TTL metering and filters

[garpet]

I've noticed a lot of conflicting information about how TTL meters interact with filters, especially orange and red filters. I'm mostly focusing on the light sensors used in the nikon DP line that comes with their F2 cameras since that's what's familiar to me, but can't find what specific sensors any use (other than the vague 'this one uses CdS, this one uses Si). Looking at some technical specs for other photosensors, it looks like CdS meters have their highest response to wavelengths between 450 and 580nm or blue to yellow light, which should at worst cause them to overexpose when using a red light filter, but I've heard the complete opposite from other photographers. On the other hand, while silicon light meters seem to very wildly in their spectral response depending on the model they tend to have their peak sensitivity closer to the infrared range, though some show a very similar response to CdS cells when designed to mimic human eyesight. This would make it seem like CdS cells should much more accurate for metering light through red or orange filters than Si cells, but again, I have heard the complete opposite before. Does anyone know where I might find the actual data sheets for the sensors used in the DP line, or have any experience or input with how they respond to different light filters? I've reached out to Nikon about getting the data sheets and will update if they send them to me, but I'd like to hear your experiences/observations or if anyone has any information.

 

[Paul Howell]

This question has been debated for years, some say that meters, TTL likr handheld meters need to be corrected using filter factors. Others say the difference is so small as not be worth the trouble. I looked at the manuals for my TTL bodies starting with Pentax Spotmatic, I have yet to see that the manufacturers recommend using filter factors, while some filter makers such as Tiffin in the past have recommended using Filter Factors.

 

[gbroadbridge]

I find a couple of test shots taken with the camera metering with and without the filters provides the necessary correction very quickly.

 

[Hassasin]

Measure through filter, never done it any other way, never noticed an issue doing so.

 

[garpet]

I find a couple of test shots taken with the camera metering with and without the filters provides the necessary correction very quickly.

This is what I've done in the past, but it wouldn't fix the issue with infrared sensitivity leading to the meter reading a higher light level through the filter than what will actually react with the film. It also requires shooting a full test roll to see the results which seems wasteful. My go to is to just read the light meter without a filter, then see if the meter reads around the expected filter factor with the filter on, but it would be more convenient to be able to predict how a given light sensor should react to a given filter prior to actually getting my hands on the filter.

 

[Paul Howell]

In this case I would use a handheld meter and use filter factors.

 

[gbroadbridge]

This is what I've done in the past, but it wouldn't fix the issue with infrared sensitivity leading to the meter reading a higher light level through the filter than what will actually react with the film. It also requires shooting a full test roll to see the results which seems wasteful. My go to is to just read the light meter without a filter, then see if the meter reads around the expected filter factor with the filter on, but it would be more convenient to be able to predict how a given light sensor should react to a given filter prior to actually getting my hands on the filter.

If I were testing a new film stock like IR my process would be to burn a roll of film and get all the variables tested.

So all possible filters and all possible lenses which I would mark (non permanent marker) with some adjustments around the expected correct focus point.

I just consider the cost of one burned roll likely to be insignificant if I'm planning to shoot a lot of that stock in future.
In my experience it just works out cheaper in the long run to get everything documented.

 

[garpet]

If I were testing a new film stock like IR my process would be to burn a roll of film and get all the variables tested.

So all possible filters and all possible lenses which I would mark (non permanent marker) with some adjustments around the expected correct focus point.

I just consider the cost of one burned roll likely to be insignificant if I'm planning to shoot a lot of that stock in future.
In my experience it just works out cheaper in the long run to get everything documented.

probably a good idea. Do you mark the filter itself so it shows in the exposure? How do you do that without affecting the metering and keeping the writing clear in the exposure? or do you mean marking the proper adjustment on the rim of the filter. Maybe I'll just write the filter and variables on a card and have that somewhere in the test shot

 

[Taylor Nankervis]

In this case I would use a handheld meter and use filter factors.

^ Way to go.

Rarely do I use a filter other than as circular polariser with my EOS 1N in Evaluative metering mode; just as with matrix, multipattern and other proprietary-named systems, metering with the filter in place will not require any additional compensation. The same cannot be said for a circular polariser however, that will require anything from 1.5 to 2.5 stop over what the meter suggests. In that case, taking careful notes of what you are doing with each exposure to "call-back" when your view your work on the lightbox is an invaluable inclusion in the workflow.

In medium format (Pentax 67), handheld multispot metering with a CIR-POL involves a +2.0 filter factor (and 0.6 to full POL) and this has served my purposes (transparency-to-print) excellently over a very long period of time now.

Historically I have botched a couple of rolls of faux-IR film, and didn't cotton-on to a lasting affinity with it. In Australia old IR stocks, well past expiry, remain popular with some LF practitioners for creative effect in the landscape genre.

 

[r_a_feldman]

This is what I've done in the past, but it wouldn't fix the issue with infrared sensitivity leading to the meter reading a higher light level through the filter than what will actually react with the film. It also requires shooting a full test roll to see the results which seems wasteful. My go to is to just read the light meter without a filter, then see if the meter reads around the expected filter factor with the filter on, but it would be more convenient to be able to predict how a given light sensor should react to a given filter prior to actually getting my hands on the filter.

I think that IR is in a different class than visible light. With IR you need to take a meter reading without the filter (e.g., with a hand-held meter)) and then manually apply the filter factor recommended for the particular filter and film.

 

[Paul Howell]

As I recall with true IR films, not the newer simulated IR like films, Kodak directions were not to use a TTL meter or any meter, but use the exposure table that came with the film.

 

[DREW WILEY]

Comparing TTL results using my Nikon FM2n with a red filter attached, and reading the scene with my spot meter and applying a 3EV filter factor instead, the latter method is much more reliable. One might get away with readings through relatively mild contrast filters; but as they get denser, the risk of significant error becomes greater. And I've tested for this many times, with a variety of films, and have even quantified it using a densitometer afterwards.

Better polarizers are basically colorless, so pose a different scenario in relation to TTL sensors. Still, I prefer to do readings with a handheld meter and apply factors. I don't even do TTL metering with my Pentax 6x7 anymore.

 

[skahde]

You never know if your film has the same spectral sensitivity as a lightmeter inside or outside your camera for a filtration until you test it with the film, camera-type/meter and filter in question. Change one variable and you are back where you started. It is much easier to work with the correction-factors given by the film-manufacturers as a starting point and meter without the filter. Then you will find that with the more extreme filters the actual light and color and structure (shadow-areas) also may play a big part in evaluating and measuring a scene.

With lighter filtrations, where you are compensating no more than one stop you will get away with measuring through the filter but latitude still helps a lot with this. With more extreme filters you need all the help you can possibly get, external metering and previous testing under different lighting conditions included.

 

[MattKing]

The answer to the question is different if one is exposing IR sensitive film than it is with regular panchromatic film.
When it comes to IR sensitive film, the meters we use don't really respond to the IR in the scene.
Instead, they respond to the visible light available, and we use our experience and judgment to determine how much IR light is likely to accompany that visible light. That determines the conversion we use to convert the visible light reading recommended exposure to an exposure that records the IR.
We don't use a filter factor per se, because the role of the filter with pure IR photography is to block all or nearly all the visible light.
If a photographer is using something like a deep red filter with currently available IR sensitive films, than it may be necessary to also incorporate into the calculations a filter factor relating to the filter's transmission of visible light.
In neither case though, will a TTL filter do what you need, because a TTL filter only reads that visible light, and the reading won't incorporate the necessary visible-IR light correspondence calculation.

The issue with non-IR film is more nuanced. Filters serve to partially block certain colours of light, and by so doing, skewing the colour response of the system. That can have a surprising effect on how much exposure we need to give film. If, for example, a scene is illuminated with a lot of blue light in the shadows, but a lot of warmer light in the highlights, a yellow filter may have little effect on the highlight exposure needed, but a large effect on the shadow exposure needed. TTL meters may be fooled by that, whereas published filter factors may be more likely to take that into account.

Personally I prefer using a hand-held meter and a filter factor.

 

[koraks]

Does anyone know where I might find the actual data sheets for the sensors used in the DP line, or have any experience or input with how they respond to different light filters?

I don't think this data is published anywhere.
Keep in mind that when comparing meter cells, you shouldn't just focus on the type (CdS vs. Si) as the spectral sensitivity will in practice be influenced very strongly by a filter layer on top of the actual semiconductor cell.

 

[gary mulder]

In practice overcompensating the exposure when using a color filter in BW photography will counter effect the filter. If for example you have a large portion of bleu sky that you want te render darker than excess longer exposure will leave you with the same grey tint in the sky as without the yellow - red filter

 

[koraks]

In practice overcompensating the exposure when using a color filter in BW photography will counter effect the filter. If for example you have a large portion of bleu sky that you want te render darker than excess longer exposure will leave you with the same grey tint in the sky as without the yellow - red filter

But the tonal relationships will still be altered. The absolute density of that sky portion may end up as high as it might have without a filter, but the green foliage below it will have a higher density relative to the unfiltered version, so the sky will still render proportionally darker in the final edit/print. The net effect also depends on how far the tonal values are pushed onto the shoulder of the film; if this happens to an extreme extent, overall contrast will of course suffer. But even then, the tonal relationships with and without a filter will be different.

 

[Petrochemist]

As I recall with true IR films, not the newer simulated IR like films, Kodak directions were not to use a TTL meter or any meter, but use the exposure table that came with the film.

Every source I have on IR film says you should bracket widely. Fortunately with digital (how I shoot my IR) this is not an issue, test shot & corrected shot for SLR cameras & direct TTL metering for my mirrorless models (where metering is done from the main sensor).

Nothing in the visible gives a reliable indication of the IR levels, unless perhaps if the lighting is entirely from known artificial sources

 

[gary mulder]

But the tonal relationships will still be altered. The absolute density of that sky portion may end up as high as it might have without a filter, but the green foliage below it will have a higher density relative to the unfiltered version, so the sky will still render proportionally darker in the final edit/print. The net effect also depends on how far the tonal values are pushed onto the shoulder of the film; if this happens to an extreme extent, overall contrast will of course suffer. But even then, the tonal relationships with and without a filter will be different.

You're right.

 

[skahde]

Every source I have on IR film says you should bracket widely.

That's how I started but with experience I found that for me exposing eg. SFX like it was ISO 6 with a filter 89B/RG695 does the trick. The more IR there is in the scene the farer up the density-curve the highlights go but you will always find at least some faint detail in the shaddows if you measure for Zone III (detailed shaddow) where you can neglect the influence of IR. This will in some cases increase the contrast you have to handle in the darkroom but as you cant reliably control the IR-level anyway, it seems logical to me to apply control where it's possible: Under the enlarger.

 

[Chan Tran]

In this case I would use a handheld meter and use filter factors.

You could use the camera meter without the filter then apply the filter factors.

 

[Oldwino]

My FM3a underexposed considerably when using a red filter and metering thru the camera. Yellow and orange were fine, but with a red filter, it was better to use a handheld meter and figure the compensation manually.

 

[Chan Tran]

My FM3a underexposed considerably when using a red filter and metering thru the camera. Yellow and orange were fine, but with a red filter, it was better to use a handheld meter and figure the compensation manually.

As I said in the post above can you just meter with the FM3a without filter then mount the filter and apply the filter factor manually instead of using a hand held meter?

 

[DREW WILEY]

Of course you can. The FM3a internal meter involves a silicon cell quite similar to Pentax and Minolta spot meters, and will give very similar readings under controlled circumstances. The problem is more the angle of view involved. I just shoot a FM2n now, but when I had an FM3a, with its extra features including a little bit of automated exposure, the results were rarely as predictable and accurate as with handheld metering. That was particularly a problem when exposing fussy slide films. I just couldn't trust it. Black and white film can be a little more forgiving, but doesn't respond as well to averaged metering. It's better to know exactly where your shadow readings versus highlight reading are going to land.

The specific question, concerning red filters in particular, is pertinent. That's because these silicon cells have their peak sensitivity at green, just like the human eye has. An additional factor can be the specific film itself, and how much red sensitivity it has before there's a spectral cutoff. Red filters differ from one another too, like the distinction between a common 25 red and 29 deep red. In other words, with my Nikon TTL metering, if I happen to have a filter in place, the odds of getting a correct reading are a lot better with a yellow, orange, or even light green filter on the lens than a red one.

 

[wiltw]

It is easy enough to TEST yourself...

1. aim camera meter at a blank uniformly illuminated wall (ideally when illumination is not fluctuating, perhaps artificially illuminated) and
2. take a reading (without filter), apply filter and take another reading.
3. Compare that experimental difference vs. the stated filter factor, and draw your own conclusion for that filter.

Repeat test method with other filters. Figure out if the meter reading varies by the same or different amount from the filter factor...then you know which filters work with TTL meters, and which ones do not.