long exposure dark frame thermal noise in Olympus cameras
- when taking photos with ANY digital camera for longer than 1 sec, an additional type of image noise called “thermal noise” needs to be managed
- this thermal noise is dependent upon:
- temperature of the sensor
- duration of exposure
- in terms of noise it is FAR better to increase the exposure time instead of raising the ISO setting by the same factor
- it is substantially removed AUTOMATICALLY when Noise Reduction = ON as the camera takes a 2nd exposure of the same duration but without the shutter open and this is subtracted from the original image to generate a much cleaner RAW file
- as thermal noise becomes problematic with very long exposures, when creating star trail images, it may be better to use Live Composite mode to take shorter exposures and automatically “stack” them in camera to create a very long star trail type of image but with far less thermal noise.
- automatic NR does NOT occur in BULB mode - you will need to manually take a dark frame exposure in RAW and in post-processing, subtract this from your RAW initial exposure
how much thermal noise is there?
- see http://www.wrotniak.net/photo/m43/em1-em5-dark.html for extensive testing at different ISO and exposure lengths for the E-M5 and E-M1 cameras which have quite different sensors
- the E-M5 has substantially better thermal noise profile than the E-M1 sensor
- strangely, the E-M1 sensor is very subject to Soylent Green effect at ISO 800 or higher with exposures above 60sec (or 15secs at higher ISO), and NR generates a magenta thermal noise pattern as presumably there is even more green thermal noise for the 2nd dark frame exposure.
- unlike earlier sensors such as the older Four Thirds dSLRs, there is no evidence of banding in either cameras in the dark frames
omd/dark_frame.txt · Last modified: 2014/10/20 12:36 by gary1