Long Exp. v's loads of stacked subs?

[roundycat]

Dave (centroid) makes a point above in this way (post 38) and I quote 'Bear in mind, you cannot put into an image, what isn't there in the original. If a short exposure, not matter how many, doesn't contain certain detail, then that detail is not going to appear by 'stacking' and processing'.

I would add this; some detail may be 'lost in the noise' but when added to other frames containing the same weak signal it is intensified in relation to the noise which is random regardless of its source. Have a look at the five pictures I posted in post 37 and you will see detail coming out in clip five that is hardly visible in clip one. The short presentation I have for our next meet might make this clearer.

Martin, in post 46 mentions dithering. If you really want to clean up your subs as you stack then dithering is a must. That, combined with sigma reject stacking, will leave you speechless, a point made in the presentation mentioned above.

Dennis

[ngc2403]

Dennis,

can you describe 'lost in noise' please. looking at your images i can see that they are post-process, but it is true that the images are better the more you stack this however was never the issue, it was that is the ten stack image, ten time better than the single image?

hard to judge in a raw, almost impossible in the processed image.

The problem with all stacking methods that are not mean, is that in removing noise, you also remove signal too. This is only a processing short-cut, that causes a loss in information from the image and not a solution to the bigger problem.

[centroid]

in removing noise, you also remove signal too. This is only a processing short-cut, that causes a loss in information from the image.

Precisely Ally, and aligns with what I said earlier:

"the most information/detail contained in an image, is in the orginal image as delivered by the camera"

Any processing that we apply to that image, destroys some of that information.

In other words "there's no such thing as a free lunch"

Dave

[CELESCOPE]

Well this has been an interesting thread , but now i think its going around in circles , best finish on this one , but thanks for all the input ,i for one have got something out of this ,and hopefully some of our newbies.

Cheers

Rog:)

[roundycat]

Ok, here goes! I hope you guys don’t think I am trying to drag this down to the level of an argument about semantics.

For the moment forget about fancy methods of combining images and just consider adding or average which is much the same. Both are strictly mathematical functions. All noise, regardless of the source is random. The actual signal from your target is not random, only the measurement uncertainty of the signal is random and with a properly exposed image that uncertainty can be quite small (wrt the signal) and will be different in each image because of its random nature.

Now for the tricky bit. I do not consider that stacking or combining is the same thing as processing. Going back to the idea of simply adding frames, the signal will reinforce and the noise will tend to cancel out. Therefore if you add a number of frames the signal will gradually take on a more definite form and the noise will be slightly suppressed. Hence it can be said that the signal is no longer lost amongst the noise as for a single frame but is rising up out of the noise and becoming clearer. See my Rosette pics. Where nebula modulation and star sharpness are both improved.

As far as processing is concerned you take the summed image stack and try to turn it into a pleasing picture. That is where all science effectively ends and artistic or aesthetic values take over. We should be guided in our processing attempts by the fact that noise is unwanted and should be suppressed as much as possible whilst not interfering with the signal as nothing of value can be added to it at this stage.

I think the Rosette clips that I posted above show this. They were all processed using exactly the same values in PS Levels and Curves so as to provide a fair basis for comparison. Once you start processing with a vengeance all the science, maths and logic fly straight out of the window. I don’t think that applies to stacking and it is not helpful to consider both stacking and post processing as the same thing.

In defence of sigma reject stacking methods I will say this. With careful use of the settings for the sigma reject filter the only things that are rejected are pixels that are outside a pre-set limit. They are replaced with the median or average value (for that particular pixel in the stack) and this process will generally only affect pixel values that are way outside the noise in the form of ‘hot’ pixels which we don’t want anyway. The vast majority of pixels in a 10- 30 frame stack are not adversely affected. I have measured actual pixel values across a stack and verified this. When stacking, the standard deviation value is that for the population (the population is the number of frames in the stack) and that makes the SD Mask (or similar) method as accurate as it can be.

I made a comment earlier in the thread about keeping one’s feet on the ground. When Gary asked his original question I am sure he was not wanting to be treated to an extended maths lesson. Whilst not decrying in any way the value to individuals of understanding what is going on from first principals we should not lose sight of the ball for the sake of playing the game.

On a slightly different, but hopefully still helpful note, I would invite anyone interested in this debate to view my web page (http://www.dens-astropics.org.uk/) and note this; None of the pictures taken by the Artemis camera had darks subtracted. An awful lot of the pictures taken with longer focal length scopes did not have flats applied. In some cases flats were used that were not dark or bias subtracted, just a hot pixel filter. If you can think of any more cardinal sins I may have committed I probably did commit them. The end result in most cases, certainly not all, is a reasonably noise free collection of deep sky images. Something I think most of us are after.

The conclusion I draw from this is one I have pretty much stated before. Most of us are in this to make pretty pictures. When a new member asks about exposure times he is generally only interested in a simple answer in order to get on with his new hobby. That is mostly where I focus my attention and will continue to do so.

Dennis

[Psychobilly]

I'm in the "art" camp and believe in the suck it and see "practical" approach for determining what works for "you"....

Billy...