transposing camera f/stops on the Log E axis

[Giles Rais]

Dear friends:

 

I am trying to create camera f/stop scales to characteristic curves, so that I can precisely place areas of exposure of my scene on specific sections of the curve for my stock. I know about measuring the Log E axis (in my case not only the straight-line portion of the curve, but from the beginning of the toe to the end of the shoulder, right before D-max) to obtain Log E units, and to divide those by .3 to obtain camera f/stops.

The problem arises when, for instance, I have a curve that measures let's say 3 Log E units which divided by .3 gives me 10 camera f/stops. I believe that since I measured my curve from where the negative begins to react to light until it stops reacting to it, my 10 camera f/stops should be placed so that middle gray (right at the middle of my f/stop scale, marked as 0.0 by Kodak) should be right in the middle of the length of my characterisitc curve (excluding the areas of D-min and D-max).

However, when I look at examples of characteristic curves that also have camera f/stops, they place middle gray in an area of the curve that is off the center of the lenght of the curve, again excluding the D-min and D-max areas (for instance, the curves for 5248 and 5246.

I need to understand the rationale behind the placing of camera f/stops along the Log E axis beyond simply how to translate Log E units into f/stops. Could someone clarify this for me? I appreciate your input tremendously.

Edited February 19, 2005 by Giles des Rais

[John Pytlak RIP]

There is a tutorial about curve placement on the Kodak website:

 

http://wwwau.kodak.com/US/en/motion/suppor...tructureP.shtml

[Giles Rais]

There is a tutorial about curve placement on the Kodak website:

 

http://wwwau.kodak.com/US/en/motion/suppor...tructureP.shtml

<{POST_SNAPBACK}>

 

Yes, Mr Pytlak, I went over that tutorial many times, but it does not say anything about the rationale behind the way Kodak places camera stops along the Log E axis, which is exactly what I am trying to find out. According to the few characteristic curves that do come with a camera stop scale and a Log E axis, the camera stops scale for middle gray is placed along the straight line portion of the curve at seemingly random places. I know how to tell how many camera stops a given curve can give me, I just need to place them along the curve properly. As far as I am reading, the placing of middle gray ("0.0") should be where the manufacturer decides correct gamma should be, usually 0.8 for negative color (?) and that seems to be the right placement for middle gray along the scale...is this correct? Does anyvbody else have a clue? Here is my problem: for 7218, the camera stops scale shows middle gray at around 0.8 density for the red sensitive layer; for 7277 the 0.0 is placed nowhere near 0.8 density... how can one properly place the camera stop scale? Thanks again for your insights everyone!!

[Giles Rais]

I think I may be on the right path. I was checking Dave Viera's book on lighting and he cites that most color negative is processed at a gamma of .65. This makes sense since labs don't ask you what speed your film is when they develop it (unless you want to push or pull).

I believe, however, that the .65 gamma figure is obsolete by now...I am checking Kodak's website but can't find the right information. Yes, I checked the sensitometric pdf data files for several negative stocks and they don't list Kodak's processing suggestions for gamma. Am I right here? Is the suggested processing gamma the point where Kodak places 0.0 for camera stops on characteristic curves? Thanks again.

[Dominic Case]

It all depends on how you understand "middle grey". It's nominally around 18%, which isn't quite in the middle of a standard grey card with white at 100% and black at 3% (5 stops): where is white (100% reflectance?) how bright a hotspot do you want to reproduce accurately; how deep a shadow?. The actual negative density recommended to do this is not just a number (e.g. 0.80R etc), it is a function of the entire curve shape.

 

The rationale is about making the film work as well as possible. Exposure placement on the curve is the main, but not the only factor.

 

It might appear that the recommended placement is too far down the curve - there is heaps of headroom on the over-exposure side and far less on the under-exposure side - but there is another concept - that of optimal sharpness. Over-exposed negative tends to lose sharpness due to image spread, so the rationale - ultimately - is to expose as little as possible consistent with good shadow detail.

[Giles Rais]

I found some info on the Kodak website that may be exactly what I was looking for. If someone could confirm my conjectures it would be great. Here it goes:

 

I checked Kodak document "Processing Kodak Motion Picture Films, Module 1"

 

In it, it describes the ECN-2 processing system, which is listed as the processing system for most Kodak color negative motion picture stocks pdfs. Under the section for "Control Parameters" they make reference to table 1.3, "calculating aim values for process ECN-2" and in that table they list the following aim values for the three emulsion layers of a color stock: R= 0.85, G=1.33, B=1.59. The 0.5 density value seems to match the placement of the "0.0" of a camera stop scale on a characteristic curve. This makes sense to me, since placing the "0.0" right above the mid point of a stock's range of densities, as measured from threshold to the end of the shoulder should correspond to the middle density, or middle gray.

 

This would mean that after converting Log E units to camera f/stops by dividing them between 0.3 (of course only counting those Log E units between the threshold and the end of the shoulder), I could then simply place middle gray ("0.0") over the point where the red layer reaches a density of 0.85, and then place the rest of the f/stops (+1, +2, +3, etc. and -1, -2, -3, etc.) at either sides of this 0.0 point.

 

This would also solve the problem of knowing how many f/stops of over and underexposure go where, since some stocks have an uneven distribution of f/stops at either side of middle gray.

 

Please correct me if I'm wrong. Thanks!

 

P.S. I think I forgot to thank Mr. Pytlak for his input. Thank you Mr. Pytlak!!

[Giles Rais]

Dear Mr. Case:

 

It all depends on how you understand "middle grey". It's nominally around 18%, which isn't quite in the middle of a standard grey card with white at 100% and black at 3% (5 stops): where is white (100% reflectance?) how bright a hotspot do you want to reproduce accurately; how deep a shadow?. The actual negative density recommended to do this is not just a number (e.g. 0.80R etc), it is a function of the entire curve shape.

<{POST_SNAPBACK}>

 

Well this is precisely want i want to do.  To create camera f/stop scales accurately from correctly measured Log E units.  I want to know exactly how to place areas of my scene at the toe, shoulder, middle gray, etc.  I can't do that until I know Kodak's rationale for placing 0.0 at 0.8 density, although I think I cracked with the ECN-2 doc. 

 

The rationale is about making the film work as well as possible. Exposure placement on the curve is the main, but not the only factor.

.

<{POST_SNAPBACK}>

 

Agreed. 

 

It might appear that the recommended placement is too far down the curve - there is heaps of headroom on the over-exposure side and far less on the under-exposure side - but there is another concept - that of optimal sharpness. Over-exposed negative tends to lose sharpness due to image spread, so the rationale - ultimately - is to expose as little as possible consistent with good shadow detail.

<{POST_SNAPBACK}>

 

I see. Does this mean, Dominic, that I am interpreting what Kodak is doing correctly though? Meaning that I can make my own camera stops scale after measuring Log E units (from threshold to beginning of D-max) by dividing them by .3 and placing the 0.0 at 0.8 density? Will this give me a proper interpretation of how many f-stops of overexposure I have to get to place some area of my scene in the D-max area of the curve? (Or how many f/stops of underexposure I must give my negative to place areas of my scene at D-min, for complete underexposure?). BTW, is your book sold in the US? Thanks again!

Edited February 21, 2005 by Giles des Rais

[David Mullen ASC]

I want to know exactly how to place areas of my scene at the toe, shoulder, middle gray, etc. 

<{POST_SNAPBACK}>

 

You're basically describing the Zone System -- that rarely works in cinematography because you can't change the relative value of objects in the frame easily. I mean, you may decide to expose a face as Zone 6, but that doesn't mean you can force a dark shirt necessarily to be three stops under that. And you can rarely get away with spot lighting every area in the frame until they are at the Zones you want them to be.

 

The whole idea of "placing" objects at different Zones works in the Zone System because you were meant to combine exposure with processing variations in b&w to adjust contrast, so if you had a shot with mostly midtone values, like an overcast foggy day, you could extend the contrast range by changing processing.

 

I personally believe that exposing a scene when making a movie should be as unscientific as possible, which is why we're lucky that modern stocks have such latitude, because it gives us the freedom to light by eye, do some basic metering, and get the scene to look the way it should. I basically meter the key light and set everything else by eye, and if sometimes I make a mistake, I still feel that the alternative is worse because metering everything can make lighting a scene too clinical. It's like the Robert Richardson hot top-back light approach -- I only started to get that effect when I stopped metering the overexposure and did it by eye, because my meter would just go through the roof when I checked the brightest area.

[Giles Rais]

Thank you for your input Mr. Mullen. Yes, I understand exactly what you're saying. However, it seems that nobody can answer exactly what I am asking regarding the placement of 0.0 at 0.85 density by Kodak. I must be expressing myself incorrectly. For this, I apologize. All I wanted to know is:

 

After getting the number of f/stops of over and underexposure a particular stock can give you (by carefully measuring the range value of Log E units from the point where the emulsion starts to respond to light, to where it stops responding to it, at the end of the shoulder and dividing the result by 0.3), does the rationale used BY KODAK regarding the placement of the 0.0 point of the camera stop scale has anything to do with their processing instructions for that stock? In this case, the ECN-2 process designed to create mid densities at R= 0.85, G=1.33, B=1.59? It seems KODAK places the 0.0 of the camera scale right at the 0.85 value. Is this correct? Is this their rationale?

 

I am sorry to keep repeating myself, but none of the answers I have gotten so far answers this question. People seem to understand I am asking about the how to light a scene using the zone system, or how to read the curve. I am not. I just need confirmation that I got the way KODAK is placing their camera stops scale correctly. Thanks again.

[David Mullen ASC]

You're missing the point -- there is no such thing as a "correct" exposure. There is only a recommended one to place the majority of midtone information within the straightline portion of the curve.

[Giles Rais]

You're missing the point -- there is no such thing as a "correct" exposure.  There is only a recommended one to place the majority of midtone information within the straightline portion of the curve.

<{POST_SNAPBACK}>

 

 

Dear Mr. Mullen:

 

I know there is no "correct" exposure. I am referring simply to the way Kodak places 0.0 along the Log E axis, and their rationale for doing so. This is the only point I am trying to elucidate. I know 0.0 can be any f/stop I choose. This misunderstanding would be funny if I didn't need this information as badly as I do. Is the midpoint density created by their processing of color negatives the rationale for placing 0.0 camera stop at the 0.85 density for those stocks? That's all I am asking...

Edited February 21, 2005 by Giles des Rais

[Giles Rais]

I am referring simply to the way Kodak places 0.0 along the Log E axis, and their rationale for doing so.  This is the only point I am trying to elucidate.  I know 0.0 can be any f/stop I choose.  This misunderstanding would be funny if I didn't need this information as badly as I do.  Is the midpoint density created by their processing of color negatives the rationale for placing 0.0 camera stop at the 0.85 density for those stocks?  That's all I am asking...

<{POST_SNAPBACK}>

 

 

So I guess nobody knows the answer to this question? :)

[Dominic Case]

Just typed a long and detailed answer to this, and somehow closed the browser before sending it. D'oh!

 

Film speed used to be measured using the ASA system, which was based on the exposure required to reach a density of 0.10 above d-min provided the average contrast of the curve between that density and a point 1.20 log E above that was 0.62. Worked OK for b/w emulsions in the 60s, (and would have given you a scientific answer to your question - although still quite convoluted).

 

In the 70s, Kodak decided that the curve shape for colour stocks wasn't well described by this system and came up with EI (Exposure Index), which was more about the exposure needed to give a good-looking image, and less about exact curve shapes and logE values.

 

As a result, an 18% grey may not appear at the same density above D-min for all stocks. However, I believe that the 0 point on the Kodak curves does indeed relate to an 18% grey card exposed at the recommended EI.

 

Fuji are more absolute in this: the curves on their brochures have the actual exposure in log (lux-seconds) on the exposure axis. As lux seconds are usually small values (below 1), the log value is negative, and the proper mathematical notation confuses most non-mathematicians. So although theur system is less arbitrary, it's also less useful to the average user.

 

BTW, is your book sold in the US? Thanks again!

Yes. Try amazon or www.focalpress.com, or have your local bookstore get it in : ISBN 0 240 51650 8

[Giles Rais]

As a result, an 18% grey may not appear at the same density above D-min for all stocks. However, I believe that the 0 point on the Kodak curves does indeed relate to an 18% grey card exposed at the recommended EI.

 

Yes. Try amazon or www.focalpress.com, or have your local bookstore get it in : ISBN 0 240 51650 8

<{POST_SNAPBACK}>

 

 

Thank you Mr. Case!

 

yes, but how do they pick that point in the curve for middle gray?

 

 

Your insights seem to confirm my understanding of Kodak's rationale for the placing of 0.0 at 0.85 density (their recommended gamma) for color negative stocks. Confusion comes when, for instance, characteristic curves for 320T have their curves placed at different densities for middle gray depending on what document you check (the .gif for camera stops has the 0.0 much closer to 0.65, while the curve included in the .pdf has 0.0 at about 0.8 gamma). This may be simply shoddy graphics work which I didn't expect to find on such an important document. Check this out, for instance (both curves are for the same stock):

 

 

I checked all other curves for Kodak's color neg stocks and they also have their 0.0 closer to 0.8 and not 0.85 gamma, even though ECN-2 has the red layer at 0.85 (!)

 

I will try to contact someone at Kodak to get final confirmation on this and will post my results here, in case someone else ever needs to understand this.

 

I simply wish there was a way to know how to place the 0.0 consistently for color neg stocks. Again, thanks Mr. Case!

[Patrick Neary]

Will this give me a proper interpretation of how many f-stops of overexposure I have to get to place some area of my scene in the D-max area of the curve?  (Or how many f/stops of underexposure I must give my negative to place areas of my scene at D-min, for complete underexposure?).

<{POST_SNAPBACK}>

 

Ok, here I go jumping in way over my head, but it seems the answer you're looking for also has to take into account some kind of post-production process (ie telecine or printing, not to mention variables in processing) to be informed, unless all you are doing is producing a camera negative that is to be viewed directly.

 

Information placed on either end of the film's exposure range, especially the extreme ends of over and underexposure, won't necessarily survive an IN/IP/Release print flow, for instance.

 

Kudos for an interesting thread!

[Dominic Case]

Information placed on either end of the film's exposure range, especially the extreme ends of over and underexposure, won't necessarily survive an IN/IP/Release print flow, for instance.

Absolutely right. The original question was about placing an exposure on the toe or shoulder of the negative curve, so the discussion so far has been on thrack - but you are right that you don't necessarily need to put something right on the dmax or dmin part of the negative for it to block out to white or black in the print (or telecine). Grading (timing) is essentially selecting the required part of the tonal range captured on the negative that you want to include in the print. A darker print will reveal highlight detail from the upper reaches of the neg curve, that would otherwise print onto the toe or dmin of a normal print for example, while pushing some well-recorded shadow detail from close to the toe of the negative, up to the upper reaches of the print curve (not necessarily on the shoulder of the print, but possibly too dark to see. It's a complex set of considerations.

 

Perhaps it's time to shoot a grey card and see what happens if you over/under expose it, or print it up or down. A grey card only goes from about 1.5% reflectance up to 100% (if you are lucky) - 6 stops: but if you light it with a 4-stop shadow along its length, you get 10 stops range.

[John Pytlak RIP]

Information placed on either end of the film's exposure range, especially the extreme ends of over and underexposure, won't necessarily survive an IN/IP/Release print flow, for instance.

 

<{POST_SNAPBACK}>

 

That's why an integral part of using Kodak's Laboratory Aim Density (LAD) system (which I developed) is using a color analyzer to do scene-to-scene timing. That puts the normal scene tonal range of 160:1 (average) or even up to 1000:1 (typical maximum) on the straight line portion of the duplicating film's curve.

 

But if you print "one-light", then grossly over or underexposed scenes may place off the straight line portion of the master positive.

[Giles Rais]

Ok, here I go jumping in way over my head, but it seems the answer you're looking for also has to take into account some kind of post-production process (ie telecine or printing, not to mention variables in processing) to be informed, unless all you are doing is producing a camera negative that is to be viewed directly.

 

Information placed on either end of the film's exposure range, especially the extreme ends of over and underexposure, won't necessarily survive an IN/IP/Release print flow, for instance.

 

Kudos for an interesting thread!

<{POST_SNAPBACK}>

 

This thread was about the rationale used by KODAK (and not me) to place the 0.0 camera stop along the Log E axis, which seems to be a complete mystery so far (unless I am right and their rationale is indeed the mid density point for the specific processing suggested by them, which for color negatives is 0.85 for the red layer using ECN-2).

Edited February 25, 2005 by Giles des Rais

[John Pytlak RIP]

This thread was about the rationale used by KODAK (and not me) to place the 0.0 camera stop along the Log E axis, which seems to be a complete mystery so far (unless I am right and their rationale is indeed the mid density point for the specific processing suggested by them, which for color negatives is 0.85 for the red layer using ECN-2).

<{POST_SNAPBACK}>

 

The exposure chosen for the middle (11th) step of a 21 step sensitometric exposure is somewhat arbitrary. For camera negative films, the sensitometric exposure time is usually 1/50 second (equivalent to 24fps with a normal shutter). The tungsten lamp of the sensitometer is filtered to the appropriate color temperature. The exposure intensity is then chosen such that the sensitometric curve has at least a step or two of "d-min" (no density produced by the exposure). That sensitometric "set up" is used for all color negative films having that Exposure Index and color balance (e.g., all EI 100T films). A speed rating (EI) can be calculated since the exposure (in lux-seconds) is known by measurement with a calibrated photometer.

 

I don't know of ANY specification that the mid density point should be 0.85 Status M red density for all color negative films. Perhaps a particular control strip had a 0.85 red density, but other color negative films would likely be different.

 

Here is a typical curve:

 

http://www.kodak.com/global/images/en/moti...ive/ti1667b.gif

 

[Giles Rais]

Dear Mr. Pytlak

 

Unless I am reading your post wrong, you are saying that the placement of the 0.0 camera stop along the Log E axis is arbitrary, and it varies depending on (in my case) the specific color negative stock we are looking at, regardless of the fact that they are all processed to the same mid-density standard (ECN-2, in this case). I must say I am surprised to find this out, especially when something as important as knowing how many stops of over and underexposure a particular stock will have in either direction would seem something very important to me, especially when some stocks have more stops of over exposure than underexposure. What about all those stocks that do not include a camera stops scale included with the characteristic curve? One could find how many stops it can hold in the straight line, toe, shoulder, D-min and D-max, but without a specific system for placing the 0.0 camera stop along the Log E axis one would be forced to say something like: "ok, this stock has a range of 10 camera stops, so that must be divided equally as having 5 stops of over and 5 stops of underexposure on either side of 0.0 (or is it 4 1/2 of over, 4 1/2 of under and 0.0?)" even though this may not be the case...interesting. Is this the final word on this? Is there no system for calculating this ourselves?

[David Mullen ASC]

Just shoot a test.

[Giles Rais]

Just shoot a test.

<{POST_SNAPBACK}>

 

I'd rather find out the info I'm looking for, if that's ok...(bit anal about tying loose ends, I'm afraid).

[Patrick Neary]

Hi Giles-

 

I'm very curious, is this pursuit an excercise in arcane discussion (I don't mean that in a bad way) or is it aimed toward some real-world practical purpose? (because I'd love to know what that purpose is!)

[John Pytlak RIP]

I'd rather find out the info I'm looking for, if that's ok...(bit anal about tying loose ends, I'm afraid).

<{POST_SNAPBACK}>

 

The exposure a sensitometer gives the film SIMULATES the exposure in the camera. The spectral quality (color temperature) of your light is almost certainly different than a low color temperature tungsten lamp in the sensitometer filtered to a nominal 3200K. Your ECN-2 process is slightly different too. The "0" point on the Camera Stop scale may not represent what you consider normal exposure for the "look" you want.

 

That is why Kodak prints a disclaimer on the published data:

 

NOTICE: The data in this publication represent product tested under the conditions of exposure and processing specified. They are representative of production coatings, and therefore do not apply to a particular box or roll of photographic material. They do not represent standards or specifications that must be met by Eastman Kodak Company. The company reserves the right to change and improve product characteristics at any time.

 

It's always best to test a film under the actual conditions of use.

 

As Mr. Mullen and others have stated: "Just shoot a test".

[Giles Rais]

"The "0" point on the Camera Stop scale may not represent what you consider normal exposure for the "look" you want"

 

OK, thanks to everyone who tried answering my question. I'll keep looking until I find the answer and post my findings back here. Thanks again!

 

Dear Patrick: in answer to your question, it's for something very important. ; )