Posted on Leave a comment

Metering Project Frequently Asked Questions

(This article is a supporting part of our ongoing testing of low-light camera metering reliability)


Jump to…

Please explain your categories.
What lens did you use for this?
If I adjust exposure compensation, my camera can meter in lower light levels, right?
Does vignetting affect the test results?
If I use a faster lens, I can meter in lower light, right?
If I use live view on my DSLR, I can meter in lower light, right?
Does metering mode matter?
Since you take photos with one-stop changes in light levels, how can you report metering failure with 1/10 stop precision?
Can I view your code? I want to understand your pipeline.

 

Please explain your categories.

Metering limit definition 1:
Metering failure occurs at the point where the camera compensates 2/3 stop for a 1 stop change in ambient brightness.

Metering limit definition 2:
Metering failure occurs when the image is 1 stop darker than the image taken at EV0. If the camera is already struggling at EV0, we report the light level that produces an image one stop darker than a photo of a well-lit white wall and consider the answer approximate.

Manufacturer claimed value:
What the spec sheet of the camera says about its metering capability. We have not attempted to translate this into ISO 100 f/2.8 units.

Moon phase equivalent to definition 1:
The illumination levels of definition 1 were converted to moon phase using the plot shown here. This should be viewed at approximate only, since the preferred exposure for moonlight conditions depends on the scene and user’s preference. Sean has a calculator for moonlit light levels on his website at http://sgphotos.com/blog/exptime.html, but this reports “expose to the right” exposure settings that aim to fill the camera’s histogram. The settings in that calculator are two stops offset from those of the plot.

Camera stops down lens to meter?
Some mirrorless cameras (Sonies) keep the lens stopped down to the user-specified settings even when not taking a picture. In other words, if you are using an f/1.4 lens and have the camera set to f/2.8, the lens sits at f/2.8 all the time. This reduces the ability of the camera to meter in low-light conditions and therefore is bad. “No” is the preferred answer to this category. There are two reasons for cameras to preemptively stop down the lens. First, it gives the user a real-time depth of field preview. Second, some lenses exhibit focus shifts when stopping down, so if the user focuses wide open and then stops down to take the picture, the picture will be out of focus. Spherical aberration causes this. The first reason can be mitigated by having a “depth of field preview” button, and the second issue can be solved with a lookup table of aperture-dependent focus shifts. It is lazy engineering to keep the lens stopped down and not give the user the option to meter with the lens wide open.

If DSLR: better in live view?
Can the camera meter in lower light when in live view? In most cases, the answer is yes. DSLRs normally split the light between the viewfinder and the AF/metering sensors. When the user switches to live view, all the light goes to the detector, and so it is all available for metering calculations.

If DSLR: benefits from <f/2.8 lenses?
I don’t entirely understand the mechanism of this, but some DSLRs do not improve their metering limits when moving to a lens that is faster than f/2.8. Metering improves from f/5.6 to f/4, and f/4 to f/2.8, but the metering at f/2.8 is the same as at f/2.0 or f/1.4. I suspect this is related to the fact that there are f/2.8 AF sensors and <f/2.8 sensors. The preferred answer in this category is “yes.”

Exposure compensation to put the histogram in the middle:
At EV0, what exposure compensation places the histogram spike in the middle of the plot? Different cameras have different ideas for “ideal” exposure, and this may be different from one camera to another even of the same model.

 

What lens did you use for this?

We used a Rokinon 50mm f/1.4 set to f/2.8. It is EF mount, which can be adapted to Sony E, Canon EF-M, Canon RF, Nikon Z, etc with a tube adapter (i.e. no loss of light). It has a mechanical aperture, so I can control the light level entering the camera. It cannot be adapted to Nikon DSLRs, unfortunately, so we used a Nikon 105mm f/2.8 Micro (also at f/2.8) for the D800E.

 

If I adjust exposure compensation, my camera can meter in lower light levels, right?

No. Look at any one of the plots, and consider what the X & Y axes actually mean. Adjusting exposure compensation would shift the curve up or down, not left or right. It would not affect at what EV the kink or drop occurs, and it would not affect our measurements for metering failure.

 

Does vignetting affect the test results?

Sort of. The lens transmission in the metered area affects the results, which is why we use a Rokinon 50mm f/1.4 at f/2.8 for all bodies for the standard test. This lens was chosen because it has a mechanical aperture that can be set by hand and stays stopped down, and also it can be adapted to many different mounts. Using one lens gives an even playing field to cameras from multiple brands. Returning to the original question, the metering of most cameras is weighted toward the center of the frame, so dark corners don’t have a big impact.

 

If I use a faster lens, I can meter in lower light, right?

Maybe. The metering of some DSLRs doesn’t improve when lenses become brighter/faster than f/2.8.

On the other hand, some mirrorless cameras (all Sonies, no Canon EF-Ms) stop down the lens in order to meter and autofocus. This is an extremely clumsy solution to the possibility that focus can shift with aperture, and is worthy of an extended blog post/rant. If you have a f/1.4 lens on a Sony but have it set to f/8, your metering is five stops worse than if the lens was set to f/1.4. So you only achieve better metering with a fast lens if you have it set to a fast aperture.

 

If I use live view on my DSLR, I can meter in lower light, right?

Probably. But your battery will die quickly.

 

Does metering mode matter?

Possibly. Most camera specs do not state the metering mode for the specified metering range. However, for example, the Nikon D800E claims EV 0 to EV 20 in matrix or center-weighted metering, but EV 2 to EV 20 for spot metering.

In case you’re wondering, our tests were all done with the camera in “evaluative” or “matrix” metering mode, and that is also what we use in real-world shooting conditions.

 

Since you take photos with one-stop changes in light levels, how can you report metering failure with 1/10 stop precision?

We interpolated between data points using a second-order spline fit. In most cases, this should be accurate to a tenth of a stop.

 

Can I view your code? I want to understand your pipeline.

Sure! The Arduino and Python codes for the project can be viewed at https://github.com/geekyrocketguy/laptop/tree/master/metering%20project. If you have further questions, Sean can be reached at sean@sgphotos.com.

 


Integrating Sphere & Camera Metering Test Project

Main Project Page – Test Results

Project Overview – What Is An Integrating Sphere, and How We Used One to Measure Cameras’ Low-Light Metering Capability

Frequently Asked Questions / FAQ
(YOU ARE HERE)

What are EVs, and What do They Mean for Different Cameras? (Non-Technical Explanation)

The Technical Explanation of EVs, and Calibration of the Integrating Sphere

So, How Did You Build an Integrating Sphere, Anyway?

Timelapse Methods Compared: Aperture Priority VS Holy Grail Method