Moving away from darks...

[ollypenrice]

I've always used darks except on ultra quiet Sony chips. That may be about to change. I know Harry and Dennis and a few others argue against them but they've worked for me till now so I have stuck with the routine I know.

However, we've been struggling with a set o 30 min subs from Yves' H36. They just came out of the stacking badly black clipped. Odd.

Since the set had been dither-guided I tried a Sigma stack using flats, a defect map and a master bias used as a dark, not a 'real' dark. Bingo, perfect. Thanks to Per for pusing me into this experiment.

I made the defect map in AstroArt. The H36 generates a lot of hot pixels so to make a defect map you use Arthmetic > Clip and clip 2000 off the bottom. Leave the top alone. This produces a black screen dotted with hot pixels, or defect map. In AstroArt 5 this defect map has its own box on the stacking page where you put the lights, flats, darks etc.

SO, dither guide, sigma stack and apply bias(as dark) and defect map but forget darks. Suits me. Thirty minute darks are a prime pain in the rear to shoot.

Olly

[johnrt]

Interesting Olly, I'm just going in the opposite direction. I have until now never used darks on my "super quiet no darks required" Atik 314l+.

But I thought I'd give them a whirl in a bit of a random experiment in an otherwise totally free afternoon of laying about. I found they made a good difference and was kicking myself for not using them sooner!

[ollypenrice]

Interesting Olly, I'm just going in the opposite direction. I have until now never used darks on my "super quiet no darks required" Atik 314l+.

But I thought I'd give them a whirl in a bit of a random experiment in an otherwise totally free afternoon of laying about. I found they made a good difference and was kicking myself for not using them sooner!

Out of interest, John, try using a master bias instead of a dark.

The dither guiding/Sigma stacking is a part of the noise reduction programme I used, though, because if you sample the same sky on the same pixel every time you don't average out or Sigma reject the rogues. That said, you'd need remarkable polar alignment not to obtain natural 'dither' through slight field rotation.

Olly

[johnrt]

I was just using a master flat frame (about 20 or so flats each time I imaged) and a master bias which was around 100 frames.

The darks I'm using are bias subtracted so as not so subtract the bias signal twice from the lights.

I get plenty of natural dithering from overloading my HEQ5 with too many scopes!

[harry page]

Hi Olly

Welcome to the Non Dark side This is exactley how I do it the defect map is excellent and the dither / sigma takes care of all the rest

Next you will be using HDR comp in Pixinsight

Harry

[Sam]

This is interesting, I guess the dithering, either intentionally or not kind of makes the noise random (a bit). So would be removed through stacking and if you got rid of the hot pixels through the defect frame then bingo it should all work. I guess I've just convinced myself to give it a go.

[Sparrow]

This is fascinating. Not sure I totally understand it, but no darks is a very attractive proposition.

Olly, I don't have Astroart. Would it be possible in any way to make the defect map with Photoshop? and if so, how? - big quesiton I know but if you could help with this I'd be grateful. I'd really like to try this out myself.

Regards

John

[johnrt]

Hmmm. Is this just a conspiracy to get me to buy AA5?

[ollypenrice]

You have to clip 2000 ADU off the bottom of the dark frame to make a defect map. That should be possible in Ps but what will you do do with the defect map when you have it? You'll need a stacking software thancan incorporoate it into the routine.

Astro Art isn't expensive...

Olly

[Sparrow]

Thanks Olly - I'll have to put this one on the back burner for now. Having just had to buy a new dishwasher and a laptop for my son's birthday, our spending budget is at zero. I'll keep it in mind for the future though.

[Shibby]

Even with my DSLR, I often don't bother with darks. With my abundant tracking errors most of the dark signal is easily stacked out! This wouldn't have worked with the awful amp glow of the old DSLR (D70). Remember darks actually *add* noise if anything, not remove it. What they remove is the dark signal (assuming separate bias).

[RikM]

Is a defect map similar to the 'select and remove hot pixels' check box in DSS?

I have to say I'm chuffed to bits with how much smoother my images look using dithering and I generally use a sigma reject routine in stacking. I do find darks help more than hinder though.

[symesie04]

Thanks Olly - I'll have to put this one on the back burner for now. Having just had to buy a new dishwasher and a laptop for my son's birthday, our spending budget is at zero. I'll keep it in mind for the future though.

Was he happy with the dishwasher?

[jgs001]

Interesting Olly... I went through a phase I didn't use darks, then started using them again... I may have to go back again... especially now I'm dithering in my guiding (does that even make sense :))...

Thanks Olly - I'll have to put this one on the back burner for now. Having just had to buy a new dishwasher and a laptop for my son's birthday, our spending budget is at zero. I'll keep it in mind for the future though.

John, take a look at pixel fixer (Pixel Fixer | Hot pixel removal utility) it would be used pre stack in DSS I guess and might be the equivalent of applying a defect map in astro art

[George]

Altho I was a strict dark user with the DSLR I moved on to the QHY8, this uses a Sony sensor so darks have been consigned to the bin

Obviously I still use flats and bias frames and Im still struggling with the perfect flat, altho the masters look really good they are tending to overcorrect

[ollypenrice]

I find the AstroArt hot pixel filter only slightly effective whereas the defect map really nails almost all of them. I've just done a back to back comparison of the two luminance panels from our recent M81/82 in the 14 inch. The original had darks, flats and the hot pixel filter. The new version had flats, defect map, bias as dark and hot pixel filter. The new system is certainly better. Smoother, quieter and slightly deeper.

Olly

[martin_h]

Nebulosity2 and now 3 includes a Bad Pixel Map routine in its calibration steps.

[Cath]

Has anyone tried this plugin for Photoshop I wonder?

Hot pixel removal

[vdb]

Glad your convinced ... I've been preaching this since long ... darks will inject extra noise ...

BPM - Bias - Flats is what in my opinion is the best way to calibrate lights. And dither is a nice bonus ...

Yves.

[ollypenrice]

Glad your convinced ... I've been preaching this since long ... darks will inject extra noise ...

BPM - Bias - Flats is what in my opinion is the best way to calibrate lights. And dither is a nice bonus ...

Yves.

Sorry to have been slow on this, Yves, but with my Atiks I haven't had any issues so I haven't tried to fix them! However, guess what my next Atik experiment will be!

Olly

[vdb]

No problem ...

I don't have the issue with my H35 either ...

But the fact remains that using darks is in my opinion worse then BPM - BIAS ...

This has been debated to death already and there are believers of both sides. I'm in the BPM camp ...

It's not that because you couldn't see the issue that you will not benefit ... my guess is that last year when you tried to get the IFN, with BPM you may have a chance ... would be interesting to see if indeed it's the case.

Especially when doing 30 minute subs, and the number of darks required it makes bpm even more attractive, and as already mentioned with bpm you do not risk injecting "extra" noise.

Yves.

[Warhen]

Cath, There's also Franks Astro Hot Pixel Studio Filter and in principal both are fine, but we need to get rid of hotties during preprocessing before alignment. If they remain- they get stretched out during registration and HP filters become ineffective.

Hey guys! Olly, I'm going to agree. Like several of us, back in the SX/Sony days I didn't do Darks. Began using them with the Atik 11000C as the Kodak is much noisier. Experimenting again however, my Light - Bias/Flat - Bias looks much cleaner than Light-Dark. One caveat, my Master Dark (25x) is 20m and I'm currently shooting 15m. Noisiest of all is Bias scaling the longer Master Dark. Better is the Master Dark. Smoothest is using the Master Bias (x50) with a BPM. This is coupled with dithering and a Sigma combine and Astroart 5's hot pix filter which is very effective at the default of 30%. I needed to clip the dark much further down to 350 for my 11002's Defect Map and it's a tricky balance of not enough/too much. Sony's are pretty right-on at 2k and I'm surprised the H36 (same sensor) responded at 2k.

Anyway, I was pretty darn surprised how good the bias route looked! One residual problem are the weak vertical discharge columns which Darks do a better job of eliminating. Any comments here, are you guys seeing this? I may stick with this route but compare to a new 15m x 50 series of Darks.

[x6gas]

Thanks for posting this Olly - have been looking at defect mapping in AstroArt myself.  I haven't used darks at all since getting the Atik 4 series cameras and have been using hot pixel correction but seems it's time to give defect mapping a go.

I understand that the defect map is made from a dark frame in the way that you say (clipping everything above an ADU value of 2000 to zero, but does it matter what dark frame you start with?  It is better to use a master dark from a stack or will one do?  Do they need to be temperature matched and does the exposure duration matter?

Thanks in advance, Ian

[ollypenrice]

Thanks for posting this Olly - have been looking at defect mapping in AstroArt myself.  I haven't used darks at all since getting the Atik 4 series cameras and have been using hot pixel correction but seems it's time to give defect mapping a go.

I understand that the defect map is made from a dark frame in the way that you say (clipping everything above an ADU value of 2000 to zero, but does it matter what dark frame you start with?  It is better to use a master dark from a stack or will one do?  Do they need to be temperature matched and does the exposure duration matter?

Thanks in advance, Ian

I can answer this with absolute precision; I don't know! I think that using single subs for anything in a statistical game like digital imaging is likely to be a no-no so I'd make a stack. I made my defect maps out of 30 minute master darks and this seems to have no ill effect when applied to 10 and 15 minute lights. This is part of the convenience of the system. One BPM fits all. I need to look into Warren's point, though, about how much to clip. I only went for 2000 because AstroArt told me too! (Baaas like a sheep...  )

Olly

[x6gas]

Thanks Olly, that's helpful.

I'm happy to experiment, but if the work has already been done and the answer is out there...

From a couple of darks taken this morning I was a bit surprised that a 300s and 600s dark had different hot pixels... I'd have expected the 600s exposure to have all of the hot pixels from the 300s sub plus a few more so I'm thinking that the safest way to go is to stack a healthy number of temperature and exposure matched darks but I think I'll start with a modest set of 15 minute darks and see where it takes me.

Cheers, Ian