Master Bias and Master Darks

[badger111]

So, I bough a second-hand Canon EOS 30D and over the past couple of nights I've piggy-backed it on to my Skywatcher with the RA motor going. As the seeing has been really good the last couple of nights (especially last night) I've been in the garden taking loads of subs of various cosmic wonders with a view to processing them to see what I can get out of my current kit.

I took 30 darks at the same ISO as the sequence.

30 flats at the same ISO as the sequence.

30 dark flats at the ISO as the sequence.

30 bias frames at 1/8000 (with the lens cap on)

I've read lots of articles around the web regarding these calibration frames; some conflicting; some not - so I thought I'd shoot everything at the same time and ask you guys what's the best way to process them.

I made Master Darks and Master Flats by Median combining them (In photoshop) from the sequences above and found a website that gave this formula - calibrated image = ((light - bias) - (dark - bias)) / ((flat - bias) - (flat dark - bias)) - I managed - ((light - bias) - (dark - bias)) / ((flat - bias) - for a first attempt and the results were better than the original, but still firmly in the "rubbish" camp (or maybe that's the best I can do with what I got? or maybe I don't know what I'm doing (more likely:))

Any suggestions graciously received.

[D4N]

Any reason why you don't want to let DSS make them for you?

Sent from my iPad using Tapatalk

[lensman57]

So, I bough a second-hand Canon EOS 30D and over the past couple of nights I've piggy-backed it on to my Skywatcher with the RA motor going. As the seeing has been really good the last couple of nights (especially last night) I've been in the garden taking loads of subs of various cosmic wonders with a view to processing them to see what I can get out of my current kit.

I took 30 darks at the same ISO as the sequence.

30 flats at the same ISO as the sequence.

30 dark flats at the ISO as the sequence.

30 bias frames at 1/8000 (with the lens cap on)

I've read lots of articles around the web regarding these calibration frames; some conflicting; some not - so I thought I'd shoot everything at the same time and ask you guys what's the best way to process them.

I made Master Darks and Master Flats by Median combining them (In photoshop) from the sequences above and found a website that gave this formula - calibrated image = ((light - bias) - (dark - bias)) / ((flat - bias) - (flat dark - bias)) - I managed - ((light - bias) - (dark - bias)) / ((flat - bias) - for a first attempt and the results were better than the original, but still firmly in the "rubbish" camp (or maybe that's the best I can do with what I got? or maybe I don't know what I'm doing (more likely:))

Any suggestions graciously received.

Why use PS to do the stacking for AP. DSS will do a better job and a lot quicker and it is free. Just load the files and it will do the job.  There is only one way of calibrating a sub with calibration frames, the algorythm might be different due to the nature of the subs but the calibration routine is always the same.

A.G

[badger111]

Any reason why you don't want to let DSS make them for you?

Sent from my iPad using Tapatalk

Well.., I've spent a LOT of my time (probably far too much time  (it's become something of an obsession)); reading books and rummaging around the internet and DeepSkyTracker is a piece of software that keeps popping-up in various searches and articles I've read; I was never tempted to download it after I'd already downloaded Registax and thought that was a complete nightmare to use for a very noob noob, such as myself.

I downloaded DSS right after you posted your comment and just ran some of my "data" through it... AMAZING (bordering on, mind-bendingly good) So simple to use!!! I spent hours and hours fiddling around in Photoshop last night, telling myself "no pain no gain" to produce, well, rubbish.

(**note: my terms for rubbish and good WILL differ from anyone who really knows what they are doing;  it's rubbish and good by my own subjective judgement

I thought I'd share a few 100% crops (converted to JPG) that show just how rubbish I am,  and just how good DSS is.. (I did smile though when PS asked me if I'd like to adjust the JPG quality for the first image when saving the crop; that's the one I spent hours fiddling with photoshop to produce... (all three are NGC869+884 with my Canon EOS 30D perched atop my 4" Skywatcher with eq1 mount and ra tracking motor - the top 2 are 6 subs @ iso 3200 @ 30secs and the last one was 15 subs at the same ISO and exposure )

The one above was what I managed to produce last night manually in PS (the lovely orange glow happened after I'd divided my light flats)

The one below is what DSS did with the same files (although I did add the dark flats to the processing simply because DSS knew what to do with them)

The one below is the best one (**note: best one by MY standards)

Well, things can (and will) only get better. I've REALLY have got the bug; currently selling everything I own to get that HEQ5 Pro mount + Apo + Atik to take some "nice" photos..........

[D4N]

I'm glad you got it sorted

That you managed to image any DSO at all with an eq1 mount is no mean feat so well done on that count.

Sent from my iPad using Tapatalk

[lensman57]

Great news, well done. DSS is a remarkable piece of free software and quite powerful and flexible once you have learnt what it can do. I do not use it any longer

but is still on my machine just incase.

A.G

[Astrosurf]

I love this thread - you're so happy! And I agree, it's great what you caught with the EQ1. That means you'll be excellent with tthe new kit when you've got it all sorted! Onward and upward.

Alexxx

[uhb1966]

Hey , the ones with DSS are great, i whish i could do so swell

[glowingturnip]

just a thought, but I think most of the calibration steps (certainly darks and bias) have to be done before debayering, but if you were doing it in PS then you're presumably debayered already ?

But yeah, I agree with the others, let DSS take the strain !

[tingting44]

awsome! glad to see you having so much fun! it only gets better, your wallet will all ways be empty in this hobby too lool.....well mine is lol

[RobH]

Don't bias subtract your lights if you're going to use darks. The bias signal is contained in the darks so you end up double biasing, which means you add noise.

Bias subtract flats or they won't work properly.

Do your calibration before debayering.

I've a section on my website under 'primers and tutorials' which you may find useful.

[tingting44]

im pretty sure its good to use lights/darks/flats/bias......i think any way lol

http://deepskystacker.free.fr/english/theory.htm#CalibrationProcess

[glowingturnip]

Don't bias subtract your lights if you're going to use darks. 

actually, just applying a bit of high-school maths:

image = ((light - bias) - (dark - bias)) / ((flat - bias) - (flat dark - bias))

simplifies to:

image = (light - dark ) / (flat  - flat dark)

which sounds about right.  If you're subtracting bias from both, they'll cancel out.  Seems a bit pointless though.

So what do bias frames do... ?!  Interesting that Pixinsight doesn't use flatdarks, I suspect it's applying a little more science

[RobH]

Correct. Subtracting bias twice has the effect of cancelling the bias out so you have effectively not removed your bias signal, and have, in effect, added noise you wouldn't have if you'd just had the bias subtracted once.

I don't know many experienced imagers who bother with flat darks, most simply use bias instead, as the flat darks are, in normal circumstances short enough to be virtually indistinguishable from bias frames.

[glowingturnip]

actually, that might be the cause of some of the OP's original problems.  

He says he managed ((light - bias) - (dark - bias)) / ((flat - bias)  which simplifies to (light - dark) / ((flat - bias)

So in that case the bias has not cancelled out and is in fact being applied multiplicatively, so will be definitely be contributing to noise.

[lensman57]

Correct. Subtracting bias twice has the effect of cancelling the bias out so you have effectively not removed your bias signal, and have, in effect, added noise you wouldn't have if you'd just had the bias subtracted once.

I don't know many experienced imagers who bother with flat darks, most simply use bias instead, as the flat darks are, in normal circumstances short enough to be virtually indistinguishable from bias frames.

A Dark flat maybe necessary if the sub's length is long enough to induce thermal noise, how long is long is anyone's guess. If Dark flats are used then Bias subtracting a Flat is not required as the Bias is already contained in the Dark Flat but the question is then asked if the Bias is  at all required for calibrating the lights if Dark frames are  used I think not as the Bias is already in the Dark frame and is then subtracted from the light. The easiest way is to just use Bias and not bother with Dark Flats at all.

A.G

[ollypenrice]

You'd need bizarrely long flats to need more than a bias to calibrate them, I think. (Use a cleaner T shirt!!  )

Olly

[badger111]

just a thought, but I think most of the calibration steps (certainly darks and bias) have to be done before debayering, but if you were doing it in PS then you're presumably debayered already ?

But yeah, I agree with the others, let DSS take the strain !

In answer to your question, I had to look up what debayering meant! I found a really good document by a Craig Stark, PhD entitled "Debayering Demystified" - It filled a lot of gaps in for me - Thanks

[badger111]

You'd need bizarrely long flats to need more than a bias to calibrate them, I think. (Use a cleaner T shirt!!  )

Olly

I took the flats after focusing the camera using the Frame and Focus in Backyard EOS and then used one of my laptops to produce the flat field by firing up notepad and putting the camera up against the screen, covering it all up with a blanket (probably a fire risk in itself) and taking them with the camera in Av mode at the same ISO...