Originally posted by Wayneheim Just got my k3 tonight and was putting it threw some tests. I was trying some high iso tests and playing with the interval shooting. I found that if you shoot a single shot at 12800 and then compare it to a 12800 interval of 8 shots set to average, the noise reduction is incredible. Way better then using in camera noise reduction. Try it at 52,000 and results were even more amazing. Could pull a much cleaner image this way then trying to use in camera noise reduction or Lightroom noise reduction. The image does end up a hair softer though but that's at 200 zoom. Tried the test at 2000 iso and it works there too but Lightroom can easily handle that noise almost as well. But interval and Lightroom is a great combination.
Anyone else confirm tried this? I tried a 19 shot average but at some point the averaging doesn't improve it any more. I think it's somewhere between 6-9 shots seems to provide the best results. Attached is a 12,800 with some Lightroom sharpening and a bit of noise reduction.
Wayne
I have not confirmed it but it seems consistent with the theory that noise is random, so averaging of several must reduce noise. The number of frames which makes sense would confront the problem of change of the image from frame to frame - either from camera movement or subject change. Analogs of this noise reduction method have been used in electronics for many decades. Interesting that the first place I encountered this was in a low noise amplifier for moving coil cartridges for playing vinyl records (remember them) - there were 8 transistors in parallel to cancel out the effect of noise in each. I never thought if there was a theoretical limit to the number which could be used or if it was just a practical 'diminishing returns' based limit to adding more - given that each additional transistor would add cost and other issues.
Also, to avoid camera movement over an extended time one would need a tripod, so why not use a lower ISO and the noise problem goes away.
This method cannot address the problem of low light AND fast changing subject.