Leica Night scene - noise and detail - Leica C-112 and D-Lux6.

[dalethorn]

Here is a comparison of noise and detail with several different ISOs. It won't be a completely fair comparison since the second C-112 image is closer than the first C-112 image and also closer than the third image (from the D-Lux6). The first image is ISO 1600 (C-112, 1/4 second), the second is ISO 80 (C-112, 6 seconds), and the third is ISO 1250 (D-Lux6, 1/4 second).

 

[Brian]

The level of detail in the lit window in shot #2 is much higher than the other two. That was a steady 6-second exposure. The lost detail in the other shots- Makes me wonder if in-camera noise reduction is killing off detail. "Like night and day"

 

[dalethorn]

The level of detail in the lit window in shot #2 is much higher than the other two. That was a steady 6-second exposure. The lost detail in the other shots- Makes me wonder if in-camera noise reduction is killing off detail. "Like night and day"

It seemed to me when I examined shot #1 (and several others taken at the same time) that the noise was just eating the image. But I suspect there are other factors. I think if I had used 1/2 second at ISO 1600 on the first, by switching to shutter priority as I did with image #2 (necessary of course to get the exposure greater than one second), then image #1's noise would look more like image #3 from the D-Lux6. So the camera probably was not saturating enough on a purely auto-exposure. Now I need to go back and retest the ISO 1600. If my suspicion proves out, I might ask Leica to include a special warning about auto-exposure under-exposure in high ISO situations.

 

[dalethorn]

The second image was updated automatically when I updated it on my website. It's now the C-112 at 1600 ISO, with exposure time of 1.3 seconds at f8.0 - I expected the camera would stick with f5.x at the somewhat long exposure using shutter priority, but instead the camera chose f8 - most peculiar. Anyway, it's a good compare now showing how much better the 'C' is with proper exposure, but just as important, that you can't depend on a straightforward calculation such as 20 times the ISO, therefore 1/20 the shutter speed.