Why doesn't my camera capture a dark picture when it looks dark to the human eye?

Question

I tried searching for this, but found nothing on this topic:

is there a way to set the camera to get images in around the brightness a human eye sees?

For example, in bright daylight, I would set it to a short exposure (lets say 1/60th), with F16 and ISO200, and get an image which kind of looks like what I see with my eyes. Then inside a room, I open the aperture, increase the ISO, decrease the target EV, and get around the result I am having from seeing. And even further at night, with rod cells active, I would get very very low EV values.

Is there some way to tell my camera to simulate the same behaviour, or is this impossible due to the higher dynamic range a human eye can perceive? I find it kind of annoying to always countercheck the images if they have the same perceived illumination, and found no way to predict what the result will be.

Answer 1

It's not possible for the camera to reproduce what you think the photo should look like.

As the dynamic range of the camera is so much smaller than that of the eye, the camera has to guess what's important in that range, and that is generally done by exposing to get a good gray as a whole average.

The camera could underexpose the image when the light is low, but that would be even more of a guess, because the camera only knows the light level from an absolute point of view, it can't know how dark you think that it is. Most of the time you would just get surprised by how big the difference it was between how dark you perceived the scene, and how dark it really was.

Answer 2

You can make your camera do this, but you'll have to tell it what you want. That's because basic exposure metering are actually pretty dumb. Unlike a human, the camera doesn't understand the context of the scene. It just takes a reading, and sets exposure to make an image which will average out to medium bright.

Modern cameras also have something called "matrix metering" or "multi-zone metering". With these systems, the camera compares multiple readings from across the frame and compares that to a database — bright in the top of the frame, dark in the middle and lower half? Probably a person with the sky in the background. These systems could judge that the entire scene is pretty dark, and decide to render the exposure in a dark key as well, but for whatever reason, they usually aren't programmed to work like that.

One reason is that underexposing means less data is collected, and especially in low light that means more noise. You're better off exposing neutrally, in the middle, and changing the image to low-key in post-processing.

However, if you really do want the scene to be dark, you can simply tell the camera you mean it. This is a standard feature called EV compensation, and it does exactly what you ask for in your comment above: it tells the camera to adjust its exposure value compensation by a balue you choose. Meter as normal (for most consistent results, with center-weighted instead of matrix metering), and then dial in a stop or or stop and a half of exposure compensation. That should give an overall lower key, without totally underexposing the image.

Note that the exposure meter built into most cameras goes down to about EV 1; mid-range and up cameras will go down to EV 0. Below that, they just guess, or refuse to set anything. I'm not sure exactly why this is, but I suspect that it's just hard to be accurate in a short time. So if it's that dark, you probably will get an underexposed image without any compensation, or will need to resort to manual settings (where with trial and error you can make your own decision about the exposure of the overall scene).

Answer 3

What I am asking is if the camera can sense automatically that it's just plain dark...Assume that for normal light condition

Phrases like "just plain dark" and "normal light condition" don't mean much. When you go indoors after being out in bright sunshine for a while, everything looks "just plain dark" at first. A few minutes later, it feels like "normal light condition" because your eyes adjust to the new light level. A well-lit city street at night might seem very bright even though there's far less light available than there is at the same spot on a dreary, seemingly dark overcast day.

Your brain is good at noticing and adapting to changes in brightness, but not good at measuring absolute light levels. A camera, on the other hand, can only measure the absolute light level. The camera doesn't "know" how you perceive the current scene, and how you perceive the current scene may well change over time. So, if you want the photo you take to look dark, you need to set the camera to do that; if you want the photo to feel very bright, you can set it to do that instead.

Answer 4

if you set it on manual and set it for the light scene you will get exactly that effect. But then you have trouble with dynamic range. and your eyes also adapt its aperture. So you should dial up or down the aperture knob as you move from light to dark and vice versa, so you get the exposure that fits your perception (through trial and error).

However, since the SNR is pretty low when you underexpose your images will be inferior to making the dark scene bright on the sensor and then in post bring it down to the darkness you desire. This way to suppress the noise, too.

It cannot do this automatically, because it could keep an eye on the absolute exposure value based on iso, aperture, and shutter, but it have no way of guessing how bright you will view the image on your monitor, screen, print, etc., and it also does not want to produce bad noisy images on purpose. It could slam dunk the exposure in post automatically (for the jpeg) but then you lose the details you might want back to quantization.

In conclusion, it is better that you do it yourself for the intended medium.