How accurate are exposure meters? Not very....

[bglick]

>You'd need to develop three parallel spectro programs fed off a beamsplitter, then come way to calibrate each channel. Last time I talked to an engineering co capable of this they thought it would have to market in the $6000 range.


Drew, if you followed the links Rob offered above, you will see this is exactly what the Nikon color metering system does. Considering they sell hundreds of thousands of cameras, and have been making this color metering system for years, (starting for use with trannie films) the cost of the R&D and the product components itself has been beaten down to consumer price levels. This is the value of Ivans and Robs contributions.

So there is no reason to custom engineer a $10k solution for something Nikon has spent 15 years perfecting and available off the shelf. If you haven't read Robs links, I suggest you do. As for the spectral response of each trannie film, I think most all of the trannie films responses are close enough, that when combined with this metering system, it will remove any large exposure errors traditional meters create.


Finally a thread has a happy ending... glad I did not bail as my instincts kept telling me to.....as this thread took the normal course for the LF forum...... badger the OP, mis interpret the OP, mis quote the OP, sarcasm, etc. etc. This explains why many of the valuable contributors left this forum long ago.. Thanks Ivan for wading through the repetitive responses of this thread, to contribute a very useful suggestion.

[Ivan J. Eberle]

bglick,

Can't say about the D60 but thanks to some of the Nikon links referenced above, it's confirmed that all the Nikon DSLRs since the D2X are using RGB meters.

Exactly what the DSLR Matrix metering algorithms are doing with that information after the fact is perhaps different in digital than what the film camera ones do. The DSLR Matrix algorithms seem to expose to the right in backlit conditions, and you can often but not always recover the highlights in PP (if you shot it in RAW, that is).

But obviously that doesn't work with transparency film. All I need or want it to do is not blow the highlights, and let the other values fall where they may. The F5 algorithms seem to get this right. (Pentax Multi-spot or whatever they called it on the P645N seems slightly biased in the other direction for print film, exposing to the right much like a DSLR regime).

So what I'm most confident using on my DSLR as a metering tool is the spot meter (no algorithms involved. CW probably doesn't compare the scene to 20K others in the memory banks, either). I've heard some argue that Nikon spot metering is looking for a 12% reflectance value instead of 18%. No matter to me as I am getting excellent exposures with the spot metering as long as I can find something middle toned. (Snow might call for incidence, sometimes hard to find a middle gray value in these settings. Reflectance off snow calls for +1 to +2 stops of compensation and I find it varies a lot with the specularity of the ambient light). High and low key are where having the instant feedback of a histogram helps. The D200 is good for this use because the DR is similar to Velvia (not as wide as Astia). It's possible that this is less than 100% reliable as the histogram may be coming off the .jpg and not the RAW, after compression. IDK for sure but this has been a topic in earlier threads.

DSLR sensors, of course, have the advantage of better quantum efficiency and no reciprocity failure. Calibrating this all precisely to a specific emulsion is something I've still not done. What I can say is it's been working well enough out to the 4 second range with my 4x5 and "Mississippi one, Mississippi two..." and Fujichrome regime. If I were doing 2-shot stereo slides, I might also have to tighten up that timing system somehow (although after many years in the wet darkroom, I'd doubt I'm as much 1/3 of stop out of whack with it).

Film is still relatively cheap for 4x5. When in doubt, I bracket 3 exposures, 2nd and 3rd being over and under 1/2 stop. Occurs to me with such good grain-free results I've gotten pushing Astia 100F, that I might as well take the exact same exposure X 3, develop the first in my Jobo, and push/pull the first developer based on the results and apply it to the next two. This wouldn't work nearly as well for Velvia 50, but it's definitely an option with Astia. Too, if the first was spot-on, I'd have 3 identical good copies for distribution. (Might this work as well for your stereo use?)

[bglick]

> Exactly what the DSLR Matrix metering algorithms are doing with that information after the fact is perhaps different in digital than what the film camera ones do. The F5 algorithms seem to get this right.


I was thinkin the same thing after my last post...the F5 was all about trannie film, I think it might have a better chance of success for this application, as reading the light is only half the battle....what the camera does with that light reading is dependent on what it exposing to...and as you suggest, the algorithms in the F5 are probably best suited for chrome film vs. the DSLR's. Exposing to the right is critical in digital, not so true with film, and we don't know what the firmware instructions are. Also, at the same time, the Lux readings from the color meters may also concur with the Nikon readings....lots to work with...thanks for sharing your thoughts on this....


Most of my stereo on 35mm / 120 film, is single shot, both shutters fire at once... I will provide some pix if you want to see these cameras.... I often to bracket.... the problem is, many scenes have moving subjects, and you have one chance to capture the image.... sure when everything in the scene is static, the options are limitless.... including your multi exposure, process, evaluate results, and adjust accordingly for subsequent exposures... an EXCELLENT strategy. Up to this point, I have not done my own processing, so this was too difficult to communicate with labs, so I just bought a few Jobo 2500's and will begin processing in house soon. Then I can experiment with what you suggest as I have control of the entire process.

The LF stereo has been dual capture on a slide bar (limited subject matter), but I will correct that soon by building stereo LF cameras. Just not enough hours in the day right now...

[Robert Fisher]

Mr. MURAMATSU (Nikon engineer) said:

"Basically, there is no need for us to change the film camera metering method for use with digital cameras. But we must remember that the allowable exposure range is smaller for digital cameras than for film cameras. At an early stage, we decided to adjust the algorithm out of concern about this difference. So, we inherited the “1,005-pixel RGB sensor” and refined its algorithm to achieve the latest “3D-Color Matrix Metering II” metering system. This new system, launched with the D2X, is now also available with the D50 and D200."

As stated earlier, I used an F5 for years shooting wildlife in every conceivable environment and was always shocked at my "kill" ratio shooting E6 film.

Is ANYBODY using an F5/F6 as a meter for LF E6 shooting?????

[Ivan J. Eberle]

I might be using either of my F5s more, if they didn't weigh as much as my medium format Pentax 645N, and if I didn't also have two DSLRs.

Now, I keep the F5s around for use as IR-triggered wildlife trail cams that are relatively expendable (in the instance that a dope grower stumbles upon one of my light traps).

[Drew Wiley]

Like I said, the meter itself is seldom the issue. Today I was so distracted by the
constantly changing light out in the redwoods that I got the filmholder in backwards and managed to double expose an 8x10 E-6. Can't blame the meter for that! And
I don't have any interest in carrying around a digital camera just to check readings, although it is perfectly apparent that that is where all the R&D money has landed.
Too bad they can't transfer some of that effort to a meter per se. But the Nikon technology you refer to isn't what I had in mind at all, though it is valid in its own
right. Makes no difference - nobody is going to develop a truly superior meter itself
because there's no money in it. Once you stick one of these inside a camera you
have all kinds of ulterior complications. I find a handheld meter superior even for
35mm photography. But to each his own.

[Drew Wiley]

I should follow up a little more just so you don't think I'm picking on someone. A
simple spotmeter fulfills a very important function. It allow YOU to place your values
right where you want them in relation to scene illumination. Every movie cameraman and a lot of Zone System photographers knows what this means. What an advanced Nikon camera contains is basically a tricolor meter with a lot of fancy probability software. It doesn't replicate the role of a spectrophotometer whatsoever. It tells you what the exposure should be based upon a lot of interpolation and stored
exposure information. What I was referring to is something which reads and mathematically calculates the linearity or curve of each sensitized layer, and then informs you of the actual placement of the values. I attempted to design something
like this for color darkroom printing and partially succeeded, but I have to plot the
values on graph paper the old-fashioned way. Of course, there's plenty of time to do
this, so the analogy to field photography is a ways off. But in the process I did
consult with some pretty high-tech electrical and optical engineers who hypothetically have the ability to design something more efficient. That's why I have an idea of what's going on and why this would be cost prohibitive for a conventional light meter. Nikon and Canon have taken a different path entirely.

[Robert Fisher]

Drew, thanks for all your input (Bill too). Effective next weekend, I will be using two meters - my Sekonic and a F5 that I purchased today.

810 E6 film is just too darn expensive to bracket plus the hassle of extra holders, space, etc. etc.

[Don Dudenbostel]

So that's half the equation. Does the effective speed of your film change at different color temperatures?

Unless your film is perfectly linear in color response then it will have slightly differing speeds at different color temps. Years ago B&W films were rated different under daylight vs tungsten due to low red sensitivity.

[bglick]

>What an advanced Nikon camera contains is basically a tricolor meter with a lot of fancy probability software. It doesn't replicate the role of a spectrophotometer whatsoever.


Drew, i applaud all your experiments and R&D... but I think you are over complicating this entire issue. Many people have reported just how finely well tuned the Nikon F5/6 performed with chrome film. That's the goal here, Nikon perfected it. If you put an equal effective fl lens on an F5/6, to achieve the same composure you have on your 810 gg, and you have full confidence in the F5/6 success with exposure with a given film, you just found a low cost proven solution to the issue being discussed here. In addition, the Nikon can be used in center metering mode, almost turning the meter into a spot meter. This is a brilliant approach IMO....


Robert I applaud your tenacity.... you are a Gettrrr done kind of guy. I can understand why you would not post your findings on this forum, but be sure to email me off list with what your findings, I am very interested. I am consumed by my day job till the end of summer so don't have a lot of time to experiment right now....