Best method of reducing noise?

[rubecula]

Imaging with a DSLR produces a lot of noise, particularly at this time of year.  So I'm wondering what's the best way of dealing with this (I know what some of you (Olly) will say - buy a CCD).  The three options seem to be:

Darks, dithering or using a peltier cooler (http://www.pampaskies.com/gallery3/Equipment/cooler_v3)

The cooler seems quite inexpensive and straightforward to make but introduces more weight.

Does anyone have any views on the respective merits of the 3 alternatives?

Thanks

[daz]

I would say all 3!

You should really be taking darks (and flats and bias) frames anyway. Dithering will help further, and as dark current is thermal noise, you really want the sensor as cool as you can..

I have a CCD and I use darks and dithering and cooling!

[Alien 13]

For me with short FL scope or camera lenses I dont find the need for darks, newer DSLRs seem to have much better noise performance so I only use bias frames.  Dithering is the best overall method to reduce noise but keeping a DSLR in its happy zone helps.

Alan 

[Jonk]

Dark noise 'should' be random, so stacking is a sure fire way of reducing noise in images.

There will always be some noise in DSLR subs. I regularly shoot static MW wide shots at iso 6400 or higher, and by stacking and median filtering 10 or more frames, reduces the noise extremely well.

[ronin]

First really is have you taken the exposure then allowed a reasonable amount of time for the sensor to cool down before the next exposure? Simplest rule of thumb is a wait/cool down period equal to the exposure length.

Next as mantioned is take some Darks, I have only played around a bit but I simply take half as many darks as I took exposures. Easiest option is camera and intervalometer in the fridge, it is dark in there, press the go button, close the door, make a coffee, then take camera out.

A peltier sounds fine but not sure how effective on a DSLR it can be. My DSLR has a live view screen on the rear, and cases these days are rubber not generally metal so heat exchange is poor. I would have thought that getting the cold side of the peltier inside and the hot part outside is not possible. So not sure if this is really a realistic option.

Always remember that a DSLR was not designed for these long exposures that imagers make them do. People are in the situation where their DSLR can managed it, but that was not the criteria they are designed/built around. So will never quite be "The Instrument" for AP. In most DSLR's any exposure over 1 Second needs a noise reduction exposure of the same length as this is what they default to if left to themselves. 1 second is a short period yet that is the general exposure duration after which the manufacturer expects noise to be too much and so needs correction applied.

 

[StuartJPP]

This is only my opinion but I'd rather have 50 frames to stack with noise in them than 25 frames with a little bit less noise in them. I believe that if you are not exposing and the skies are clear then you are wasting precious imaging time, I leave 5 seconds between frames, basically enough for the subs to be downloaded to the PC.

Like I said, this is just my opinion YMMV.

[uhb1966]

...and find out the ISO "sweet spot" , ie the highest ISO which does not add too much noise. With my Canon, i stay around 400-800.

[wimvb]

In no particular order:

longest exposures your setup/conditions can handle at "low" ISO

as many subs as you have patience for, then double. Seriously: the more, the merrier

darks may or may not add to quality. Good cosmetic correction and pixel rejection can sometimes replace darks. It seems very individual, so experiment.

dither, at least 12 pixels for a DSLR

an external cooler may be useless if your camera is mainly made of plastic. It will cool, but how much??

image during low noise conditions: winter. It's a natural cooling system for the electronics. The skies are darker, so less "sky noise" as well.

I won't advise you to get a CCD, get a cooled ASI1600 in stead. It has cooling built in, is CMOS (almost DSLR), and has rediculously low noise to start with. There is a very active thread in the equipment section of this forum.

[wimvb]

3 hours ago, StuartJPP said:

This is only my opinion but I'd rather have 50 frames to stack with noise in them than 25 frames with a little bit less noise in them. I believe that if you are not exposing and the skies are clear then you are wasting precious imaging time, I leave 5 seconds between frames, basically enough for the subs to be downloaded to the PC.

Like I said, this is just my opinion YMMV.

I agree; you ever know when clouds roll in. And since some frames are lost due to satelites, planes or gusts of wind, you'd want as many as you can get.

[rubecula]

Thanks Guys, some really good advice here for me to work through.

[wuthton]

The following video from Craig Stark is an excellent overview of how to reduce noise. It's quite long but well worth watching from beginning to end.

 

 

[MARS1960]

11 hours ago, rubecula said:

Imaging with a DSLR produces a lot of noise, particularly at this time of year. 

 

It sure does, especially a Canon. 

Aftrer many years as a Canon diehard and only a few months into astrophotography i'm looking to buy a Nikon D5300 to have modified, the noise that these produce is miniscule at certain ISO's compared to Canon.

[MattJenko]

Dither and sigma clip in stacking. I have never found darks to be useful with a non cooled DSLR, it just makes the eventual noise worse. Dither + stacking removes hot pixels and as long as you have enough subs, noise is manageable without the need for epic post processing noise reduction processes.

[alacant]

Add flat and bias frames to your dithered snaps and lose the dark frames. HTH.

[rubecula]

Thanks again for all the help guys.  As a first step I'll try to get my head round dithering.  I thought it was just a matter of selecting the option in BYE but it looks a bit more complicated than that.

[wimvb]

There's a video clip on youtube with Tony Hallas, about astrophotography with a DSLR. He talks about the advantages of dithering. Just google "Tony Hallas astrophotography".

[ollypenrice]

If you mention noise to Olly he will not instantly say 'CCD,' he will say, 'Take a lot more data.' So will Sara! Only then will he say 'CCD.' I think the key points are all well made above. I think DSLR darks are next to useless. I don't even use them any more with set point CCD. Bias as dark, big dither (Craig Stark), sigma clip, a good hot pixel filter in the stacking software and many many many sub exposures. 

To be honest I've lost interest in calibration files, other than flats which are vital, because I'm using an ancient bad pixel map, an ancient master bias, a very aggressive hot pixel filter stacking in AstroArt 5, and I have never had so few issues to clean up at the end. This is why I haven't updated my bias or BPMs. What problem would I be trying to solve?

Olly

[Alien 13]

The only DSLR dark that I have known to work is the in camera long exp noise reduction it does a very good job of matching temperatures and is fine for single shot widefield where you have to get both the sky and landscape together.

Alan

[Uranium235]

Nobody has yet mentioned focus... that has to be as good as it can be in order to define between what would be a very dim signal, and what is noise.

Also, pick your nights well. Camera electronics arent the only source of noise - the sky can be pretty noisy sometimes as well, depending on air moisture content or high level, thin cloud.... often made worse when its lit up by the Moon. Its not always a good idea to go out whenever it looks clear, especially when adding to an existing project because any data taken on a bad night will just degrade the data you took when the conditions were more favourable.

[rubecula]

16 hours ago, ollypenrice said:

Bias as dark, big dither (Craig Stark), sigma clip, a good hot pixel filter in the stacking software and many many many sub exposures

Thanks Olly, that's now my mantra for the next imaging sessions.  Not sure how long it will take me to sort out dithering.  Need to determine the parameters to give me a 12 pixel dither (thanks Wim) followed by a suitable calm down period.  Hopefully I can get it working and tuned to my kit in a single night and then follow that up with a full night's session imaging. 

[alacant]

Hi. Dither is very simple to set up. It is built into APT and if you guide with PHD2 you can refine it exactly to your taste. Simply choose how many pixels to dither and start imaging. You can't decide exactly how long to settle until you have seen the first few exposures and seen how the dither hits the mount. Take the default values to begin with. They rerely need altering. Again, if you have PHD2, you can see how long it takes to resume normal guiding by observing the (alarming the first time you see it) guiding spikes to settle after each exposure; it will not contuniue until guiding has settled. Example on my rusty EQ6, a 10 pixel dither rarely takes longer than 20 seconds to settle even if it's jolted DEC which has backlash. HTH