NR in camera or PP

[vincent_su]

Hi,
If this question has been discussed before, please feel free to point me to the thread.
I mostly shoot in raw and am wondering if the noise reduction better be done in camera or with the PSE 6.0.
Look forward to your comments.

[canonloader]

I'm not sure you can do NR in RAW mode. At any rate, I shoot only RAW and do NR in the newest download of Photomatix, which now has it available as an option in the steps of generating an HDR file, and the second to last step once I get the final jpg into CS3. Last step is sharpening.

[vincent_su]

Thanks for the tips. I'll make sure that I do this early.

[Bill Boehme]

I'm not sure you can do NR in RAW mode. At any rate, I shoot only RAW and do NR in the newest download of Photomatix, which now has it available as an option in the steps of generating an HDR file, and the second to last step once I get the final jpg into CS3. Last step is sharpening.

It doesn't make any difference whether it is RAW or JPG -- the custom function noise reduction can be set for exposures one second or longer. Alternatively, you can do it yourself by using dark frame subtraction in Photoshop. Contrary to popular folklore, the dark frame does not even need to be taken immediately after the image is made. It can be made at any time -- the only requirement is that the ISO and shutter exposure time need to be exactly the same.

Here is a thread (http://photography-on-the.net/forum/showthread.php?t=598343&highlight=dark) where I explain using dark frame subtraction using Photoshop.

[René Damkot]

Are we talking long exposure NR (dark frame subtraction) or (high) ISO noise reduction?

First I do in camera, since that's easiest IMO. If I had to do a lot of loooong shots behind each other, I'd probably turn it off.
Second: Never in camera, rarely in post.

[Bill Boehme]

Are we talking long exposure NR (dark frame subtraction) or (high) ISO noise reduction?

First I do in camera, since that's easiest IMO. If I had to do a lot of loooong shots behind each other, I'd probably turn it off.
Second: Never in camera, rarely in post.

That is a very good question, René.

In both cases, the root cause of the noise has the same origin which is photosite-to-photosite variation in sensitivity to light which results in a fixed characteristic known as pattern noise. With high ISO, the analog gain of the photosites are increased which means that low-level signals in the nonlinear region of their response curve get boosted up above the LSB cutoff point. Long exposures are typically made using low ISO (at least that would be the case for really long exposures in astrophotography). But, the very long integration time is similar to increasing ISO gain without digging down into the nonlinear response region. In both cases, I think that the pattern noise is repeatable, but somewhat better noise reduction results can be achieved for low ISO long exposures. The problem with high ISO is that nonrepeatable noise starts to become noticeable. Things like quantization noise during A/D conversion, Johnson noise, and thermal noise all begin to muddy things up (literally).

[canonloader]

I don't think this is going to be needed much for HDR images. Rarely will you shoot more than a second, and the 40D and newer cameras can handle shutter speeds up to 30 seconds without that much noise. At least mine does.

[ArcticShooter]

That is a very good question, René.

In both cases, the root cause of the noise has the same origin which is photosite-to-photosite variation in sensitivity to light which results in a fixed characteristic known as pattern noise. With high ISO, the analog gain of the photosites are increased which means that low-level signals in the nonlinear region of their response curve get boosted up above the LSB cutoff point. Long exposures are typically made using low ISO (at least that would be the case for really long exposures in astrophotography). But, the very long integration time is similar to increasing ISO gain without digging down into the nonlinear response region. In both cases, I think that the pattern noise is repeatable, but somewhat better noise reduction results can be achieved for low ISO long exposures. The problem with high ISO is that nonrepeatable noise starts to become noticeable. Things like quantization noise during A/D conversion, Johnson noise, and thermal noise all begin to muddy things up (literally).

Too long text to read. Lost it in the middle ;)
Can you make an tutorial for this?
Sounds interesting