Hacker News new | ask | show | jobs
by hnov 1351 days ago
Phones outperform cameras in suboptimally lit shots. They also distort reality, hard. Good enough for IG but pretty different results otherwise.
2 comments

Depends on what camera you're referring to. Many modern CMOS cameras can basically see in the dark.
There’s an empty parking lot near my house that is pitch black at night. One morning early, I took a photo of it, on a lark - with a new Canon at high ISO and slow shutter.

It was unreal, like someone had switched on the sun for a second. It’s grainy from the high ISO, but you can see full details in full colour and with the naked eye it’s as dark as a pool of black ink.

I used to do a similar thing many years ago. Since my then DSLR was atrocious above ISO 1600, and barely usable at 800, I had to learn to be patient.

But the results of minutes-long exposures in the dark were absolutely crazy. Colors were very deep, and there was an incredible nuance in the tones of even the darkest corners.

Highly recommended to anyone not used with taking pictures in such conditions. This could even be tried in one's home at night. All the random LEDs shining around in the dark actually create an eerie atmosphere if you look long enough.

dslrs can use a similar technique that phones use—exposure bracketing—with good or better results depending on a variety of factors.

It does make you wonder if fullsized camera bodies will figure out better bracketing. Exposure time is really important, and you need good computation to align the images and correct for stability.

it's better to get it right during shooting than in post.

> dslrs can use a similar technique that phones use—exposure bracketing—with good or better results depending on a variety of factors.

This. My 2016 Olympyus Pen-F does this (a small m43 mirrorless body). It takes 4 captures at different exposures and combines them in-camera. I don't usually use this mode, but the results aren't bad at all for snapshots.

It can also take up to 7 different exposures, but the user needs to combine those manually. This is the mode I use most often, since it allows me to tweak the result how I want.

Right, my camera even goes as far as offering in-body HDR, it's just not as good as what I get in my iPhone, especially when the scene has moving elements, the camera isn't on a tripod, etc. Phones can paper over a lot of things with "AI" which means that you can get a serviceable shot without trying, but it doesn't really correspond to reality and sometimes (especially with the moving elements case) has artifacting.