Jump to content

Stargazers Lounge Uses Cookies

Like most websites, SGL uses cookies in order to deliver a secure, personalised service, to provide social media functions and to analyse our traffic. Continued use of SGL indicates your acceptance of our cookie policy.

  • Announcements

    sgl_imaging_challenge_banner_solar.jpg

     

Datalord

So close to ordering, but then...

Recommended Posts

Datalord    19

Ok, there was a patch of clear outside late evening last night, so I set it all up and went out to figure it all out. Before getting to the questions, I'll just mention that the weather trolls are very strong in this country. I managed to get about an hour to one and a half of playtime before it was all white again.

Anyways, my initial test was just with my 6D attached to the RASA. Three things occured:

1. Very clear vignetting of the image. This surprised me quite a bit considering the claim it has a huge imaging circle.

2. Which lead me to think "flats", which led me to borderline panic. My usual flats are a simple piece of paper in front of the WO, how on earth am I going to do flats with this monster?

3. Due to the clouds trolling, I didn't get it properly PA'ed. Even so, I was surprised to find star streaks on 30s exposures. I thought I took my time to balance the setup, so is there anything else I have to look for? Or is it just the PA?

With all that said, the scope clearly has immense lightgathering ability. Looking forward till tonight, looks like I might have a chance to tinker more with it.

Share this post


Link to post
Share on other sites
Datalord    19

Got some more time tonight and the camera is ticking away right now. Just to give an idea of the vignetting I'm talking about, this is what it looks like:

5990d4a201b43_LIGHT_30s_400iso_21c_20170814-00h01m14s578ms.thumb.JPG.b7e751e1174888a83fdb241131774054.JPG

I really need ideas on how to make flat frames when I can't cover the scope. What to do?

Also, I managed to get the PA done a little better. Still have some work to do, but it seems like I can do 30s frames now.

Share this post


Link to post
Share on other sites
Davey-T    8,799

You can make a light box to mount on the scope or I think Pixinsight can do some sort of artificial flat, don't know for sure because I haven't used it.

Dave

 

Share this post


Link to post
Share on other sites
Datalord    19

First light declared. Even without flat calibration, I like it. Just 31 minutes of 30s frames and a bit of messing with colours.

First_light_de-noised-layeredAndcoloured-scaledDown.thumb.jpg.dcab99a97bea377abd45b0712fd5ee69.jpg

And 2 hours on M31, still no flats:

RASA_-_First_process-gimped.thumb.jpg.0075a92ec6dd0b3af040f0ca8c0dba59.jpg

Edited by Datalord
  • Like 1

Share this post


Link to post
Share on other sites
AKB    305

Looks like a good start... time to change the list of gear in your signature!

  • Like 1

Share this post


Link to post
Share on other sites
Datalord    19

So, I just waited till morning and pointed the scope up and took the flats. A bit of a challenge, to be fair, this is SO fast I had to take the flats at 1/4000.

RASA_-_second_process-gimped.thumb.jpg.fbdc363343e0f641027f887af311a435.jpg

I'm quite happy with this on my first night. Looking forward to getting my cooled camera soon.

Edited by Datalord

Share this post


Link to post
Share on other sites
AKB    305

What a difference a flat makes! (plus other processing.)

I've only just started using my Hyperstar and really it's the same flats challenge.  I use 1/1000 seconds, because that's as fast as my CCD camera goes.  Good luck with the cooled camera – I have yet to take that step.

Share this post


Link to post
Share on other sites
ollypenrice    16,505

I think the flats have damaged the image. The M31 background was much flatter without them.

Getting good flats is easier in theory than in practice and, for no reason that I have ever identified, I sometimes get bad ones which I can't use.

The image you posted to show the vignetting shows it beautifully and your flats should look like a paler version of this, minus the stars. If they don't they're not correct. Could you post a stretched JPEG of a flat? It should just have a brighter centre with darker corners and whatever dust bunnies have been created. A typical flat for me on our most vignetted rig looks like this:

O%20FLAT%20web-M.jpg

The stretch visually exaggerates the light fall off, which is really from around 19000 ADU in the corners to 23000 in the centre. Images come out fine.

I'm not an expert on this but I think there might be a danger, in very short flat exposures, of not getting up into the real linearity range of the camera. If you could contrive a dimmer light source it might be a big help. In CCD imaging with a shutter there is also the problem of the flat suffering from 'shutter wipe' in short exposures. How or if this works with a DSLR I don't know.

Flats also need to be dark-frame subtracted. Because the exposures are so short you don't need 'real' darks, though. Just give a master bias to your stacking software as a 'dark flat.' It's all you need.

I would never throw my raw flats into the same stacking/calibrating operation as my light frames. Always make a master flat first and apply that. If you don't you can't see what it looks like. The same with bias and darks. Make masters.

Olly

Share this post


Link to post
Share on other sites
Datalord    19

How do you see the background is damaged? I have no doubt it is, I just don't know how to see it.

Like I said, I had problems getting the flat dark. This is a shot at 1/4000 I used. Haven't stretched it, though:

IMG_5393.thumb.JPG.3447de9fb39058e74607265bea3323aa.JPG

On a practical note, how do I make the master flat with the dark subtracted? I thought DSS handled all that.

Share this post


Link to post
Share on other sites
AKB    305

The flat you posted hides the extent of the vignetting, partly because of no stretching, and partly because it has (naturally) non-aligned R,G, and B histograms.  Splitting the channels and stretching a bit shows this clearly.

lrgb-flat-stretch.thumb.jpg.05e4739c869134b6b4a33363e58c0169.jpg

...and this is just with a small stretch.  The stretching you normally need to do for images is massive, with tiny amplitude differences being exaggerated, so it will be much worse.

You can make a master flat simply by processing it as a separate image using your darks, and remembering not to apply that again when using it on actual images.

 

Share this post


Link to post
Share on other sites
steppenwolf    3,927

I'd be tempted to simplify this a little by not concerning yourself with a Dark for the Flats - I tend to only consider a Dark for the Flats if they are long (say 15 seconds or more) - at 1/4000 sec I really wouldn't bother. However, you should Bias subtract your Flats.

  • Like 1

Share this post


Link to post
Share on other sites
AKB    305
37 minutes ago, steppenwolf said:

you should Bias subtract your Flats.

Indeed.  Said "dark" meant "bias".  Apologies.

 

Share this post


Link to post
Share on other sites
ollypenrice    16,505
2 hours ago, Datalord said:

How do you see the background is damaged? I have no doubt it is, I just don't know how to see it.

Like I said, I had problems getting the flat dark. This is a shot at 1/4000 I used. Haven't stretched it, though:

IMG_5393.thumb.JPG.3447de9fb39058e74607265bea3323aa.JPG

On a practical note, how do I make the master flat with the dark subtracted? I thought DSS handled all that.

There are two bright magenta triangular patches here.

datalord.thumb.JPG.2e184e53b0649135a808e8c252a9616c.JPG

Your flat looks OK so I don't know what's going on here.

One thing to do is balance your three colours' histograms. If you have Ps try to get the top left of each histo peak to the same point in Levels by moving the black point in. Your background is extremely magenta here, suggesting green is weak. Ps also has a colour sampler tool which you can set to 3x3 or 5x5 average and it will tell you the colour brightness in each channel. I like them to be equal, some go for a higher value in blue.

DSS and other programmes will do the entire data set in one pass (lights, darks for lights, flats, bias as dark for flats) but the trouble with this is that when it throws up a problem it's hard to isolate the source. Given that DSS is very slow anyway I would advise working from master darks, master flats and master bias so you can try with and without your flats, darks, whatever, to find out what is doing what.

What I would really recommend is ditching DSS entirely and buying AstroArt 6!

Olly

 

Edited by ollypenrice
  • Thanks 1

Share this post


Link to post
Share on other sites
Datalord    19

I know I need to find a better software package, I just don't know which one to get. Some say AA, some say PI, some do just PS and then there is Registar etc. And all of them have a considerable learning curve. 

Share this post


Link to post
Share on other sites
ollypenrice    16,505

PI can do the jobs I do in AstroArt (which I use only for calibrating and cosmtic correction as an alternative to DSS.) However, I find PI obscure and complicated and I don't get on with it. I do run every colour image I do through it as a first step, though. (I use DBE to remove colour gradients and SCNR green to kill green noise.) After that I head for civilization, Photoshop CS3!

The PI and Ps approaches are radically different in pursuit of the same goals. We all want to process our images in zones, each zone wanting a different process. (Dark parts may need noise reduction, bright parts but not stars might want sharpening, stars want slight stretching, faint parts want hard stretching.)

In PI you need to create masks to separate the zones. In Ps you use layers, modifying the bottom layer and erasing the top one where you like the bottom one better. For me, it is way easier to use the layers approach because you have real time views of what your eraser is doing. However, it is against PI's religion to touch an image with your hands. You must only touch it through numbers. (I think the authors avoid walking on cracks in the pavement as well...)

:eek:lly

  • Like 1

Share this post


Link to post
Share on other sites
steppenwolf    3,927
20 minutes ago, ollypenrice said:

What I would really recommend is ditching DSS entirely and buying AstroArt 6!

I won't add my personal choices in here as we all have different ideas based on personal experiences and I don't want to cloud the issues by piling more options on you but I would echo what Olly says about DSS and his software suggestion ain't a bad one! However, something to try with DSS is to separate the alignment phase from the stacking phase as this will give you a greater amount of control over the whole process while you consider what software route you would like to take in the future. Whatever route you do take, make sure that there are many users using the software so that you can dip into a large knowledge pool.

  • Like 1

Share this post


Link to post
Share on other sites
Datalord    19
11 hours ago, AKB said:

The flat you posted hides the extent of the vignetting, partly because of no stretching, and partly because it has (naturally) non-aligned R,G, and B histograms.  Splitting the channels and stretching a bit shows this clearly.

lrgb-flat-stretch.thumb.jpg.05e4739c869134b6b4a33363e58c0169.jpg

...and this is just with a small stretch.  The stretching you normally need to do for images is massive, with tiny amplitude differences being exaggerated, so it will be much worse.

You can make a master flat simply by processing it as a separate image using your darks, and remembering not to apply that again when using it on actual images.

 

I really wish I could comprehend how to put this into something actionable. What tool(s) and what do I do in those tools? I'm quite willing to read about it, if I just knew what to search for.

Share this post


Link to post
Share on other sites
ollypenrice    16,505
11 hours ago, Datalord said:

I really wish I could comprehend how to put this into something actionable. What tool(s) and what do I do in those tools? I'm quite willing to read about it, if I just knew what to search for.

I would assume that any differences between colour channels on an OSC will arise from the underlying chip's different sensitivity to different wavelengths. I don't think this is a problem?

To make masters, just lie to your software programme.

Put all your bias in as lights and average combine them without alignment or any other filtration like hot pixel correction. Call the this Master Bias and make a copy called Flat Darks. (This is just for your sanity's sake. They are actually the same thing.)

Put all your flats in as lights and your Flat Dark master in as dark for lights and average combine them, again without alignment or any filtration. Call this Master Flat.

Darks (if you want to use them, but they might do more harm than good.) Put all your darks in as lights and average combine them without alignment or filtration. Call this Master Dark.

Olly

Edited by ollypenrice
  • Thanks 1

Share this post


Link to post
Share on other sites
ollypenrice    16,505
10 minutes ago, Datalord said:

so simple. Thanks!

No one thing in AP is particularly difficult. It's just that there is a vast and sprawling web of interconnected not-very-difficult things to do. :Envy:

Olly

Share this post


Link to post
Share on other sites
Datalord    19
9 minutes ago, ollypenrice said:

No one thing in AP is particularly difficult. It's just that there is a vast and sprawling web of interconnected not-very-difficult things to do. :Envy:

Olly

Yes. And you learn all of them only by failing over and over again. Yet, this is the most rewarding hobby I've ever had.

Share this post


Link to post
Share on other sites
Datalord    19
13 hours ago, ollypenrice said:

I would assume that any differences between colour channels on an OSC will arise from the underlying chip's different sensitivity to different wavelengths. I don't think this is a problem?

To make masters, just lie to your software programme.

Put all your bias in as lights and average combine them without alignment or any other filtration like hot pixel correction. Call the this Master Bias and make a copy called Flat Darks. (This is just for your sanity's sake. They are actually the same thing.)

Put all your flats in as lights and your Flat Dark master in as dark for lights and average combine them, again without alignment or any filtration. Call this Master Flat.

Darks (if you want to use them, but they might do more harm than good.) Put all your darks in as lights and average combine them without alignment or filtration. Call this Master Dark.

Olly

Well, I tried. With DSS. The master bias went smooth, once I found which boxes to uncheck. However, making the master flat did not happen, because the master bias can only be saved in RGB 16 bit and apparently it expects to have a Gray 16 bit image to stack against. I'm getting this error:

DSS_shit.png.23c3f45abe0e349815e1049fa27a6fa2.png

So, I'm going to have to make the plunge. I have to get a proper software package. 

Stacking: AstroArt 6

Colour fun: Photoshop.

I'm solid with those two, once I learn how to use them?

Edit: downloaded AA6 demo. Stack went ok, any kind of further processing ended in total mayhem. Will watch tutorials...

Edited by Datalord

Share this post


Link to post
Share on other sites
ollypenrice    16,505
10 hours ago, Datalord said:

Well, I tried. With DSS. The master bias went smooth, once I found which boxes to uncheck. However, making the master flat did not happen, because the master bias can only be saved in RGB 16 bit and apparently it expects to have a Gray 16 bit image to stack against. I'm getting this error:

DSS_shit.png.23c3f45abe0e349815e1049fa27a6fa2.png

So, I'm going to have to make the plunge. I have to get a proper software package. 

Stacking: AstroArt 6

Colour fun: Photoshop.

I'm solid with those two, once I learn how to use them?

Edit: downloaded AA6 demo. Stack went ok, any kind of further processing ended in total mayhem. Will watch tutorials...

I don't do any real post processing in AA. The 'remove line' tool is great for sat trails and if the hot pixels haven't been dealt with in the stacking (page three of the pre-processing section) then that can be useful too. It now has a 'remove gradient' tool which looks OK but I do that in PI so haven't tried the one in AA.

There are some excellent astro plug-ins for Ps. Noel's Actions (now found on a website called Pro Digital) are used by many imagers. There is also a gradient removal tool called Gradient Xterminator, by Russell Croman, if you struggle with the one in AA. And an excellent freebie is to be had on Rogelio Bernal Andreo's Deep Sky Colors website. His 'Hasta la Vista Green' gets rid of unwanted green bias and noise. It is a 'must' if you don't have PI.

There are some incredibly bad web tutorials out there. Anything by Warren Keller or Adam Block will set you in the right direction.

I would just advise two guiding principles. 1) Don't do things in processing which you don't understand. While experimenting with an intervention think about what it is doing and how it works. 2) Try to hone your skill at looking at an image beause in the end it is the image which will tell you what you need to do. So keep looking at the background sky for flatness and colour and also at the stars. (Background and stars are the hardest parts to get right.)

Olly

  • Like 1

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.

×