existing curve to place anything on. The curve construct after the event just shows what happened and in the two scenarios I gave above, the first would produce a different curve than the second.

Your concept is that there is a pre-existing curve which youn are predicting. But if you change the parameters to the curve then it should be obvious that the curve will change. What I don't think you have grasped is that a single exposure, and the basis of your curve prediction, follows the exponential laws of exposure. By adding a pre-flash (double exposure) you break those laws so that the rules for constructing your predicted curve change. Result = new curve.

Actually my premise is just that for a given film, in a given development scheme, that: a given amount of exposure, will produce a given amount of density.

I don't care how the film gets the exposure, 1, 2, or 10 shutter cycles.

The only conditions I'll lay down to keep the discussion "fair" is that reciprocity failure must be accounted for to determine the actual usable exposure (it is after all a known) and that only one measurement scale be used on either axis of the H&D chart.

 

[Xmas]

Actually my premise is just that for a given film, in a given development scheme, that: a given amount of exposure, will produce a given amount of density.

...

But that is not a valid hypothesis?

 

[markbarendt]

But that is not a valid hypothesis?

Really?

Please, explain why not.

I will clarify that I mean actual film exposure.

 

[Xmas]

Hi Mark

That is how I read posts #56 and #59 - the key word is linear - YMMV
Don't forget the OP wants to know if preflash is a waste of time Athirils results ie ISO says no it is not a waste of time.

Noel

 

[markbarendt]

Hey Noel,

Any pre-flash that is done in a productive manner, say within Rudy's and mother natures rules for success in creating developable silver, should do what I suggest.

I agree that pre-flash works.

 

[Xmas]

Well a different analogy is a pin ball machine.

If you need four balls in the win zone and they don't stay in the zone long the game goes on 'for ever', but if you cheat and super glue in three the first time another drops in you win cause you cheated.

A preflash is cheating but the process is more complicated.

And to be fair Athirils results look better than any I get.

 

[Gerald C Koch]

BTW even if the characteristic curve was a universal law of physics, fully continuous and whatnot, a preflash would not translate into a linear shift, neither sideways nor upwards. Remember, it's a logarithmic scale in both axes.

Not true. The characteristic curve describes the relation between density (linear) to the log of exposure. So the plot is logarithmic on one axis and linear on the other.

I would also like to comment on a subtle erroneous notion that pops up from time to time on APUG and elsewhere. That is that a relationship is somehow "unreal" because it is non-linear. Not all relationships in nature are linear. For example the decay of a radio-isotope is not linear but logarithmic.

 

[Athiril]

Well a different analogy is a pin ball machine.

If you need four balls in the win zone and they don't stay in the zone long the game goes on 'for ever', but if you cheat and super glue in three the first time another drops in you win cause you cheated.

A preflash is cheating but the process is more complicated.

And to be fair Athirils results look better than any I get.

I find it ineffective for simple underexposure and normal processing times, combined with a strong push is what gets it there. I based it off an article about Tri-X and heavy pushing to get detail you otherwise simply cannot get.

 

[Stephen Benskin]

A film's characteristic curve is a construct. It is the graphic representation of the film's response to exposure and given development conditions. The X-axis is exposure and it increases from left to right. The Y-axis is density. For tone reproduction purposes, the idea is to isolate the different variables into their own quadrants (have their own curve). The following example is a multiple quadrant diagram which includes in addition to the film and paper curve, printer characteristics, viewing conditions, color efficient, illumination level, and objective reproduction.



For the sake of simplification, sometimes the variables are combined. I include flare in my paper curve by projecting a step tablet instead of contacting. To be clear, projecting a step table does not represent the paper's characteristic curve. It represents the paper curve with enlarger flare. In order to produce a true representation of how the paper responds to exposure, the step tablet would need to be contacted.

A film's characteristic curve represents the resulting densities from exposure under given development conditions. A film plane exposure, H, of 0.0032 lxs will fall at a specific point along the X-axis. Please see the below example. The resulting density for that exposure falls directly above that point. For an exposure of one stop greater, 0.0064 lxs, the exposure will fall 0.30 log-H units to the right of 0.0032 lxs. No matter how the exposure 0.0064 lxs is reached, if the film receives 0.0064 lxs, it will fall at a specific point and the resulting density will always be the same. In the below example, the density resulting from an exposure of 0.0064 lxs 0.10. Whether the extra stop is from opening up the f/stop, a slower shutter speed, flare, or pre-exposure, if the resulting film plane exposure is 0.0064 lxs, the resulting density will be 0.10. This becomes a lot clearer when using a calibrated sensitometer and a film curve with actual log-H values.



The below example shows how the camera/flare curve contains the effects of exposure from a scene luminance range of 2.40, one stop flare, one stop pre-exposure. I've included Zone references. Without flare or pre-exposure, the exposure will fall way down at 0.0020 lxs. Flare and pre-exposure bring is up to 0.0080 lxs. The scene luminance range of 2.40 becomes an effective illuminance range of 1.81. The right two columns in the Scene/Camera Image Data table shows the exposure with flare, Hf, and without, Hnf.

 

[Athiril]

I should mention that my preflash was Zone III at EI 12800, meaning, it was a Zone -1 preflash if you consider it at box speed.

I also did other preflashes, but this was the effective one at the push time, the Zone II (12800) preflash is effectively no different in shadow detail rendition and density than no pre-flash, which is only 1 stop out from the Zone III (12800) preflash.

I might redo this experiment because I remember all densities jumping up.

Zone II (12800) preflash
Superia 800 @ 12800 Zone 2 preflash by athiril, on Flickr

vs Zone III (12800) preflash
Superia 800 @ 12800 Zone 3 preflash by athiril, on Flickr



Going back to that example, it doesn't appear as expected behaviour if


Zone 1 = 1
Zone 2 = 2
Zone 3 = 4
Zone 4 = 8
Zone 5 = 16
Zone 6 = 32
Zone 7 = 64
Zone 8 = 128
Zone 9 = 512
Zone 10 = 1024



+Zone II preflash
Zone 1 = 3
Zone 2 = 4
Zone 3 = 6
Zone 4 = 10
Zone 5 = 18
Zone 6 = 34
Zone 7 = 66
Zone 8 = 130
Zone 9 = 514
Zone 10 = 1026



+Zone III preflash
Zone 1 = 5
Zone 2 = 7
Zone 3 = 9
Zone 4 = 13
Zone 5 = 21
Zone 6 = 37
Zone 7 = 69
Zone 8 = 133
Zone 9 = 517
Zone 10 = 1029



Then it's unexpected behaviour, even if you account for that it looks like it might be cut off at Zone IV or Zone III (nopreflash) it's still unexpected from the little difference between no-preflash vs Zone II preflash vs Zone III preflash, where the Zone III preflash has more than 1 stop of improvement vs the other two.

I should also mention I shot it also at 3200, with no-preflash on the same roll, which went through the same developer at the same time.

The Zone III (12800) pre-flash is pretty close to that, maybe slightly better, but they are quite similar, like the Zone II (12800) pre-flash and 12800 no-preflash shot are quite similar.

3200 no pre-flash
Superia 800 @ 12800 Zone 3 preflash by athiril, on Flickr

If the above example is 12800, then below would be 3200
Zone 1 = 4
Zone 2 = 8
Zone 3 = 16
Zone 4 = 32
Zone 5 = 64
Zone 6 = 128
Zone 7 = 256
Zone 8 = 512
Zone 9 = 1024
Zone 10 = 2048




I don't think the examples match all the postulating.

 

[RobC]

possibly not because it depends what your main exposure is. If you exposed for highlights, midtones or shadows. I assume you didn't expose for the shadows so it must have been the mid or higher tones. The fact you have effectively bracketed all your shots to arrive at a sucessful one with your push indicates to me that that there are a lot of unknown(to me) variables in the process to arrive at your result. My example numbers were just to illustrate a point and not intended as an exact match for your result.

Also if we assume your colour film normal development produces a 7 stop range onto paper then how many stops did you close down for preflash? Was it really zone 3 of 10 or 2 stops less than middle of 7 which would be zone 1.5. All depends on your metering and normal film constrast range.

 

[Xmas]

I find it ineffective for simple underexposure and normal processing times, combined with a strong push is what gets it there. I based it off an article about Tri-X and heavy pushing to get detail you otherwise simply cannot get.

Fair dinkum Athiril - but I only tried/try for reduced contrast in shadows or highlights in prints...

 

[Rudeofus]

Not true. The characteristic curve describes the relation between density (linear) to the log of exposure. So the plot is logarithmic on one axis and linear on the other.

Jerry, as you sure know, density is a logarithmic measure of absorption. My statement stands as I wrote it. No densitometer I am aware of can measure density without forming the logarithm of measured light.

I would also like to comment on a subtle erroneous notion that pops up from time to time on APUG and elsewhere. That is that a relationship is somehow "unreal" because it is non-linear. Not all relationships in nature are linear. For example the decay of a radio-isotope is not linear but logarithmic.

Many people associate the term "linear" with straight line curves, but I thought about something else: L(a*X + b*Y) = a*L(X) + b*L(Y). Such a relationship will obviously not hold across the whole characteristic curve, but in most sections it will hold in infinitesimal regions where a linear interpolation is possible. In the lower toe region, though, we have this "three silver atom" threshold, and that relationship above will not hold even locally. This is where the characteristic curve falls apart, this is where Mark's theory "it doesn't matter how many exposures contribute to the total amount of light" falls apart, and this is likely where Athiril got his extra stops of real sensitivity from.