Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Firecapture Teething Problems

  • Please log in to reply
12 replies to this topic

#1 kevinbreen

kevinbreen

    Vostok 1

  • -----
  • topic starter
  • Posts: 136
  • Joined: 01 Mar 2017

Posted 19 June 2017 - 12:15 PM

I recently upgraded to ZWO ASI224MC and Firecapture, having previously used NexImage5 and iCap. I'm getting poor results, which are evident when I try to post process the stacked TIF files. 

 

I have found that if I set the histogram too high at capture (say, 80%) then I've only got 20% to play around with in post processing (Registax). Am I right in thinking that the primary reason in keeping the histogram at a chosen level (let's say 55 - 60% as per Kokatha man's suggestion) is to allow for enough elbow room in post processing so that the wavelet adjustments, gamma adjustments etc will still result in a final histogram that is less than 100%? Any final histogram over 100% is undesirable, right? This modus operandi served me well with the NexImage 5 in iCap.

I know that more frames captured means a better final picture (as there are more to choose the best 40% from, say). This is why I upgraded to the ZWO ASI224MC.

My way of thinking is this -
The histogram is a function of fps and gain. The higher the fps the better, clearly. Once I've set the fps, I set the gain next, but not so that it means exceeding the preferred % histogram. Would you agree that this is the right methodology?

My settings are as follows
For my Jupiter captures in Firecapture, Red is 54 and Blue is 96. Brightness is down at zero, gamma is set to 50 (OFF). Autoexposure is OFF. HighSpeed is OFF(for now, anyway). HardwareBin is OFF, as is AutoHist. Are these the settings that you super-imaging folks use?

..... is there something I'm overlooking?

Post Processing:
I open the avi in Autostakkert!2.
I analyse it.
Quality Estimator is set to Local. Noise Robust is set to 6. Gradient for Jupiter.
I set align points, about 30. On the planet, not on the edges.
I stack it.

Question: Should I locate the bar to the highest quality location (the gray graph, not the green one) in the quality graph first (so at the highest quality frame, right?) and set my align points there?

I open the TIF in Registax.
Wavelet Scheme is set to Linear
Wavelet Filter is set to Gaussian

I run the histogram
The Red, Green and Blue frequency lines extend beyond the 100% mark. But hang on! I captured with a 50% histogram! What's going on?
The planet's image looks way too exposed. Now, obviously, any trimming of the histogram makes it even worse.
There is no clearly defined left hand side bump in the histogram which depicts the abundance of pixels which captured the black background. There was when I used NI5 and iCap.
OK, in the  RGB Balance window I can pull the red, green and blue lines back into the histogram.... but why are they beyond 100% in the first place?

Another issue - Processing Area
In Registax, under Settings, I can adjust the size of the processing area, because who wants to process a whole expanse of black sky when you can adjust to the processing area to snugly fit around the planet, right? No matter what size, I set the processing area to, the four right-angled corners are always in the corners of the window. Help!

Other factors:
I collimated the scope a week ago. I got it spot on.
I now have a motorized focuser. My focussing is better now than before.
I can thus rule out collimation and focussing as contributing factors.

I don't do anything else in the right hand window.

No point in going near wavelets now as they're just going to blow out the planet even worse....
I'm sure the ASI224 is way better than the NI5, the proof is in the pictures all over CN.
Here's the result - it's pretty bad.
If you have any pointers I'd be extremely grateful.

Many thanks folks

K

Attached Files



#2 RedLionNJ

RedLionNJ

    Surveyor 1

  • -----
  • Posts: 1672
  • Joined: 29 Dec 2009
  • Loc: Red Lion, NJ, USA

Posted 19 June 2017 - 12:59 PM

There are many, many ways to skin a cat.

 

When seeing is good, you can use a longer exposure (smaller fps), lower gain and a higher peak histogram. When seeing is mediocre, increase the fps to the max you can get, even if it means dropping the histogram down to 50-60% (even less on Saturn).

 

I've never explicitly adjusted the processing area in Registax. But unless i click on 'Do all', it generally processes a small square within the image. This can be moved by clicking on any part of the image.

 

Is it possible you have autostretch enabled in Registax? This could stretch the histogram to the full width. Subsequent application of wavelets would then push it too wide.

 

In general - seeing is going to be the #1 factor in any form of high-resolution imaging. It can vary wildly from night to night, hour to hour or even minute to minute.  In your image, you've definitely deconvolved far more than the original data allows.

 

Other questions:-  focal ratio?  Are you capturing in RAW?

 

In AS!2 (try AS!3, btw), I use about 15-20 APs on Jupiter, well away from the edges. I first set the APs, then do the quality estimate and my percentage then totally depends on the appearance of the resulting quality graph.



#3 kevinbreen

kevinbreen

    Vostok 1

  • -----
  • topic starter
  • Posts: 136
  • Joined: 01 Mar 2017

Posted 19 June 2017 - 01:40 PM

There are many, many ways to skin a cat.

 

When seeing is good, you can use a longer exposure (smaller fps), lower gain and a higher peak histogram. When seeing is mediocre, increase the fps to the max you can get, even if it means dropping the histogram down to 50-60% (even less on Saturn).

 

 

 

I've never explicitly adjusted the processing area in Registax. But unless i click on 'Do all', it generally processes a small square within the image. This can be moved by clicking on any part of the image.

 

Is it possible you have autostretch enabled in Registax? This could stretch the histogram to the full width. Subsequent application of wavelets would then push it too wide.

 

In general - seeing is going to be the #1 factor in any form of high-resolution imaging. It can vary wildly from night to night, hour to hour or even minute to minute.  In your image, you've definitely deconvolved far more than the original data allows.

 

Other questions:-  focal ratio?  Are you capturing in RAW?

 

In AS!2 (try AS!3, btw), I use about 15-20 APs on Jupiter, well away from the edges. I first set the APs, then do the quality estimate and my percentage then totally depends on the appearance of the resulting quality graph.

Thanks for getting back to me.

There's no autostretch as far as I can see in Registax. But the "direct" checkbox beside the "stretch" button in the histogram window was checked. Should it be checked or should it be unchecked?

 

Focal Ratio:

I'm using an 8SE (focal ratio is 10). ASI224MC has 3.7 micron pixels. Using the rule of thumb that 5 (to 7) x pixel size in microns = optimal focal length, it works out at about f20. So I use a 2x Barlow to get close to f20.

 

Capturing in RAW?

emmmm, I'm not sure what you mean. To me, RAW is an image format that I can select on my DSLR. I'm using Firecapture, I'm capturing in AVI format. ASI224MC is a colour camera, so I capture in colour. Here is where I'm confused with the whole FILTER and DEBAYER thing....

 

If I want to capture in colour, what filter should I choose?

If I want to capture in black and white what filter should I choose?

If I want to capture in IR, I should use the IR filter, right?

Am I totally wrong here?

 

I wasn't aware of your method of first setting the APs and then running the quality estimate. I'll give that a try. In the meantime if you could shed some light on the above issues I'd love to buy you a beer. Thanks!



#4 APshooter

APshooter

    Mercury-Atlas

  • -----
  • Posts: 2570
  • Joined: 14 Oct 2013
  • Loc: Camby, In.

Posted 19 June 2017 - 03:31 PM

 

If I want to capture in colour, what filter should I choose?

None.  Just make sure debayer is not selected during capture.  AS2! will automatically put the color back in when you stack the files.



#5 kevinbreen

kevinbreen

    Vostok 1

  • -----
  • topic starter
  • Posts: 136
  • Joined: 01 Mar 2017

Posted 19 June 2017 - 03:40 PM

Thanks APshooter. What's the advantage in doing this over capturing in colour, apart from the smaller file size?

#6 APshooter

APshooter

    Mercury-Atlas

  • -----
  • Posts: 2570
  • Joined: 14 Oct 2013
  • Loc: Camby, In.

Posted 19 June 2017 - 04:13 PM

 

Thanks APshooter. What's the advantage in doing this over capturing in colour, apart from the smaller file size?

If you don't select debayer and make the image 'monochrome' before capturing, you will lose all your color data.  My best Jupiter shot EVER was ruined by capturing with debayer checked.  I've never seen conditions that good again....and the great red spot was centered perfectly too!   The color data is in the file, it just needs AS2! to bring it out.

 

Color is very hand if you're using the ADC and Firecapture.  Use debayer and click on the ADC align feature.  I always enable autoalign as well, to keep the planet in view while working the prisms.

 

https://www.youtube....h?v=DJ1SpbCmM9U

 

https://www.youtube....h?v=mLDduTLPMJU

 

https://www.youtube...._aMmO4DBA&t=75s

 

https://www.youtube....h?v=5hgx7Lb16EU


Edited by APshooter, 19 June 2017 - 04:17 PM.


#7 kevinbreen

kevinbreen

    Vostok 1

  • -----
  • topic starter
  • Posts: 136
  • Joined: 01 Mar 2017

Posted 19 June 2017 - 05:42 PM

Hi, you said "If you don't select debayer and make the image 'monochrome' before capturing, you will lose all your color data"

but if I do select debayer I get colour anyway...ok, I think I get it - I'm actually not using an ADC so there's no advantage to shooting in colour. Is that what you mean?



Let's suppose I DO want to shoot in colour. You said to select no filter, but there's L, B, G, R, CH4 etc etc. So where do I select no filter?


Sorry if I'm asking inane questions - it's difficult to catch nuances in written posts. Thanks, Kev

#8 kevinbreen

kevinbreen

    Vostok 1

  • -----
  • topic starter
  • Posts: 136
  • Joined: 01 Mar 2017

Posted 19 June 2017 - 06:12 PM

Last night I shot an avi with debayer not selected and yes, the image was monochrome before capturing, and I got the colour back in AS!2, like I did with avis I took with NI5 and iCap.
I also shot an avi last night with debayer selected and it was automatically in colour.


I don't understand how you lost all your colour data as you outlined above? In my understanding, whether debayer was selected or not, you would have got your colour anyway.

What am I missing here?

#9 APshooter

APshooter

    Mercury-Atlas

  • -----
  • Posts: 2570
  • Joined: 14 Oct 2013
  • Loc: Camby, In.

Posted 19 June 2017 - 06:23 PM

Hey Kev.

 

I think I confused things.  To see color on the screen to aid in focusing or with ADC setup, you have to select debayer. 

 

But you do not want to capture debayered images.  The image files will be 3x as large.  If you want to record in color, uncheck debayer.  The image on the screen will appear B&W.  But it will process in color.

 

If you record with debayer on, the image on the screen will appear in color, but process B&W.  Hence my saying you will lose color data.  I have not found a way to retrieve the color data with debayer on.

 

Edit:  Not sure how you got the color with debayer on.  On my Jupiter file it recorded in color, but would never process in color even when forced in AS2!  I may have set something incorrectly on that particular outing.

 

Here's a link to a similar discussion:  https://www.cloudyni...new-asi-224-mc/

 

And:  https://stargazerslo...866-debayering/

 

 

 

Here's a link on the advantages of the ADC:

 

http://skyinspector....-corrector--adc


Edited by APshooter, 19 June 2017 - 06:31 PM.


#10 kevinbreen

kevinbreen

    Vostok 1

  • -----
  • topic starter
  • Posts: 136
  • Joined: 01 Mar 2017

Posted 19 June 2017 - 06:57 PM

Great, thanks for clearing that up, my mind was in knots!
Finally, how does one select "no filter"..?

#11 APshooter

APshooter

    Mercury-Atlas

  • -----
  • Posts: 2570
  • Joined: 14 Oct 2013
  • Loc: Camby, In.

Posted 19 June 2017 - 07:35 PM

I'd just select Lum...since it's clear. 



#12 Kokatha man

Kokatha man

    Cosmos

  • *****
  • Posts: 9778
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 19 June 2017 - 08:13 PM

Briefly bigshock.gif  - go to "Settings" at the bottom of the control window in FC & select "Debayer" & check (tick) both boxes..."Debayer & capture color data" and "Always record undebayered (RAW) but preview in color"

 

This way you will see the onscreen image in colour but as soon as you hit "capture" the image will turn into a mono image during the capture sequence...this saves a lot of disk space for starters & as Torsten has said repeatedly he advises users of FC to debayer in other programs later. (eg, AS!)

 

For the other 2 windows therein, you don't worry about the "Debayer algorithm" since you will be leaving that for another post-capture program...& choosing the correct Bayer pattern from the dropdown there is pretty obvious - unless you desire some bizarre colour-rendering of a planet! lol.gif

 

If we're employing the colour camera we gauge the histogram (50% - 60%) by using the mono image onscreen by un-ticking the Debayer box on the left-hand side of the large FC image view window (camera window)...this is the little matrix array or r,g & b squares which should be by default about halfway down that side...from when you originally loaded FC. We have of course already ticked the boxes in that "Settings - Debayer" window I speak of at the start of this post. wink.gif

 

Un-ticking this little coloured icon on the main camera/planetary view window will turn the onscreen image mono for you to adjust your histogram...although you can also just use the green histo of the colour image to set this value. (it's the same)

 

Just make sure you tick this again once you've set the histo if using the mono image before hitting "capture."

 

Those 2 boxes you check in "Debayer" under "Settings" in the smaller FC control window will remain that way for any future operations of FC unless you've fiddled with them in the meantime... wink.gif

 

Picking "L" just gives the files a handle for you...obviously if you're using a colour camera it is the most sensible label for remembering what they are! (ie, colour captures - even though if you play any of these avi's before they're debayered they look like mono recordings)

 

On a related issue re your other thread (even I get tired of typing at times rofl2.gif ) you do "want some headroom" as you put it in your histograms you set - or do your initial processing in Registax apropos that thread - this is one reason for not setting the histogram too high in capture or stretching it too much in preliminary processing...& in capture especially it is also very easy to "blow out" the EZ areas of Jupiter (or Saturn) with too high an exposure/histogram. wink.gif

 



#13 kevinbreen

kevinbreen

    Vostok 1

  • -----
  • topic starter
  • Posts: 136
  • Joined: 01 Mar 2017

Posted 20 June 2017 - 06:07 AM

Thanks Kokatha man, much appreciated


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.







Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics