Noise Ninja custom noise print- worth the effort for stacked photo??

I am an amateur astronomer and use a digicam to take pictures of the solar
system. I take up to 100 photos and then stack the photos into one to
reduce noise levels. However, sometimes a 100 photo composite isn't enough
for the amount of noise I get. I have Noise Ninja, but don't use it much
because I am never happy with the results as it seems to smooth the images
too much, washing out details. However, I have never created a custom noise
print for my camera, using the checkerboard pattern and a defocused shot.
My big question is: would a custom noise print created this way be worth the
effort and would the results be better than just the box sampling I do now
from the existing image? Also, I'm guessing that I would need to take 100
such samples since 100 images make up a composite, but not sure how to go
about it. I would think that each image would have to have noise reduction
before stacking for this to work and that's how I was going to do it. So,
yes, a lot of effort involved, but would the result be worth it?

Thanks,
Jason Sommers

"Jason Sommers" wrote in message
nk.net...
I am an amateur astronomer and use a digicam to take
pictures of the solar system. I take up to 100 photos
and then stack the photos into one to reduce noise levels.
However, sometimes a 100 photo composite isn't enough
for the amount of noise I get.

If reducing the random noise by a factor of 10 isn't enough, you
should indeed investigate post-processing noise reduction per image
(assuming you cannot reduce noise by preventing it). Do you do
darkframe/offset/bias/etc. subtraction?

I have Noise Ninja, but don't use it much because I am
never happy with the results as it seems to smooth the
images too much, washing out details.

I'm more familiar with Neat Image, but these programs are probably
equally suitable. Spend some time in tweaking the settings (often
means reducing the default amount of reduction).

However, I have never created a custom noise print for
my camera, using the checkerboard pattern and a
defocused shot.

You should, as there is no suitable featureless area in deepspace
images you could use to create a noise profile. Do make sure that you
don't make this profile until the darkframe subtraction etc. has been
done first.

My big question is: would a custom noise print created
this way be worth the effort and would the results be better
than just the box sampling I do now from the existing
image?

Although it's hard to judge without examples I am almost certain it
will improve things a lot. A program like Neat Image can run
unattended in batch mode once you've figured out the optimal settings.

Also, I'm guessing that I would need to take 100 such
samples since 100 images make up a composite, but
not sure how to go about it. I would think that each
image would have to have noise reduction before
stacking for this to work and that's how I was going to
do it.

Yes, individual noise reduction per image (in batch mode), not on the
stacked result.

So, yes, a lot of effort involved, but would the result be worth it?

Again, make sure you've done all you can to prevent the noise in the
first place, because all post-processing carries the risk of signal
loss. There would be not much benefit if the noise reduction doesn't
improve the S/N ratio.

Bart

"Jason Sommers" wrote in message
nk.net...
I am an amateur astronomer and use a digicam to take
pictures of the solar system. I take up to 100 photos
and then stack the photos into one to reduce noise levels.
However, sometimes a 100 photo composite isn't enough
for the amount of noise I get.

If reducing the random noise by a factor of 10 isn't enough, you
should indeed investigate post-processing noise reduction per image
(assuming you cannot reduce noise by preventing it). Do you do
darkframe/offset/bias/etc. subtraction?

I have Noise Ninja, but don't use it much because I am
never happy with the results as it seems to smooth the
images too much, washing out details.

I'm more familiar with Neat Image, but these programs are probably
equally suitable. Spend some time in tweaking the settings (often
means reducing the default amount of reduction).

However, I have never created a custom noise print for
my camera, using the checkerboard pattern and a
defocused shot.

You should, as there is no suitable featureless area in deepspace
images you could use to create a noise profile. Do make sure that you
don't make this profile until the darkframe subtraction etc. has been
done first.

My big question is: would a custom noise print created
this way be worth the effort and would the results be better
than just the box sampling I do now from the existing
image?

Although it's hard to judge without examples I am almost certain it
will improve things a lot. A program like Neat Image can run
unattended in batch mode once you've figured out the optimal settings.

Also, I'm guessing that I would need to take 100 such
samples since 100 images make up a composite, but
not sure how to go about it. I would think that each
image would have to have noise reduction before
stacking for this to work and that's how I was going to
do it.

Yes, individual noise reduction per image (in batch mode), not on the
stacked result.

So, yes, a lot of effort involved, but would the result be worth it?

Again, make sure you've done all you can to prevent the noise in the
first place, because all post-processing carries the risk of signal
loss. There would be not much benefit if the noise reduction doesn't
improve the S/N ratio.

Bart

Jason Sommers wrote:

I am an amateur astronomer and use a digicam to take pictures of the
solar
system. I take up to 100 photos and then stack the photos into one
to
reduce noise levels. [...]
My big question is: would a custom noise print created this way be
worth the
effort and would the results be better than just the box sampling I
do now
from the existing image?

You take a hundred images and wonder if a few more is worth it? You
have a weird "effort" calculus.

But whatever.

I can't comment on "Noise Ninja" (which, in the absence of other data,
sounds more like marketing than principled noise reduction -- but
perhaps silly names are an indication of a saturated market than
anything else), but whatever "custom noise print" you may record with
your sensor most likely won't hold up once you start stacking up
mis-aligned frames.

If it was me, I would just stack up several hundred images, not just a
mere hundred. This is, after all, what everyone else does:

http://www.trivalleystargazers.org/k...rn/Saturn.html

Many other hits. Also note that stacking is the only way to reduce
noise without sacraficing image detail (or, equivalently, making
assumptions about the structure of the image -- assumptions which Noise
Ninja and its ilk necessarily make.)

Jason Sommers wrote:

Not sure what you mean by this "effort calculus" but in astronomy,
one is
limited by the rotation of the object being imaged.

My meaning should be clear: if you can find the time to take 100
images of something, an extra few for a "noise print" are in the
"effort noise".

[...]

I've seen Saturn images from stacks of over a thousand frames. What
kind of "digicam" are you talking about here? Basically everyone is
mangling web or video cameras into some shape or another for this work.
If you aren't doing likewise, then either you have innovated something
markedly new and worthy of publication (e.g., stack projections of your
images, not the images themselves, thus solving the rotation problem),
or you are mis-engineering a solution.

Jason Sommers wrote:

Not sure what you mean by this "effort calculus" but in astronomy,
one is
limited by the rotation of the object being imaged.

My meaning should be clear: if you can find the time to take 100
images of something, an extra few for a "noise print" are in the
"effort noise".

[...]

I've seen Saturn images from stacks of over a thousand frames. What
kind of "digicam" are you talking about here? Basically everyone is
mangling web or video cameras into some shape or another for this work.
If you aren't doing likewise, then either you have innovated something
markedly new and worthy of publication (e.g., stack projections of your
images, not the images themselves, thus solving the rotation problem),
or you are mis-engineering a solution.