Saturday, January 10, 2015

Noise comparison of A6000 versus 5N / Rauschvergleich für die A6000 und Nex 5N

since I stumbled upon statements claiming that the A6000 to be inferior in regards to noise to the old 5N, and I own both, I made a quick test comparing both.

Common Settings

Tripod, 2 sec delay, steadyshot, center field AF, lens SEL50F18, f8, WB set manually to the white paper.
I then set the A6000 with exposure to the right and checked the histogramm for no clipping. This was (for the test scene) at Exp. compensation of + 0.3 EV.
Now everyone knows that the old Nexes tend to heavily underexpose, but I had to set the 5N to exp.comp. of +1.7 EV for similar histogram and shutter speed!
Similar shutter speed indicates that both sensors are of similar sensitivity, with a very minor tendency of the 5N not need longer shutter speeds at higher ISOs.
Different exposure compensation means that the program logic of the A6000 is much better by default.
NR in camera was set to Low.

Test scene

Test Scene, note that phone and bottle are in the front and therefore not as sharp as the mid section where the AF was directed at.

This is NOT the real file with the full resolution, use the download link below
The comparison consists of four crops at 100% from left bottom, right bottom, mid bottom and mid section.

Download from Google drive and view locally on your PC for correct rendering

The pictures from the A6000 were resized to 16 Mpx.

In the mid horizontal lines from left to right are the crops from the OutOfCamera-Jpgs, A6000 on top, 5N below. The crops run from left to right from ISO 100, 800, 1600, 3200 and 6400.

For ISO 100, I also have added the JPG files obtained from Adobe Camera Raw, applying USM with 100/1.2/3. For ISO3200, the same, with NR of 30/20/0/50/50.

Lastly, for ISO 100 and 3200, I also added on top the crops from the not resized 24 Mpx JPG OOC file. Please note that for ISO3200, there is one typo listing it as "ISO100 1/5 sec" instead of the correct "ISO3200 1/5 sec".


If both cameras are properly exposed à ETTR, the noise level of both cameras is very similar from ISO100 to 6400. The noise of the A6000 is more fine-grained, the 5N is more blotchy. The 5N applies stronger sharpening, which can be seen at ISO100 through the slight halos indicating oversharpening. Using Raw, both are very similar again. Remember that this was with incamera NR setting to Low. I do not recommend to use the NR-high setting in the A6000.

However, as the RAW-JPGs prove, the cameras themselves are almost identical in noise and detail, with the edge going to the downsized files from the A6000. Also, this proves that there is NO incamera RAW NR as some have claimed. In reality however, the 5N might even be more prone to higher noise in case the user does not use ETTR to compensate for the 5N's strong tendency to underexpose, and consequently has to compensate in post-processing.

No comments:

Post a Comment