Noise Reduction Canon 450D

[damo_Melbourne]

Hi all,

Should I have the custom function setting on the 450D Long Exposure noise reduction turned on or off?
Im just starting off with 30sec and 60sec shots unguided at this stage.
It kicks in for anything over 1sec.
I would have thought to have it turned off and apply numerous darks or is it benefical having it left on?

thanks
Damian

[Tilt]

Depending on your ISO setting, I would probably leave it on. Or alternatively, you could take a 60 sec manual dark right after each of your 60 sec lights. Using ICNR just makes things easier, as you can start your image run and darks will be taken for you automatically.

Michael

[darrellx]

Damian

I have a 400D and I leave the ICNR option turned on. From the comparisons I have done, the images appear to look better in my eyes when I have it on, compared to having it off and applying darks.

I have also tried leaving it on, AND apply darks. I think this just adds noise.

So I leave mine on most of the time and apply flats.

Darrell

[damo_Melbourne]

Hi guys thanks for the fast response.

Darrell, when you say just apply flats what do you mean exactly ?
I have INCR turned on for ISO800 then Im applying multiple 60sec darks for my 60 secs exposures so is that why my pics are noisy?

If I have INCR turned on I should not be applying darks at all?

Im using DSS and it has a flat dark option so Im not sure what a flat is now - I thought it was a bias frame taken at the fastest exposure setting possible?

Then I read the following and decided I needed a drink! :)
http://www.iceinspace.com.au/forum/showthread.php?t=41588&highlight=flats

Excuse all the stupid questions!

thanks for any advice!
cheers

[White Rabbit]

I have mine on Auto, I have a 1000d. I was wondering the same thing a while ago and read somewhere that it's more effective to use the inbuilt noise reduction as it takes the dark frames when the camera is at the same temp as when the photo was taken. Makes sense to me.

Thanks
Sandy

[darrellx]

Yep, I was told that it is more effective/efficient to use the inbuilt feature to remove noise. Essentially a dark is applied to the photo after every shot. The alternative is to take a set of darks at the end of the run of lights (or subs) - which is what you do with ICNR turned off. Taking the darks at the same temperature as the light frames (subs) is important - well critical.

So I take my series of lights with ICNR on. I take my flats - 10 or 12 of them. Then let the software do the combining. I use Deep Sky Stacker (DSS).

Damian, I would suggest that you maybe introducing noise by having ICNR on and then applying darks as well. Thats what happened with mine. So I stopped.

The Flats that I use are "light flats". The scope and camera are left the same as when I took the lights (subs). I then put a light box over the end and take the flat shots. I previously used a white t-shirt, then progressed to a cotton sheet. Now I have a box with some leds in it.

I believe you can take "dark flats" but I don't bother.

Darrell

[Octane]

Damo,

Leave ICNR off and take dedicated dark frames. You'll thank me later.

Also, take dark frames for your flat frames, too.

As for bias/offset frames, some people use them. I have found that they introduce noise into my final composite, so, don't use them. Furthermore, your bias noise should be contained within your dark frames, as well as flat darks.

I guess it all depends on how critical you are and what your expectations are.

Regards,
Humayun

[Tilt]

If you have ICNR turned ON then you don't need to take darks.

On my next image run I plan to do the hard slog and take darks after each light frame, I have not done this before (I have always used ICNR) and I want to see how it turns out.

Michael

[bojan]

Darks are taken to deal with hot pixels, not to deal with noise.
Noise is averaged out later in stacking process.
That is why it appears there is more noise if you use darks AND internal long exposure compensation - that means darks were subtracted two times from frames, and that defies the purpose of taking darks.
Now, when to take them, that depends.. I have the feature enabled (400D) and I have no problems.. (it only takes 2x longer to actually take images, but then again, I am not in a hurry.. and this way I know the darks were taken at the same temperature as frames.
It is good idea to take flats with the same procedure (with darks).. for the sake of consistency of the whole process.

[damo_Melbourne]

Can I just check that we are talking about the same thing here... my exposures with "Long exposure noise reduction" as per the 450D manual page 155 = ICNR correct? INCR is not mentioned anywhere in the manual from what I can see.
The reason I am asking is if I take a 60sec pic then thats all it takes - its not taking another 60sec exposure ie a dark.
Im firing off multiple 60 secs lights one after the other or is this happening in the background and I have immediate access to further shots?

Thanks for all the advice, sorry if this has been covered elsewhere but I have done a heap of reading up on this here and Im still a tad confused

cheers
Damian

[luka]

Absolutely correct!!!
(and you can thank Humayun straight away :D)


When you subtract your darks from your lights, the final image will contain the noise from the light and from the dark added together.
In other words, the ICNR reduction will double the noise.

The fact is that any dark subtraction will add noise to your lights. However, if the amount of noise in your darks is less, the amount of noise added to your lights will be less. The same applies to flats.

To minimise the noise in your darks take multiple darks and stack them. The darks have to be taken at the same temperature as your lights, i.e. straight after. Do not use ICNR.



If you want to know more how it all works, see here:
http://deepskystacker.free.fr/english/theory.htm

Luka

[bojan]

Absolutely incorrect, I am afraid.

Noise is a random signal, and it is pretty much irrelevant if it is subtracted or added in averaging process (the AC part of course). The result of combining noisy frames on top of each other is not multiplying the noise level, but averaging (smoothing, due to th enature of noisy signal). Only permanent features (like hot pixels and bias), however, can be removed (by subtracting) or multiplied (by adding dark frame by mistake).

The ONLY important thing is that the darks are taken at same temperature as light frames (because those permanent features are VERY dependent on temperature of the chip), and that the ALL other settings are identical.

So, INCR per se WILL NOT double the noise, provided the other parameters were the same. What could go wrong here is if you have Internal long exposure compensation ON, and if you apply darks during stacking later, then the whole math done by stacker will not be done correctly and the final image will still contain the hot pixels and other permanent nasties (inversed, though)
Also, the problem will be there if the temperature was different when taking lights and darks , or any other parameter was not of the same value (ISO, or exposure time). That is why I take darks at the same time as exposures (by having long comp ON), and I do not worry about them later.

As far as INCR is concerned, I will have to have a look at 450D manual to clarify the definition.. But I am pretty sure it is the same as what I meant by "Long exposure Compensation". Canon (and others) usually do not change those things from model to model without a lot of announcements, so I am pretty sure 400d and 450D are very similar/same in this respect.

That means, when this feature is ON, camera takes exposure, then takes dark (same "exposure" time) and subtract them from light (to remove permanent features from final frame). The only drawback is double time needed for this.. a problem when taking a long exposures, and/or plenty of them.

However, if INCR in 450D world is something else (like some sort of low pass filter), it is better then to set it to OFF..

[luka]

Hi Bojan,
Read the deep sky stacker page I have attached in my previous post. I am afraid you are not correct.
From the article:
...Of course when you subtract one dark frame to one light frame you will remove the hot pixels which may give the false impression that the dark subtraction did its job.
However, at the same time the subtraction doubled the noise of the calibrated light frame and ruined it thoroughly...

And ICNR would be exactly one dark subtracted from the light unless some other filtering is applied (which I am not aware of).


Remember, while the noise is random, it still follows the standard error analysis and it is added when two signals are added or subtracted. If you have 16+-4 and 9+-3 the result from the subtraction is 7+-7 (and not 7+-1 as many people think).

[damo_Melbourne]

I think I have started something here! :)

What I will do is turn off the noise reduction and take dedicated darks at the time of the lights. The trick will be to stack the darks and make a master dark first then apply that. as far as flats are concerned I'll tackle that one next!

I found this too which may be of use ...
http://www.covingtoninnovations.com/dslr/EOS300Dastro.html

Will see how I go.

Its an interesting topic.

Will thank Humayun...now! thanks mate and thanks to everyone for your comments

Will post some second light pics to show the difference. It will be interesting to compare them with my first light pics where I had the noise reduction turned on.

Its a huge learning curve!
:)

[bojan]

Luka, this statement in the article is misleading .
It is true, though, that hot pixels are removed.
Also (as I stated) the mean value of noise is also removed (or, shall we say, its DC component, or bias).
However, noise (which is superimposed on bias mean value and which is randomly distributed around it) will not double with subtraction of the dark frame.
If we are strict here, it will double the noise POWER (which is proportional to mean amplitude, squared) but that results in 1.4x (=SQR(2) higher noise amplitude, and not linearly, as in your example. Actually, even this statement is not quite right because noise distribution with CCD (and I think CMOS is the same) is binomial, and not Gaussian...
Now, during the stacking process, those noise peaks are averaged out (stacking adds the luminances of every pixel and then takes mean vaule from them or takes median - depends on settings) and the visible noise disappears very quickly.. only 3 - 4 frames are enough for first approximation.
I still think the benefit of not having to deal with dark frames later justifies having the INCR ON (for short exposures, of course..)

[Geoff45]

There seems to be some confusion about what a dark actually is. First of all. the dark current is not noise. It is unwanted signal. Noise is random, signal is (more or less) repeatable. It is a signal because it is telling you something about the nature of your imaging chip (which are the hot pixels for example), which of course you are not interested in when doing a picture. However, being a signal, it also has a random component, which is noise. That is, if you look carefully at individual darks, you will find that overall they are very similar, but that they vary slightly from frame to frame, due to the noise in the dark signal. So we have the dark current and noise in the dark current. The reason we take many darks is to reduce the noise in the dark current by averaging. This is why you'll get a better result than by using ICNR. The ICNR only uses one dark and so will be much noisier than the average of many darks. Your light has the dark current in it and this is subtracted out in calibration. If you use ICNR and then subtract the average of many darks, you are subtracting the dark signal twice. This won't get rid of hot pixels for a start, it will replace them with "cold" ones.

Have a look here (http://www.photomet.com/resources/encyclopedia/darkcurrent.php)for more info

I recommend a read of Handbook of Astronomical Image Processing by Berry and Burnell for those who really want to get their heads around this stuff, which is not an easy thing to do.

[bojan]

Geoff,
you are subtracting darks twice ONLY if you have INCR ON, AND if you supply dark frames to stacker.

[Geoff45]

Isn't that what I said? Maybe it's not clear. I'll edit my post

[deadsimple]

I've got a 450D and I use ICNR especially for timelapse/sequences.

Why? I don't want to disrupt my sequence timing by using up a timeslot for a dark, nor lose focus by pushing on the lens when I put the cap on. I end up halving my 'framerate' (so to speak), but at least they are regular and the camera is untouched between shots.

If someone can suggest how to take a dark without touching the camera I'll give manual darks a go :)

[Octane]

Ash,

Take the darks /after/ your imaging session is complete.

If you're really concerned about the temperature changing that much from the beginning of your imaging run to the end of it, then, take a few (half-an-hour's) worth of dark frames before you begin imaging. Combine them with the ones at the end of the run.

I've never done this; always taken darks at the end of my imaging run and never had any problems, at all.

Regards,
Humayun

[EzyStyles]

I have to agree with Humayun. Take your darks at the end of your imaging run. ICNR off. Have a library set of different dark exposure time, ambient temp and ISO settings. Once you have gathered enough darks from various temp, expo, iso, all you have to do is measure the first light frame temp to the last of the particular image you are taking then average the image temp.

example: first light frame of M42 measured at 10 degrees
last light frame of M42 measured at 8 degrees.

Averaging 10 from 8 = 9 degrees.

grab a 9 degree dark set at the same expo & iso time. load everything up in Deepskystacker and let it do its job. It works everytime. :thumbsup:

Don't worry about ICNR unless it "works" for you. I know it didnt for me. Save the explanation and research about the differences between hot pixels, darks and go all nitty gritty on it etc etc. It is not required to learn the fundamentals behind it if it doesnt work for you. test it out for yourself first. :)

Your library of each different dark set should consist of 10 or more dark frames (20 will be better) and have an average of those 10 dark frames which you use to subtract from your light frame. Hope that explains it .

cheers

[Tandum]

I'm with eric on this one. It's not dark long enough to be doubling your exposing time using ICNR. Your far better off using that time to take lights. Build up a dark library when it's overcast, you don't need the scope setup to do it.