#1  
Old 09-10-2019, 08:29 PM
Martin Pugh
Registered User

Martin Pugh is offline
 
Join Date: Nov 2006
Location: Canberra, Australia
Posts: 1,339
OSC - Image calibration method

Hi folks.
This switch to OSC cameras has been rather bumpy so I am hoping some long term users can answer my questions:
Darks, Bias and Flats - do you:
a. debayer all of the individual files, then stack to create your masters to then calibrate a light frame, then debayer the light frame or
b. stack the raw images into master calibration frames that contains the bayer matrix, calibrate a light frame, then debayer the light frame

or some other approach?

Next:
It appears rather critical that your bias/darks/flats are taken at the same gain, offset and temp as your light frames.

So while I can gauge the gain level from the specification charts supplied with the camera (depending on what read noise, dynamic range and full well I am aiming for), I guess I am completely out of luck when it comes to deciding what my offset value should be.

certainly looking forward to some insight into this lot.

cheers
Martin
Reply With Quote
  #2  
Old 09-10-2019, 09:35 PM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
Image calibration is largely the same whether you're working with OSC or mono. If you think of a OSC as just a mono camera then it makes it a lot easier

It is best if all calibration is done without debayering. A Master Dark is just stacked raw subs like you would with a mono.
Making a Master Flat is the same, calibrate your flats and then stack without any debayering.

So, debayering is the step between calibrating your lights and registering.
- Calibrate your Lights
- Debayer
- Register
- Stack

As for your second question, calibration frames should always be as close to the lights as possible. Having the same gain, offset and temp is the same regardless of whether you have a mono or OSC, CCD or CMOS.

The Offset is basically your bias level so as long as you're getting bias frames somewhere between 200-1000 ADU you're doing good.

What camera are you using?
Reply With Quote
  #3  
Old 09-10-2019, 09:35 PM
LewisM's Avatar
LewisM
Novichok test rabbit

LewisM is offline
 
Join Date: Aug 2012
Location: Somewhere in the cosmos...
Posts: 10,388
Using PI? Or CCDStack? Or?

In PI, life is a LOT easier to use BatchPreProcess honestly, as it does everything for you. Just remember to tick CFA images

I know how to do it step by step in PI, but its a headache, and I honestly - with my data anyway (which doesn't say much) - see any difference. Rick Stevens will know better.
Reply With Quote
  #4  
Old 09-10-2019, 09:57 PM
Martin Pugh
Registered User

Martin Pugh is offline
 
Join Date: Nov 2006
Location: Canberra, Australia
Posts: 1,339
Thanks for the replies. Appreciate that.

I am using a QHY600C, on an f2.2 36cm RASA....what a challenge.

Cheers
Martin
Reply With Quote
  #5  
Old 10-10-2019, 05:49 AM
The_bluester's Avatar
The_bluester (Paul)
Registered User

The_bluester is offline
 
Join Date: Feb 2011
Location: Kilmore, Australia
Posts: 3,342
I don't know about the IMX455, but for my ASI294 (Using the IMX294 sensor) ZWO have at least at one time, specifically recommended not to use bias frames.

Using APP, the cal frames are debayered, if you do them undebayered and feed it lights to be debayered it will error out and tell you that the light and cal frames don't match, one channel in the cal frames and three for the lights.

Definitely your darks will need to be at he same gain and temp, flats I am not so sure it is important, bias I think would need to be if you are going to use them.

What sort of calibration problems are you having? If it is weird colour casts in different parts of the image it might be similar to the tricks in calibration for the 294, flats (And dark flats) need to be longer than 4 seconds for that camera, they are a SOC design (System On a Chip) and under about 4 seconds exposure (I don't know the precise figure, I aim for 8 seconds for flats just in case) the exposure is controlled by the SOC where longer ones are controlled externally to it by the circuitry of ZWO design. That means flats shorter than the threshold use different circuitry and the noise profile is different, so short flats don't work properly with long duration lights.
Reply With Quote
  #6  
Old 10-10-2019, 08:41 AM
Nikolas's Avatar
Nikolas (Nik)
Dazed and confused

Nikolas is offline
 
Join Date: May 2012
Location: Melbourne
Posts: 3,253
In DSS I just throw in all the files lights darks etc, obviously allocated as such and make sure the debayer settings are rggb
DSS does the rest
Reply With Quote
  #7  
Old 10-10-2019, 06:09 PM
ChrisV's Avatar
ChrisV (Chris)
Registered User

ChrisV is offline
 
Join Date: Aug 2015
Location: Sydney
Posts: 1,737
Wow. Someone's got one of these beasties.

I'm using PI. With my colour cmos I only use darks, and flat darks for my flats. I don't use bias frames. I found that that there was clipping when the bias was also subtracted - this seems to be an issue with the low read noise cameras
Reply With Quote
  #8  
Old 12-10-2019, 07:19 AM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
A long time ago I imaged with a Canon 20D and later an SBIG STL11000 one shot colour.

Firstly CCDstack tends not to handle colour images well. So keep that in mind.

Images Plus is a very refined program for one shot colour. I would be inclined to use that again.

Greg.
Reply With Quote
  #9  
Old 12-10-2019, 12:04 PM
h0ughy's Avatar
h0ughy (David)
Moderator

h0ughy is offline
 
Join Date: Apr 2005
Location: NEWCASTLE NSW Australia
Posts: 33,156
Quote:
Originally Posted by Nikolas View Post
In DSS I just throw in all the files lights darks etc, obviously allocated as such and make sure the debayer settings are rggb
DSS does the rest
I do the same however the tiff file gets processed with pixinsight afterwards. I use a asi071 and a RASA 11, soon to be a asi294 and RASA 8 in the doghouse as the focus can now be automated as there are no focus locks on the 8.

I always try to make the right hand side curves on top of each other to get the colour right. But im no expert just something Mr trimarchi showed me many many years ago.
Reply With Quote
  #10  
Old 13-10-2019, 09:16 AM
Martin Pugh
Registered User

Martin Pugh is offline
 
Join Date: Nov 2006
Location: Canberra, Australia
Posts: 1,339
thanks again for the replies.
i will certainly take a look at Images Plus.

I have not yet been able to measure the linearity of the camera so as a test I took flats at 10K, 20K, 30K, and 40K. None of them corrected the very slight vignetting and all appeared to overcorrect i.e the corners became bright.

The start up drivers have been awful for the 600 but they have moved fast to fix those issues but I am overseas at present doing astro processing lectures so cannot resume. The QHY600 has 3 readout modes necessitating at least 3 different bias/gain/offset values making things more complicated.

Cheers
Martin
Reply With Quote
  #11  
Old 13-10-2019, 09:34 AM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
What exposure were you using for the flats?
Did you use darks for the flats or a bias?

I’ve found that the best way to do flats with a CMOS is to NEVER use bias but have darks for the lights and darks for the flats. It does mean you want to have the same exposure for both darks and flats as best as possible though.

Over correction on flats in my experience is usually from poor flat calibration or trying to correct for too much vignetting. If the vignetting is too strong it appears as overly bright corners, at times not because it’s actually too bright but because of the lack of data and the artificial brightening.

I noticed this when using a DSLR on the RH200, the mirror box within the Nikon caused heavy vignetting that made the outer portions of the frame uncorrectable via flats.
Reply With Quote
  #12  
Old 13-10-2019, 10:08 AM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
I don't know if this applies to the QHY600 but the Sony A7riv (same sensor)
has a dual gain analogue to digital converter as do the earlier A7rii and iii cameras.

At ISO320 the read noise drops from this dual gain converter.

So if that is the same for the QHY600 (ADC is on the sensor back face so it probably is the same) then you may need a set of flats for before ISO320 and after ISO320. The bias would be different.

I don't think the QHY600 uses ISO but I wonder how that works with the ADC probably programmed by Sony to include that dual gain?

Perhaps this is not relevant but worth mentioning in case it is. I suppose you would need to know what gain setting that ISO320 threshold is.

As far as flats go I think they are part science and part Art! I tend to play with my calibration files to make sure they are working before I finish an imaging run. I don't bias subtract my flats when making a master flat. I list it separately when doing the dark subtract etc. I also tend to use adaptive darks and use the bias for them as well.

Overcorrecting flats in my experience is from overexposed flats. It can also be traced to a mismatch of bias (not fresh enough) or slightly off darks.

What can complicate things are the different readout rates and resultant different read noise available with some cameras (notably FLI, they offer a fast readout for focusing but higher read noise and a slower download but lower noise). If you mix them up flats won't work well.

These days with most scopes having lensed correctors/flatteners etc there are often bright centre spots and increased vignetting which can be a challenge to correct.

Greg.
Reply With Quote
  #13  
Old 13-10-2019, 10:56 AM
Martin Pugh
Registered User

Martin Pugh is offline
 
Join Date: Nov 2006
Location: Canberra, Australia
Posts: 1,339
I used CCDAP so the exposure was automated and I did calibrate the flats using a bias so I can try eliminate that and switch to darks only.

The issue with an f2.2 scope and a back illuminated sensor is that it has to get really quite dark to avoid saturating the sensor, even with a minimum flat exposure length of 1sec. Consequently, flat fields are littered with trailed stars. Nonetheless, the combination algorithim gets rid of them and the flats ‘look’ really good, but are not currently working as I described so work to do and an EL panel is on the way.

The QHY600 does not use ISO.
Cheers
Martin
Reply With Quote
  #14  
Old 13-10-2019, 12:24 PM
The_bluester's Avatar
The_bluester (Paul)
Registered User

The_bluester is offline
 
Join Date: Feb 2011
Location: Kilmore, Australia
Posts: 3,342
I have always done my flats as flats of about 8 seconds duration, and dark-flats shot in the same session with all the same settings and exposure time as the flats. I use them in APP to make a master flat then only use the master flat.
Reply With Quote
  #15  
Old 15-10-2019, 11:24 AM
sil's Avatar
sil (Steve)
Not even a speck of dust

sil is offline
 
Join Date: Jun 2012
Location: Canberra
Posts: 1,474
I use the process outlined in Light Vortex Tutorial on Preprocessing when I'm putting in lots of effort to a dataset. Typically I often use APP for a quick image, feeding it the sub sets I have and it does what it can with what I provide (RAWS from DSLR usually). Its usually my way to preview a sessions data by running it through APP first to decide if I want to spend the time in PixInsight to try to wring every bit of signal I can out of the data. Its an evolving process.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 09:40 AM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Testar
Advertisement
Bintel
Advertisement