filter equivalents in printer lights

[casey tompkins]

The reason for this question arose after getting two bad prints of a negative back from the lab. I will be sending the negative back for a third print and this time, instead of requesting they time to grey card, I am going to call my lights. Essentially, the lab, rather than basing the printing lights off my grey card, has printed out the color in my negative that was created by a tobacco filter I was using. To clear up any misunderstanding, I did not shoot my grey card with the tobacco filter on. Here is my question. When calling my own lights on a print, is there a printer light combination that is equivalent to a tobacco filter but that does not change the brightness of the print? For instance, if the first print's exposure was dead on but they timed out my tobacco filter, is there a known combination of printer light adjustments that I could adjust my first print's lights to without effecting the brightness of the print. Say, -4 in red, +1 in green, and +3 in blue.... or something along those lines.

 

It's interesting how in modern digital color correcting it is just a matter of clicking a button to apply common filter effects to an image, but, to my understanding, there is no literature available that discusses the relationship between printer lights and lens filters.

 

Thanks in advance for any advice/discourse.

 

Casey Tompkins

[K Borowski]

Well, a printer light isn't a digital unit either, comparing one "trim" to a point in Photoshop or another graphics program they are all different.

 

It's possible, you're right to replicate a good deal of subtle filters digitally. I'm not sure what the equivalents are. With optical printing you don't have as fine of control as with a subtle filter (of course if it's too subtle you can end up with what you've ended up with, the lab timing it out as a film stock color shift or color spill from not shooting under the proper color temp.)

 

The advantage of filters are, that they are DONE. You don't need to stare at a computer monitor and tweak.

 

I'd keep hounding them to follow directions and not time the filters out. The problem is your timer, not your approach. If your timer doesn't know how to follow directions with leaving a filter effect in, maybe he or she doesn't know how to count either ;-)

[casey tompkins]

Well, a printer light isn't a digital unit either, comparing one "trim" to a point in Photoshop or another graphics program they are all different.

 

It's possible, you're right to replicate a good deal of subtle filters digitally. I'm not sure what the equivalents are. With optical printing you don't have as fine of control as with a subtle filter (of course if it's too subtle you can end up with what you've ended up with, the lab timing it out as a film stock color shift or color spill from not shooting under the proper color temp.)

 

The advantage of filters are, that they are DONE. You don't need to stare at a computer monitor and tweak.

 

I'd keep hounding them to follow directions and not time the filters out. The problem is your timer, not your approach. If your timer doesn't know how to follow directions with leaving a filter effect in, maybe he or she doesn't know how to count either ;-)

 

 

Thanks for the reply Karl.

I wasn't trying to compare printer lights to digital units or points in photoshop... I wouldn't even know where to start if that were the case. The digital comment at the end of my post was merely a comment comparing the process of adjusting prints digitally compared to photochemical.

 

Anyways, I completely agree with the advantage of shooting with filters is that they are "done", that an also I am not planning on scanning the negative so I need my print to be as close to how I shot the negative as possible.

 

I'd send it back and ask for a third print timed to grey card but I have simply lost faith in getting the results I am after. Also, I want to use this event as a learning experience so I have a better understanding of calling my own lights even in situations like this.

[K Borowski]

Threaten to take your neg. to another lab. That should elicit the desired response! Also, if you have paid for a print that, multiple times, was not produced to your explicit written instructions, they should make you a new one for free.

 

Make sure, of course, that your filter factors were right before calling them out though!

 

 

As far as point equivalents, I forget. I want to say 0.30 (the equivalent of a T-stop exposure difference) on a densitometer is 20 points in photoshop? Dominic Case wrote the correlation between a printer trim and a point on the densitometer, but I forget.

 

I know a lot of additive color printers are/were based on 50R 50G 50B being maximum exposure intensity through the negative. So the scale is a lot smaller. Even a 1 trim adjustment was .02 or .03 at least if I recall Dominic's explanation correctly.

 

 

Putting that in terms of your tobacco filter, that'd be, what, a subtle red addition and blue subtraction (er, wait opposite that), maybe less than a single trim on each!

 

I guess it depends on the lab's printers too. I'm sure there are printers with very subtle control, fractions of a trim. My knowledge of additive printing comes from the days of primitive '80s video analyzers and "computers" of that nature. I even have a densitometer that has individually lighting up numbers, the digital pattern found on a calculator being too high-tech to include.

 

So I'm sure improvements have been made, whether or not your lab has them.

 

Even with the greatest optical printer, digital printing is generally more accurate and far more controlled, almost like you could get if you could easily adjust processing AND printer lights without any film hassles like crossover or density shifts or reciprocity.

[Brian Pritchard]

A printer point is 0.025 Log E and equivalent to a 2.5 CC filter. A 30 CC filter is equivalent to 12 points.

 

If you are using filters to work out what correction you will need when printing you have to remember that because of the uplift in contrast between the negative and print, a smaller correction will be required on the negative to give a larger change on the print.

 

Each change of a printer point will make a change of 0.07 density on the print (at the LAD density). So if your print requires a 30cc yellow correction, 0.3 density correction, then the correction to your negative grading would be 0.3/0.07 = 4.28, giving around 4 points. You would increase the blue light by 4 points and reduce the red and green by 2 points (if you want to keep the density the same).

 

The other point to bear in mind is that when using filters to colour correct you have to judge by the middle tones because the filter will overcorrect the highlights and undercorrect the shadows.

 

Brian

Edited April 22, 2010 by Brian Pritchard

[Dominic Case]

A printer point is 0.025 Log E and equivalent to a 2.5 CC filter. A 30 CC filter is equivalent to 12 points.

Not so fast, Brian :unsure:

 

You are correct when it comes to converting printer points to a filter correction in the printer - and in the later part of your post when you discuss laying a colour filter over a print to extimatge a regrade correction.

 

But I think the discussion here is about the effect of a filter over the camera lens, in terms of printer lights when it comes to make a print. So the gamma of the negative comes into play.

 

In plain terms, a 0.30ND filter (a 30ND) on the camera is equivalent to one stop (we know that). It will have the effect of reducing negative density by about 0.17 (which is the 0.30 times the gamma of the negative, around 0.55). To correct that in the printer lights, you need a change of 7 printer lights (at the value you mention of 0.025logE).

 

It's the same for individual colours, so any filter that has a density of 0.30 to one primary colour (R, G or B) will correspond to a printer light change of 7 lights in that colour. A yellow filter would reduce the blue printer light etc.

 

I'm not sure what the density readings of a tobacco filter would be, but I suspect your guess of -4R +3B wouldn't be far off. Probably more on the red if anything. To grade out an 85 filter takes about a 20-point shift.

 

Bear in mind that, especially for creative filters like tobacco (asdistinct from CC filters) the exact spectral profile is significant and you can't make an exact comparison with RGB grading, or even with more complex correction in Photoshop or the Lustres and Apple Colour systems.

 

Finally, despite all this knowledge, I have always recommended that you request the grader to give you what you describe,, rather than giving her/him instructions on how to grade - which they are supposed to know better than you! As others have suggested, you need to be very explicit. If you shot a grey card, unfiltered, then have them grade that to neutral and print the rest of the roll AT THAT LIGHT. If the print comes back with a neutral greycard and yur filtered shots also neutrl, then maybe you are overestimating the effects of your filter. If the greycard isn't neutral, then they haven't done as you ask.

 

But if you call the lights and you still don't get what you hope for, then you are finished. You have nowhere left to go.

 

Sorry this is long - it's a complex, but not at all unfamiliar problem.

[Brian Pritchard]

But I think the discussion here is about the effect of a filter over the camera lens, in terms of printer lights when it comes to make a print. So the gamma of the negative comes into play.

Dominic

I am not sure that is what Casey meant. I read it as the lab had graded out his tobacco filter and he wanted to put it back by regrading the print. However I would agree that telling a grader what ights to use is not a great idea because if you make a mistake and the print comes out wrong you will have to pay for another print whereas if you requested a certain colour correction and the grader gets it wrong you should get a reprint - at least that is the way the labs I have a worked in resolved the problem

Brian.

[Alexandra Wesche]

Hello all,

 

during my research concerning relationships between density, filters, printer points and film stocks I keep returning to this topic to understand them.

 

I understood the relationships between lense filters and negativ gamma, but I often also read that a 30ND filter at 0,025 Log E equals one F stop and 12 printer points.

 

When is that the case?

 

 

I also have a question concerning CC filter grading on a print and the translation into printer points.

 

Brian writes above that, because of the contrast increase from negative to print CC filters can't be seen as 2,55 CC = 1 printer point.

 

Do I understand the following correctly?

 

A 30 CC/ND filter on a camera lense leads to a density change of 0,17 on a negative with the contrast of 0,55 and can be corrected with 7 printer points when grading for the print. That means 0,025 Log E on the printer = 0,025 density on the film.

 

When (sight) grading with a 30 CC filter (I know that's too much) on a print, we have to include the contrast increase from negative to print. Which would mean 0,025 Log E on the printer = 0,07 density change in the film.

 

The process is the same: a correction of the negative for the print. In the second case, do we have to calculate with 0,07 instead of 0,025, because we're judging the correction by the look of the print?

Do I understand it correctly that 0,07 can only be used, if the density change in a LAD patch is actually exactly 0,07 for 1 printer point on a specific printer.

 

What can I do if I feel that a print is only 1 point too magenta (or whatever) judging by the CC filters.

 

 

Please excuse my complicated expression of this topic. I still have to think very hard to express these thoughts... I hope it's still somewhat understandable.

[Brian Pritchard]

Alex

The use of filters for re-grading is really only an approximation. As I mentioned in a previous post the filter will over correct on the highlights and under correct the shadows so it requires a lot of experience to decide on the correct filtration. The filters are most useful to decide whether the correction is blue, cyan or a combination of both, again for example.

 

You also have the difficulty that printer points are discrete amounts; you can't correct by half a point, for example.

 

If you are re-grading using filters you would look at the print over a light box, use the filters to get the approximate correction and then use your experience to judge whether a four point correction suggested by filters is really perhaps 3 or 5 points. This is the reason that the introduction of Video Colour Analysers in the late 60's was such a revolution for colour grading. Up to that point the first and possibly the second print ended up in the bin (or the bulk release if you were a little bit sharp) after the introduction the best graders were able to show the client the first print as the answer print.

 

Before analysers some labs used special printers that were able to print a few frames of each scene at different lights to enble the grader 5o get a reasonable first print without the expense of printing a complete test print first. They were called scene testers; they were also used in the BW days.

Brian

[Simon Wyss]

Here is my question. When calling my own lights on a print, is there a printer light combination that is equivalent to a tobacco filter but that does not change the brightness of the print?

No, there isn’t. No lab technician in the world, not even Rembrandt if still alive, could add to a photographic colour picture what’s not in it. RGB printing light systems influence the copy as a whole, I want to say: you alter the base for the exposure, not the colours alone. These interrelations go overlooked a bit with the electronic simulation because there contrast, density, and hue seem to be variable separately. In fact, one can change a single colour channel’s value alone. Chemical film responds broadbandly, we have to set more distinct points than the computer-educated client expects.

 

You see, bigger steps on the printer light scale alter the colour balance so much that a subtle filter is just outdone. Personally, I’d never dare attempting to time for anything else than a combined colour-grey board. The rest must be before the camera.

 

Isn’t this what cinematographers demand from a lab, repeatable accuracy?