Tiny 1-inch Sensor in Low Light

misterodd wrote:

Night photography is hard on any camera. I wanted to test this out on an Nikon 1 camera. In my opinion, at least, on a screen, it didn't turn out too bad. Any thoughts?

It might take 2nd place to the noise in the "Original", but it is also obvious that you had no stability at all in this shot; you have a mostly-horizontal (8/2 o'clock) long, s-shaped motion blur. That disadvantages the captured signal vs the noise right away; motion blur does not smooth noise; only detail. Unless you took this from a moving perspective, camera-based stacking and merging could have given you more image-stabilization, as others have mentioned.

As far as the noise itself is concerned, one should keep in mind that when you shoot B&W on a color camera, the software may be including a very weak color channel, which contributes mostly read noise without contributing much signal. Using a workflow that compares the RAW color channels might suggest a workflow that omits the red and/or blue channel if it is really weak, or at least running a strong noise filter like a wide median filter in just that channel, if that channel is otherwise helpful.

The blue channel under sodium vapor lights, for example, may have an extremely low SNR compared to other channels, and worthy of omission, or extreme filtering of just that RAW color channel. Shooting under a red LED will render 3/4 of the pixels almost read-noise-only.

If cameras had only photon noise, adding in a little bit of exposure from a weak channel or weak exposure in a series would still increase SNR; in the presence of read noise, however, the read noise is always contributed even when there is no signal.