This a7RII vs a7S low light video comparisons pits Sony’s two video-forward cameras against each other. Shot by the folks at Cinema5D, this video looks at the two cameras as they capture videos in very low light situations. Their companion article descries their testing methodology. \
For this test we used two identical Zeiss Loxia 50mm F/2 (Sony E-mount) lenses on the Sony A7RII and Sony A7s and filmed our test chart at the same time. For the purpose of the test we zoomed into lowlight critical areas of the frame. 400% on the Sony A7s’s HD footage and 200% on the A7RII’s 4K (UHD).
On first sight it might appear as though the noise floor is similar, but in our video above you can upon close inspection see that the A7s retains better detail throughout. At around ISO 8,000 the Sony A7RII shadow areas get more and more washed out and some detail is lost. Noise performance is still good and in many situations the footage beyond ISO 10,000 and even up to 25,600 might still be usable for you, but look out for those washed out dark areas.
The Sony A7s in comparison holds a very clean image all the way up to ISO 25,600. Noise gets severe beyond that point, but detail is retained quite well in comparison to other cameras, which makes it such an impressive lowlight tool.
If you want to be on the safe side you should be careful not to expose beyond ISO 6,400 on the new Sony A7RII, but if your final output is HD and web content you might find that even ISO 25,600 is possible without too much noise on the Sony A7RII. The fact of the matter is that you should make your own tests, find out and get a feeling for how far you can and want to push your footage to get the images look the way you like.
Obviously these notes are about video–you can shoot the A7RII way above ISO 6400 and have a great image. But it’s good to know what the ceiling is for video shooting.
Cinema5D has been doing a great job of doing reviews of the a7RII.
You can read their reviews of the Sony A7RII (see their field review here and preliminary lab test here). .