M82 M83

[earth titan]

Need a little advice if possible, and thanks in advance.

I've attached two images, both subjected to pretty much the same processing. The first has no darks, flats or bias where as the second has everything.

I know there is only 55 minutes of data (ran out of clear sky) and a lot of noise (it was ISO 1600) but when run through DSS there is a noticeable difference. It's as though the addition of calibration frames etc. makes things worse.

[Shibby]

How many dark subs did you use to calibrate the image, Tom? If not enough, they will add noise to the image (and the second one does look noisier).

[earth titan]

How many is a good number?

[Stephen]

What happens if you use just the darks and flats (leave the bias out?)

I can't find it, but I was reading a thread the other day that suggested DSS does it wrong when it comes to bias and does a double subtraction (hence introducing noise!)

That thread could of been out of date though

[earth titan]

Interesting. Not tried that combination.

Back in a mo...

[earth titan]

Nope. Still makes the images worse.

I guess I need LOADS more data.

[Demonperformer]

I have seen figures that suggest the optimum number of darks/flats is 20 ... don't really understand the math, but something to do with getting the best results without falling into the realm of diminishing returns.

There are regular outcrops of 'has dss got it wrong' threads. One solution suggested was put all your flats and bias frames and one light in dss and process. This gives you a master flat. Then put that with all your darks and lights into dss and process as normal. However, I am sure that someone got in touch with the writer of dss, who explained how it did all work properly really. [grabs tin helmet and runs for cover]

BTW: surely that's M81 rather than M83

[Shibby]

Try stacking without the darks. How many where there?

Apparently one dark addes 41.4%, two adds 22%, three adds 15% and so on... Check out this graph here: Astrophotography

[earth titan]

Found this on another post:

This is the link:

http://stargazerslounge.com/imaging-discussion/129762-calibration-dss-doing-wrong-response-luc-dss.html

Goes onto to discuss DSS in great depth, thanks to the OP, about the manner in which DSS adds calibration frames. Will be having a go at some of these this evening and post my results.

[RikM]

I try to take lots of light frames. At least 20, better 60 or so. I have found that 20 Darks seems to be enough. I can't see any difference compared to 30. Normally I take 30 flats and 50 bias because these are so quick to do.

I only temperature match the darks though. Flats I do the following morning and a take a master Bias only once every few months. I have one for each ISO setting but I never heard that Bias needs to be temperature matched though?

[earth titan]

Thanks Rik.

This is a dumb question, but is it possible to 'fool' DSS and use the same dark but copy and paste in the same folder and rename 20 times, so it 'thinks' it has 20 darks....... I know this is a cheat and not really the purpose, but I wonder what the effect would be?

Perhaps I'll try it....

[Stephen]

Thanks Rik.

This is a dumb question, but is it possible to 'fool' DSS and use the same dark but copy and paste in the same folder and rename 20 times, so it 'thinks' it has 20 darks....... I know this is a cheat and not really the purpose, but I wonder what the effect would be?

Perhaps I'll try it....

It would be exactly the same as providing one dark. As you point out, this is not the point

[RikM]

If you just copy one, the darks will all be the same. You don't have any randomness to average out, which is the point of getting 20 or more. You would get the same result as If you just used the one on its own.

There is no shortcut to this I am afraid, but if you put the effort in, the result is worth it.

[earth titan]

As always, brilliant answers. Thanks chaps.

Like most things in life, there are no shortcuts....

[Demonperformer]

This is a dumb question

No such thing!

The dumb thing is when we think we have all the answers without asking questions.

Keep them coming.

[Spikey]

For the amount of time on the subject thats a pretty good image.

[Shibby]

Like most things in life, there are no shortcuts....

Well.. the great thing about darks is that you can take them after-the-fact whenever you like, so long as the exif temperature is within about 4degrees of you light frames. Then, you can build up a library of master darks to use whenever you need them - a great shortcut!

[RikM]

My Monkey Head image, I shot the darks the following day with the camera in the fridge.

[pixueto]

My Monkey Head image, I shot the darks the following day with the camera in the fridge.

LOL you made me laugh. I never thought of that, that's quite clever.

[ollypenrice]

Diminishing returns really come in at around fifty exposures, not twenty.

The problem with DSLR darks is that they may not be temperature matched and can, therefore, be all over the place. The bias frames will help the software to scale the darks, at least as I understand it. I use set point CCD so I've no personal experience.

What I think is happening here is that the darks are too aggressive and are clipping the black point. I might be quite wrong but try stacking without flats and then without darks. Something is clipping the stacked lights. You need to find out what is doing it. Maybe the darks were taken at a slightly higher temperature.

Taking them in the fridge is only useful if the fridge is close to the temp of the lights to which they are applied. It can be a good idea or a waste of time. The thing is to temp match the darks and lights, which is vital.

Olly

[RikM]

I checked the exif temps. My lights were 12-13°C and my fridge darks 17°C so I wasn't expecting a good result but I think I got away with it. I asked on another thread how closely matched they needed to be and the advice was: within about 4 or 5°C.

[Demonperformer]

The thing is to temp match the darks and lights, which is vital.

Don't know if this is a good idea or a waste of effort, but I tend to do half of my darks at the start of the session and the other half at the end, to try to compensate for any temperature drop or rise while I am taking the lights, in the hope they will sort of balance out right. If I actually notice the temperature changing a lot, I will sometimes stop and take a third half(!) in the middle and combine each set of lights with the half taken before it and the half taken after it (if that makes sense).

[Shibby]

There is a handy tool called Dark Master (*roots around*) here: Dark Master / Wiki / Home

...it organises your darks & lights automatically by exif temperature.

[George]

I used to shoot my darks at the end of the session when I was using a DSLR, minimum darks I used was 21 but on average it was 25-31. I read on the DSS user group that Luc suggested an odd number of dark/bias was better (something to do with the algorithms used).

It can also be useful to have a darks bank, this can be done on those cloudy nights or indeed any time a useful temperature range is present.

Just collect your darks at 2C intervals and keep them for re-use.

[Cloudwatcher]

I'm not a puka imager so the techno speak is way over my head I'm afraid........ I mean,Dark Flats.....what's that all about......

Not the same as Gloomy Apartments are they?

Seriously though,there is some detail to be found in your image if dealt with initially as separate layers on a large scale.

Having said that and looking at my effort the star layer looks a bit rough.....Still,can't win em all,as they say........