This is part of a long series of posts about the Sony a6300. The series starts here.
I don’t expect any surprises here, since I’ve dealt with a lot of Sony cameras, but you never know, so I looked at the histograms for all the dark field captures that I made for the plots in the previous a6300 posts. I’ll spare you the tedium of going through all of the ISO settings and both semi-mechanical shutter modes, and just post the most interesting ones.
ISO 100, single shot shutter mode:
This is typical Sony raw compressed performance. The camera advertises itself in the EXIF as a 14-bit device, but every other bucket is empty, so it’s only delivering 13 bits of precision.
In continuous mode at the same ISO:
Just like most modern Sony mirrorless cameras, the precision has dropped to 12 bits, and the noise level has risen.
At ISO 320, just before the change of conversion gain, in single shot mode:
At ISO 320, in continuous mode:
Everything looks normal.
At ISO 400, just after the increase in conversion gain, in single shot mode:
And in continuous mode at the same ISO:
As we saw in the graphs, turning up the ISO in this particular case actually lowers the read noise. Nicely played, Sony.
Looking at one of the middle-high ISOs to get a baseline, first in single shot mode:
And now in continuous mode:
There is an odd dropout in the continuous case, but otherwise it’s pretty much what you’d expect.
Taking a big step up in ISO to the point just before the digital low pass filtering starts to take place, first in single shot mode:
And now in continuous mode:
The odd dropout in continuous mode still appears, with a little energy in the blue channel — what’s that about? — but otherwise it’s normal looking.
Now to ISO 12800, where the filtering starts to rear its ugly head. First in single shot mode:
Note that the histogram is no longer Gaussian.
In continuous mode:
Lots of drop outs.
At very high ISO, in single shot mode:
Not even approximately Gaussian, with a very weird, but beautiful in its mathematical symmetry comb effect.
In continuous mode:
We’re basically down to 10 bits here.
It gets even worse as you go higher:
It looks to me like in continuous mode the camera is applying digital gain, causing the dropouts, which then get partially filled back in by the digital lowpass filtering.
Jack Hogan says
About the single shot histogram at ISO 10k: it appears that the raw data is being multiplied by a number that is not a multiple of two. Not hard to figure out which.
Max Berlin says
Thanks Jim.
Was considering the A6300 for video but now I’m out.
Jim says
Max, I don’t see how you can draw conclusions about video from the dark-field histograms for stills. The processing is completely different. There is no raw video in the a6300.
Jim
Max Berlin says
I get it. But a video camera that is worthless for stills means I have to carry another camera. I am trying to simplify and keep the compromises to a minimum.
Jim says
Why do you say it’s worthless for stills? For an APS-C camera, it’s doing a great job on still image quality, as far as I can tell.
Jim