TTL metering and filters

[Andrew O'Neill]

Well, I do mostly large format, so it's handheld metering, and filter factors for me... same with the RB. I have also found that some films (TMY, HP5) respond a bit differently to filters, especially the red #25.

Edit: forgot to add that David Kachel wrote an excellent article about film's response to different filters, in a 90's issue of Photo Techniques. If you can find it, I highly recommend it.

 

[MattKing]

What do you do with a polarizing filter where the filter factor can vary 1 to 2 stops depending upon the position of the filter?

Generally speaking, a polarizing filter has little effect on the shadows and a variable effect on the highlights.
So if you are basing your exposure decisions on highlight rendition - as you might be doing @Alan Edward Klein when you are using slide film - then the exposure adjustment (filter factor) will vary with the effect you are choosing with the adjustment of the polarizer.
If you are basing your exposure decisions on shadow rendition - as many negative film users do - then a less variable exposure adjustment (filter factor) is probably in order. I would suggest one stop.
This highlights the challenge of using TTl metering through the filter - at which position of the filter do you take the reading when you are exposing black and white negative film?

 

[DREW WILEY]

Polarizers typically require 1-1/2 stops correction. The problem is, that TTL metering can be fooled. And in color film usage, you potentially change the response of colors themselves, especially blue.

A few days ago, I stumbled onto a completely pristine 67mm B&W linear polarizer that I had completely forgotten I owned. And it was just the right kind of day to play with it, with some stunning clouds on the shoreline following a storm which cleared the air nicely and left the sky blue. It grabbed my Pentax 6x7 and took a few relevant shots on a nice afternoon walk beside the Bay. Since that's an SLR, it's easy to see the effect of the polarizer at any particular angle of view; it would have been a lot tricker if I had taken my Fuji "Texas Leica" rangefinder instead. But I'll have to await developing the film to know if I ciphered all the exposures correctly or not.

 

[wiltw]

Interesting what one can discover with just a bit of experimentation taking place at Latitude 38, 2:30pm Pacific Daylight Time ...I just took a Minolta Spotmeter F, and a B+W CPL (and ensuring that I was reading from the lens side of the CPL)

1. Indoors, with absolutely flat light I aimed the meter without CPL, I took a reading of an absolutely not g;ossy surface I was able to register relative readings -1.0EV with the CPL
2. in the SHADE, I aimed the meter , I took a reading of an absolutely flat (no sheen) cloth with no filtration set that as baseline 0.0EV
   placing filter in front of the spotmeter lens and rotating the CPL I was able to register readings ranging between -1.0EV and -2.0EV
3. in the SUNLIGHT, I aimed the meter, I took a reading of an diffuse paver surface with no filtration and set that as baseline 0.0EV
   placing filter in front of the spotmeter lens and rotating the CPL I was able to register readings ranging between -1.0EV and -2.0EV
4. in the SUNLIGHT, I aimed the meter , I took a reading of the blue sky oriented where polarizer would most significantly darken the sky, with no filtration and set that as baseline 0.0EV
   placing filter in front of the spotmeter lens and rotating the CPL I was able to register readings ranging between -0.7EV and -2.6EV

Now the speculation can begin...

1. (works as expected)
2. as to why -- in the shade -- we can measure any change as we rotate the CPL measured in the diffuse light of the shaded area
3. as to why -- in the sunlight pointed at a diffuse surface on the ground we can measure the change as we rotate the CPL
4. as to why a smaller minimum loss is observed (bullet #4 vs bullet #1) -0.7EV vs -1.0EV

 

[Alan Edward Klein]

Interesting what one can discover with just a bit of experimentation taking place at Latitude 38, 2:30pm Pacific Daylight Time ...I just took a Minolta Spotmeter F, and a B+W CPL (and ensuring that I was reading from the lens side of the CPL)

1. Indoors, with absolutely flat light I aimed the meter without CPL, I took a reading of an absolutely not g;ossy surface I was able to register relative readings -1.0EV with the CPL
2. in the SHADE, I aimed the meter , I took a reading of an absolutely flat (no sheen) cloth with no filtration set that as baseline 0.0EV
   placing filter in front of the spotmeter lens and rotating the CPL I was able to register readings ranging between -1.0EV and -2.0EV
3. in the SUNLIGHT, I aimed the meter, I took a reading of an diffuse paver surface with no filtration and set that as baseline 0.0EV
   placing filter in front of the spotmeter lens and rotating the CPL I was able to register readings ranging between -1.0EV and -2.0EV
4. in the SUNLIGHT, I aimed the meter , I took a reading of the blue sky oriented where polarizer would most significantly darken the sky, with no filtration and set that as baseline 0.0EV
   placing filter in front of the spotmeter lens and rotating the CPL I was able to register readings ranging between -0.7EV and -2.6EV

Now the speculation can begin...

1. (works as expected)
2. as to why -- in the shade -- we can measure any change as we rotate the CPL measured in the diffuse light of the shaded area
3. as to why -- in the sunlight pointed at a diffuse surface on the ground we can measure the change as we rotate the CPL
4. as to why a smaller minimum loss is observed (bullet #4 vs bullet #1) -0.7EV vs -1.0E

At minimum, you lose some light even at minimum due to the way the CPL blocks light. It increases in blockage as the angle offset to the light, the closer you get to 90 degrees to the light. That would be a 2-stop factor. I wouldn't use the -2.6 adjustment because that's the sky darkening. You want that darkening effect, so set it for 2 stops only.

Interesting that my 1 to 2 stop guestimate given earlier closely matches your results. So just check to see how much you turn the filter to estimate a factor. No polarizing would be 1 stop. Halfway would be 1.5 stops. Full 2 stops.

 

[DREW WILEY]

Depends what one is hoping to achieve, relative to the scene, as well as what kind of film is involved, as well as if any additional filtration is in place.

 

[Taylor Nankervis]

Is it more accurate to use an exposure meter through the polarizing filter?

It is an option.
If using a handheld multispot/incident meter, in spot metering, a polariser can be placed over the lens and the meter will not require a FF correction in that useage case. But finding a CPOL in such a tiny size to fit a spot meter's lens is quite a mean feat; I was lucky to find just a Skylight 1B as a "protection" filter in a "lucky dip" for the front element!

 

[Taylor Nankervis]

What do you do with a polarizing filter where the filter factor can vary 1 to 2 stops depending upon the position of the filter?

Individuals approach this in various ways. But there must be method and observation of results rather than a slapdash approach and tears on the lightbox at images where the indiscriminate use of a polariser has well and truly killed the atmosphere. We've all seen this in rookie imaging yet it is surprisingly still one of the most commonly least understood skills for beginners.

In the early parts of using a CPOL with a Pentax 67 (and RVP50 prediominantly), metered with a Sekonic L758D, full POL shots were recorded in detailed notes and progressively bracketed (same scene for each shot). The scenes with varied stepped filter factors (beginning with +0.5 and ending with +2.5) were then scrutinised on the lightbox and came out as +1.5 (start) for mild hazy and/or overcast conditions to +2.0 for heavily overcast flat light. The situations I shoot in are unique, involving enclosed rainforests where the effect of full polarisation excludes any detrimental effect that could be shown in the sky. Now, those early tests were many, many years ago and I have never, not once, felt a need to give anything more than +2.0 as a filter factor! On occasion, such as printing yesterday, it was found that an additional +0.5 stops of brightness (viewing 6 history steps on screen) was required due to illumination loss at the scan-step. This is not a reflection the exposure with a polariser, but an endemic feature in analogue-to-digital printing.

 

[DREW WILEY]

I only used polarizing filters in relation to copystand work when glossy originals (like Cibachrome) needed cross-polarized light to tame the reflections. But there was always some penalty in accurate color repro. Same goes in nature. Their use outdoors can be overdone, and start looking fishy.

As I already mentioned, I was playing around with one outdoors for fun a few days ago; but that involved b&w Tmax film and a supplemental 25 red filter. I removed my meter prism from the P67 long ago, and now use just the ordinary prism finder, since I use a handheld meter for everything anyway. I factored 1-/2 stops for the pola filter itself, based on earlier tests, but fudged a little on the overexposure side, just in case.

 

[Petrochemist]

It is an option.
If using a handheld multispot/incident meter, in spot metering, a polariser can be placed over the lens and the meter will not require a FF correction in that useage case. But finding a CPOL in such a tiny size to fit a spot meter's lens is quite a mean feat; I was lucky to find just a Skylight 1B as a "protection" filter in a "lucky dip" for the front element!

The filter doesn't need to be the right size. A suitable step up ring will allow normal filters to attached or you can just hold the filter in place
Either way though matching the alignment of a meter polariser to that used on the camera can be important.

Note polarisers are available in both high extinction & high transmission types, the amount of light passing through will vary significantly between these.

 

[wiltw]

Interesting what one can discover with just a bit of experimentation taking place at Latitude 38, 2:30pm Pacific Daylight Time ...I just took a Minolta Spotmeter F, and a B+W CPL (and ensuring that I was reading from the lens side of the CPL)

1. Indoors, with absolutely flat light I aimed the meter without CPL, I took a reading of an absolutely not g;ossy surface I was able to register relative readings -1.0EV with the CPL
2. in the SHADE, I aimed the meter , I took a reading of an absolutely flat (no sheen) cloth with no filtration set that as baseline 0.0EV
   placing filter in front of the spotmeter lens and rotating the CPL I was able to register readings ranging between -1.0EV and -2.0EV
3. in the SUNLIGHT, I aimed the meter, I took a reading of an diffuse paver surface with no filtration and set that as baseline 0.0EV
   placing filter in front of the spotmeter lens and rotating the CPL I was able to register readings ranging between -1.0EV and -2.0EV
4. in the SUNLIGHT, I aimed the meter , I took a reading of the blue sky oriented where polarizer would most significantly darken the sky, with no filtration and set that as baseline 0.0EV
   placing filter in front of the spotmeter lens and rotating the CPL I was able to register readings ranging between -0.7EV and -2.6EV

Now the speculation can begin...

1. (works as expected)
2. as to why -- in the shade -- we can measure any change as we rotate the CPL measured in the diffuse light of the shaded area
3. as to why -- in the sunlight pointed at a diffuse surface on the ground we can measure the change as we rotate the CPL
4. as to why a smaller minimum loss is observed (bullet #4 vs bullet #1) -0.7EV vs -1.0EV

Since there have been no responses to the above post, I will state specific questions in the hopes of getting responses to my puzzlement...

1. Why does a diffuse (with no surface sheen) surface exhibit any range of readings, between -1.0EV and -2.0EV, in the sunlight or in the shade? (above bullets 2,3)
2. If the fundamental light loss thru a CPL is -1.0EV, why does a measurement of the sky have a range minimum loss reading of -0.7EV , as the CPL is rotated (above bullet 4 vs bullet 1)

 

[wiltw]

At minimum, you lose some light even at minimum due to the way the CPL blocks light. It increases in blockage as the angle offset to the light, the closer you get to 90 degrees to the light. That would be a 2-stop factor. I wouldn't use the -2.6 adjustment because that's the sky darkening. You want that darkening effect, so set it for 2 stops only.

Interesting that my 1 to 2 stop guestimate given earlier closely matches your results. So just check to see how much you turn the filter to estimate a factor. No polarizing would be 1 stop. Halfway would be 1.5 stops. Full 2 stops.

Makes sense that the overall exposure loss is not represented by solely what occurs to the darkening of the blue sky. But post #61 is the reason for any puzzlement on my part.

 

[koraks]

Why does a diffuse (with no surface sheen) surface exhibit any range of readings, between -1.0EV and -2.0EV, in the sunlight or in the shade? (above bullets 2,3)

Because it's not actually diffuse. Not 100%.

If the fundamental light loss thru a CPL is -1.0EV, why does a measurement of the sky have a range minimum loss reading of -0.7EV , as the CPL is rotated (above bullet 4 vs bullet 1)

Because your CPL is as perfect as that diffuse surface - i.e., not entirely.

I think there's a technical reason for the latter as well; ever noticed how even a cheap pair of polarized sunglasses is just a little more effective than a good CPL dialed in just right?

 

[DREW WILEY]

Depends on how much you want to delve into it, Wilt. There are plenty of web resources. Edmund Scientific has a good one, Introduction to Polarization.

Polarizing filters can certainly make look surfaces look "flat" if you overdo it; that's one of their liabilities.

 

[Petrochemist]

I think there's a technical reason for the latter as well; ever noticed how even a cheap pair of polarized sunglasses is just a little more effective than a good CPL dialed in just right?

I suspect that would be because the sunglasses use high extinction polarisers, while most filters are high transmission types.

 

[Taylor Nankervis]

[...]

Note polarisers are available in both high extinction & high transmission types, the amount of light passing through will vary significantly between these.

Yes indeed. In both my 67 and 35mm kits, on all L-series or highly corrected lenses, I don't skimp on the quality of filters slapped on the front. I have a few of B+W's Kaesemann high transmission polarisers. In any professional practice, mishaps occur and I have dropped (and shattered) at least three historically. At more than $500 in Australia, that is an unwelcome hit...

 

[koraks]

I suspect that would be because the sunglasses use high extinction polarisers, while most filters are high transmission types.

That makes sense; I assume high extinction in practice comes at the cost of lower transmittance for light polarized parallel to the filter's direction, resulting in a net light loss in all scenarios, which would be undesirable for most photography. I've sometimes wondered about the availability of filters more similar to those used in (cheap) sunglasses since they are much more effective on e.g. sunlit clouds than polarizers generally used for photography (including the B+W Käsemann filters that I used years ago as well).

 

[Alan Edward Klein]

As some have mentioned, too much polarization can wash out the life in some pictures, especially in the foliage. I've found it's better to back off a little from full polarization.

Here's a digital picture with polarization at maximum. Notice how flat the greens are and how lifeless. It looks unreal. The filter at 100% removes all the reflections from the leaves.

 

[BrianShaw]

As some have mentioned, too much polarization can wash out the life in some pictures, especially in the foliage. I've found it's better to back off a little from full polarization.

Here's a digital picture with polarization at maximum. Notice how flat the greens are and how lifeless. It looks unreal. The filter at 100% removes all the reflections from the leaves.

I believe you, but can you show an example of both conditions: full polarization and same scene with less polarization? To my eye, and looking at it in small size on the monitor, the greens in that image looks pretty good.

 

[DREW WILEY]

One reason I generally eschew polarizing filters is that they can crush underfoot all the sparkle and life out of a scene.
In Alan's example, you have the color green everywhere, yes; but it looks like paste, as if monotonous tempera paint,
and not like a realistic luminous setting at all. All the nuance is gone. Even the reflections in the water look dull and lifeless.

 

[Taylor Nankervis]

As some have mentioned, too much polarization can wash out the life in some pictures, especially in the foliage. I've found it's better to back off a little from full polarization.

Here's a digital picture with polarization at maximum. Notice how flat the greens are and how lifeless. It looks unreal. The filter at 100% removes all the reflections from the leaves.

I'm viewing the post on a non-colour corrected monitor (spare, at home) so I could be some way out by way of accuracy. Having said that, the tone and differentiation of the greens in Alan's pic do look quite flat and junked; the reflection in the water is what I would expect to see, but again, the tone is flat and uncomplementary. Open landscapes like this in bright sun using a polariser are a subject I would (if so inclined, and I am not) bracket several times, also with varying degrees of polarisation, and likely not full polarisation on account of the presence of the sky (especially irksome with e.g. a Pentax 67 45mm lens) and the "kill effect" endemic with indiscriminately applied POL. Most of my (enclosed) rainforest photography is on full polarisation (and filtor-factored exposure, +0.5) to bring out the character of the greens that are endemic to such scenery — there is usually no sky, unless I point the camera straight up to the crown of the forest, but that informs the viewer of nothing much in particular.

Sometimes I aim to preserve some spectrals in rainforests, particularly where rivers and creeks are involved; generally if they are present — glaring spectrals — I come darned close, if not fully nuking 'em!

________________________________________________________________________________
• Partial (0.6) POL to preserve small spectrals, enclosed, low light environment.