a6k Loc: Detroit & Sanibel

If you are not a perfectionist or close to it then the stuff below will not be worth your time.😇

I've been trying to do well-controlled experiments to understand in precise terms what my camera's exposure and control system(s) do. I don't have enough results yet for an unassailable answer but I've stumbled on something that some of you may want to consider or even try out yourselves.

My camera for these experiments is my Sony a6500 which has interchangeable lenses. I have a Metabones "dumb" adapter for Nikon lenses and it's mostly just a spacer. I use it without a lens for my tests and I use my monitor for well-controlled light sources as an alternative to a reflective surface. My idea here is that when light hits the sensor, the sensor does not know or care if the light came from an emitting source or from a reflecting source. Omitting the lens is just a way to reduce "fall off" in the image but I only evaluate the center anyhow. Just to be clear, the light source is about 5100 to 5300 Kelvin.

My camera allows me to set 1/3 stop increments (or 1/2). I chose to use 1/3. Obviously, I can only vary the ISO and/or the shutter since I am not using a lens. If I did use a lens then its ability to move exposure between stops would be design-specific. Like perhaps all digital cameras, mine reports exposure variables in the EXIF. Both the settings that I see on the camera's screen and those that are reported in the EXIF are always aligned to the "standard" stop and fractional stop intervals. In other words, my shutter can do 1/320, 1/400, 1/500 and so on. But it can't do 1/345 for example. My iPhone does do that but my Sony does not. I don't know what Canon, Nikon, etc. do.

What I discovered is that when the shutter is being controlled by the camera (Aperture priority or full-Auto for example) then it is being controlled to a much greater level of precision. At least, that is what I think is happening. What I know for sure is that for many exposures of a known light level, the automatic behavior gets me a different level of exposure in the image than if I enter the reported settings manually and disallow any automation. The reverse is also true (see below). I think this is just a form of rounding error. If so, then the error can be almost 1/6 of a stop. That's not huge, but it's easily observed and measured by any or all of my post-processor applications and I can see it, too.

Which is more "correct"? I have to conclude that the automated or semi-automated version is the more correct one. BTW, the error is not always in the same direction which adds to my comfort level in saying that it's a rounding error.

This is not about how good my meter is or what it assumes about the average reflectance of the world. In fact, I've come to the conclusion that those questions are very unimportant. I want to know what my exposure system will do in terms of the eventual levels in the raw file relative to the ends of the scale. I look at it with both the RGB 8-bit levels 0~255 and with CIELAB's L* values which are up to 100 (%).

For those who may want to experiment on this themselves I will point out that a solid white target that fills your monitor's screen will turn out some form of middle gray in the eventual image if you let the meter either recommend or actually set the exposure. My monitor has brightness levels from 0 or 1 up to 100 and I've measured them to discover that they are linear. In other words, there is one stop between 50 and 100. I suspect many monitors can do something similar. In addition, Sony EXIFs report "Brightness Value" which is actually in the normal log form. In other words, a BV of 4 is one stop lower than a BV of 5. These BV's are part of the APEX exposure system which can be looked up. They correspond to a formula in which an EV of 15 is normal daylight, corresponding to the Sunny 16 rule. This is all very helpful but may be TMI. YMMV. Because each image reports not only exposure but BV, it is easy to see that there is a range of BV's that will have the same reported exposure (shutter, ISO, aperture) but will produce unequal values in the gray result.

Taking it the other way around, I shot a series of exposures on "full Auto" (but the camera could not adjust the missing lens). As it happens, it stuck with ISO 100 which is what I wanted. I varied the brightness level of the monitor and took about 7 stepped exposures. The resulting images all had the same CIELAB brightness to the second decimal value (same integer for all). This tends to prove that the exposure system on the camera, at least in the middle of the envelope, is accurately repeatable and predictable.

NCMtnMan Loc: N. Fork New River, Ashe Co., NC

I'd rather just go take pictures.

donrent Loc: Punta Gorda , Fl

Do you always go thru life looking how to do things more complicated ?

dsmeltz Loc: Philadelphia

The most telling part of your post is "I've been trying to do well-controlled experiments." If you are a studio photographer (you always shoot under controlled conditions) this approach makes perfect sense. However, I shoot in the real world under conditions that are never controlled. The most important settings in that environment are planning and preparation.

a6k Loc: Detroit & Sanibel

I suppose I should have expected a few ad hominem remarks. I decided to post this for the few readers who shoot manual and eschew camera automation. If what I am reporting is of no interest to you, why not just move on instead of cluttering up the thread with personal criticism?

Adams tried to be precise in how he metered and placed his subject tones in the scale (AKA Zone System). It seems to me that the deeper knowledge of the way that the tool works and the way that the digital version of film responds is still worth some study. The people who design, engineer and build our cameras have to know this stuff so why criticize a layman for being curious and wanting to improve his results?

An experiment needs to be well controlled in order to have any validity but this kind of understanding is certainly not restricted to studio shots. In fact, landscapes are the most likely category to benefit. Those who shoot ETTR need to understand the distance between blown highlights and what the meter is doing. My experiments, reported elsewhere, with Zebra settings were aimed at that. This set of experiments is helping me pin down the point on the light-dark scale that the meter is set for and thus to pin down the remaining room above.

For scenes which don't have huge dynamic range (includes many bird shots which are my main interest) this can help me to know how much to increase exposure in order to better capture both the shadows which are often present and to reduce noise - all without blowing highlights.

I am not suggesting that this little point (automation vs full manual) is all that there is to taking pictures and I wonder why anyone would presume to lecture me about such an obvious point.

larryepage Loc: North Texas area

I believe that this is a very interesting experiment that you have designed. It is a very good digital adaptation of the process we learned in my first photography class to fully know and understand how to move from our meter readings to a properly dense negative. Your process is actually the first one I have seen that effectively adapts our old film "calibration" process to digital equipment.

What you describe is able to verify whether your camera's metering system and shutter work together properly to deliver the proper exposure. And your experience comparing various automatic modes to manual settings verifies what is very common in digital cameras...the automated system avails itself of shutter speeds and aperture settings that fall between the one-half or one-third stops available manually. I do not use Auto ISO, but suspect the same thing would happen with that function as well.

The next step, then, is that once camera performance is verified, each lens should be verified to understand three things... 1) Is the diaphragm set properly by the camera, whether automatically or manually, 2) Is the lens working properly (no sticky diaphragm, etc., and 3) Does the lens properly handle adjustments to exposure at various focus distances, especially in very close focus when the lens is longer and an improperly compensated lens can consistently underexpose images. This can certainly be done with your

Also...now that I have been convinced (mostly by this forum and by a trusted workshop instructor) to plan on post processing all retained images, I no longer worry about brightness falloff, since it can be fixed with one click in Lightroom for essentially any lens imaginable.

So what you are doing is valid and is historically part of the process for disciplined photographers. I'm guessing that identified problems will be rare, perhaps extremely rare, but the equipment we use is very complex, and there is not necessarily any reason to presume that it is all absolutely correct. I remember finding that my Olympus OM-2n consistently underexposed by a full stop. That certainly had helped me avoid blown highlights, but it was a pretty big deal when trying to properly apply the zone system on 35mm film.

Thanks for this insightful investigation and report. Please keep us posted.

a6k Loc: Detroit & Sanibel

Thank you for making my day! BTW, I agree that a full understanding of the camera's predictable behavior should include the attached (or fixed-on) lens and in the case of zoom lenses, at various zoom settings.

Right now I'm struggling with a related issue. When I examine the raw files in these tests (I'm up to 9 series so far) with various apps on my Mac I get different results and sometimes those results are greater than I'm comfortable with. I also get (no surprise) a significant difference between a "linear response" setting and the various available curves in a given app.

I've been using - in no order - CaptureOne, Luminar, Preview, RawPhotoProcessor 64, Darktable, RawTherapee, FastRawViewer and the Mac's Digital Color Meter. And when I want to see what the EXIF contains I use either Preview or "exiftool" which requires "Terminal" AKA command-line. Some of the above, notably Luminar, are mostly useless for the purpose with the exception that it and FRV can display the image and then I can use the Digital Color Meter app on that.

Not having a credible starting point is a major barrier to understanding the output. We Never had that kind of problem with B&W film. I built a densitometer back in those days and it was exactly what was needed for matching negatives to grades of contrast in printing papers. I may break down and invest in RawDigger because I can't see how, without it, I can make all this sensible.

Any contributions from those with kind intentions and a lot of knowledge would be welcome.

larryepage Loc: North Texas area

This additional information is very interesting. It sounds like there might be a need to 'calibrate' our post processing software also. Come to think of it, we did do some work once we got our negatives correct to understand our enlargers. Mostly lamp brightness variations from one to another, as I recall. And of course lenses could get switched around as well...

Rathyatra Loc: Southport, United Kingdom

pila

The road to perfection is a slippery slope

traderjohn Loc: New York City

Why else would you have a camera?? I think you hit the nail on the head. Anal seems to be a new way to pass the time

Architect1776 Loc: In my mind

Shooting a blank wall or what ever is an interesting exercise. And you can see what a camera is doing on a blank monochromatic wall.

In real life though most of us shoot subjects with a fairly wide range of colors and shade to highlight in them.
We make a best guess at where we want to set the exposure for what part of the subject and go from there. 1/6 of a stop is nothing in real world shooting, even studio, and a very slight tweak in PS or your favorite program can make final adjustments.
So now you know this information how will you apply it to real world pictures that you take?
As you know there are sports/action shots that require very fast response or landscapes with a wide range of light values. How does what you learned apply to the various situations at 1/6 stop?
Now take this data and go do real photos in uncontrolled situations or with subjects in the studio with a wide range of colors and brightness reflecting back to the camera.

rmalarz Loc: Tempe, Arizona

I've done, and do, very similar to your experiments. I rarely take a camera out to shoot without doing a series of tests. These involve careful metering of a Macbeth color chart and shooting with each of the lenses I can use with a particular camera. I also use each of the metering modes, as well, Then, I run the same tests using a handheld light meter. At that point, the results will provide me with the capabilities of the combinations of cameras and lenses.

I'm not sure of your choice to choose to bypass the lens. Although each of our lenses have f-stops, the amount of light transmitted through the lens is equally important to know. And, yes, I'm a believer in The Zone System and use it regardless of shooting film or digital.

I'm often puzzled by those who post a series of random photographs stating they're testing a new <fill in the blank>. What exactly were they testing? I guess some are just too impatient to use their new equipment rather than to know their equipment. I prefer to know and understand what my equipment will do under a variety of situations.
--Bob

larryepage Loc: North Texas area

Digital cameras and lenses today are made so well and are so reliable that most of the time it is possible to take them out of the box, attach a lens, and capture a really good image. All you really have to do, especially if you are saving your images in raw form, is avoid blowing out your highlights. Almost everything else, except serious underexposures, bad focus, and sometimes wrong composition, can be recovered to at least some degree in post processing. But if you make a claim to be doing careful spot metering and then shooting in manual for perfect exposure, and particularly if you claim to be using the zone system, you have to know where zone 5 is on your camera. In particular, you have to know where zone 10 (or 11, if you use that version) falls with a given exposure.

So is this experimentation required? No...as the OP said. But to maintain really precise exposure control, knowing what the camera actually does is critical. And yes, that does include the lens as well. But it is necessary to understand the camera first, then understand error contributions from each lens.

We all know about the exposure triangle which this process (or a similar one) helps address. But there is also another triangle...the photography triangle, with Vision, Technique, and Equipment at the corners. This is certainly part of both the Equipment and Technique pieces.

olemikey Loc: 6 mile creek, Spacecoast Florida

I find it interesting - but then I'm retired and have time for technical things. I do some of the same, but not as disciplined in approach. Do what you like!