Exposure compensation with color filters: is TTL meter reliable?

[Kabouik]

Hi there,


I'm a happy new owner of a Bessa R3A. I've purchased an orange filter to mount on my 50 F/1.1, and I'm now considering a medium yellow one fot the 35 F/1.4 (yeah I know, 35 mm is a weird choice for a R3A, but it was a used one at a good price, and I like the 1:1 viewfinder of the R3A).

My question is in the title. I have read everything and its opposite about AE compensation through the lens for color filters, here and on other boards like photo.net. However, each time people have not really discussed about it, only one or two users answered the same thing and that was it, so it is difficult to figure out who was right and who was wrong. The topics are old (and I can't find them anymore), so I'm trying to get up to date advice here! :]

I know films are quite tolerent and can deal with subtle deviation to exposition, but hey, I'm interested in an answer anyway. Until now I've been overexposing by +1 on the R3A when using the orange filter. I've read somewhere here that the TTL will detect light differences with the filter on, but not the -2 it actually induces; hence the manual compensation by 1 stop. Is that right or should I change the technique? What about a yellow filter, will the TTL meter be more accurate?

Thanks in advance for your answers!

 

[furcafe]

In my experience, meters do not have same spectral response to colors as do your choice of sensor (eye, film, digital sensor) & there's also the fact that ambient color temperature is variable, so I would start w/compensating per the manufacturers markings after testing your camera's meter. Is your R3A's meter giving you a 2 stop difference when you mount the filter in the lighting conditions you're shooting?

 

[Kabouik]

Actually you make a good point that I have read that the meter will partly compensate on these forums, but had not received my filter yet at this time. Since then, I did not even check it!

Here's the result: no compensation at all (inside light, should try with sunlight tomorrow). I have scanned two Tri-X 400 rolls overexposed by one stop when using the filter, developped with Rodinal 1:100 (one stand and one with 16° C R09 and frequent agitations) and they came out fine, especially with cold Rodinal (I got good tonality and no overexposed lights). Now I'm wondering whether I should overexpose more if I'm already getting good lights (orange filter are supposed to induce a 2 stop loss). Perhaps lights could be higher without burning, but trying is a bit risky. I'm even more wondering what to do with the medium yellow filter I will receive soon.

 

[Dwig]

In my experience, meters do not have same spectral response to colors as do your choice of sensor (eye, film, digital sensor) & there's also the fact that ambient color temperature is variable, so I would start w/compensating per the manufacturers markings after testing your camera's meter. Is your R3A's meter giving you a 2 stop difference when you mount the filter in the lighting conditions you're shooting?

+1

TTL meters will do a great job of compensating for filters that correct bad lighting (e.g. daylight <> tungsten light balancing filters, ...) but don't do a very accurate job of correcting for narrow spectrum band pass filters (e.g. filters like the orange and red used with B&W). With the lighter B&W filters, like the #8 yellow, the compensation is small and TTL meters generally do an adequate job. With the stronger (read: narrower portion of the spectrum passed) filters like the #25 red there can be a significant error. How large an error will vary with the meter cell type and the film in use.

Also, metering through the filter can lead to "subject failure". If you intend to darken the sky by using a #25 red and meter with a fair amount of sky in the meter's field of view the result will be a more normal sky with the rest of the scene lighter than expected.

 

[Kabouik]

Thanks a lot for your answer. So I guess I'll keep correcting the orange filter by one stop, while I won't apply compensation for the #8 yellow I mount on the 35 F/1.4. We'll see the results.

So far my results with the 50 F/1.1 + orange filter + overexposure by one stop have actually more light thant the shots from the 35 mm without finder nor compensation. Therefore, the +1 compensation is probably exagerated, but actually there are no burnt lights, everything seems properly exposed (with stand dev at least), and I find it slightly better than the 35 with no compensation. Perhaps I'll start applying a +0.5 stop with the 35, just as a trial.

 

[Sparrow]

TTL meters never give a correct reading because they're reflective ... if you take a photo of a black cat and then a white cat the meter will do it's best to make them both of them come out the same shade of grey ... the effect the filter contributes to the inaccuracy is probably quite small, variation in the overall brightness of the subject is probably much more significant