Topic: |
Long exposures and
sensor noise with Canon 5D2 |
Author/Copyright: |
John Jovic |
Shooting long exposures with
digital cameras can result in images with unusually high levels of
background noise. This noise is a side effect of the sensor itself
heating up. Some cameras are better suited to long exposures than
others. This is a brief comparison of images shot with a Canon 5D2 to
see what happens during very long exposures, between 5 and 60 minutes.
These tests also look at what happens when Live View is used and if this
affects sensor noise.
The test images where all shot with the cameras body cap fitted, ie no
lens, and the files are shown below for comparison, both as full frame
and 100% crops of each image. All tests where performed at room
temperature. The images where shot in RAW and where converted with Phase
One Capture One using the default noise setting in Capture One with no
exposure, contrast or any other adjustments at all. The images where
simply resized for web use.
The following images where shot with a cold camera, ie one which hadn't
been used at all for several hours before the exposures where made.
What happens when the sensor
gets hot? The following images where shot once the camera had been used
for over 1/2 hour so was effectively a normal temperature for the
camera/sensor. Although a 60 minute exposure does generate considerable
background noise it appears that 5 minute exposures are quite
comfortable for the 5D2, even after consistent or constant use, such as
immediately after the 60 minute exposure.
|
|
|
Not so good, but it could be much worse. |
|
|
|
|
|
This exposure was made immediately after the 60
minute exposure was made, ie with no delay, so the camera was hot
due to exposure, not Live View. |
|
The resulting 5 minute
exposure shows very little if any noise and is almost the same as
the 5 minute exposure with a Cold camera. It appears that the 5D2
does not generate a great deal of sensor heat during the exposure
itself. |
What happens when you use
Live View? The camera certainly draws quite a lot of current as the
batteries typically drain quite quickly, and the camera certainly heats
up. These tests where performed using Live View Stills only, Silent
shoot, Mode 2. The camera was left in Live View mode for a full 25
minutes before each of these test images was shot. The camera warms
considerably in Live View, more so than during the 60 minute exposure
but this might simply be due to the LCD screen itself rather than due to
the sensor being active.
It appears from the above tests that Live
View contributes to long exposure sensor noise quite significantly so it
may pay to minimise using it if noise performance needs to be optimised.
The Table below sets out the same images show above (100%
crops) but in a way
that allows them to be easily compared.