Video of me processing, please help

Questions and answers about processing in StarTools and how to accomplish certain tasks.

Video of me processing, please help

Postby rottielover » Mon Jun 26, 2017 5:05 pm

I created this video of my trying to process 3 targets I took recently.

Each target is about 1 hour's worth of exposures. Canon 70D at 1600 ISO using a kit lens EFS 55-250 @250mm on an iOptron tracker (so no dithering)...

Each was stacked with Deep Sky Stacker using lights , darks, flats, bias (30 each on the calibration frames)..

So here is a link to the YouTube live recording I did, I broke these up into 3 video's and I'm uploading those separately so I will post sep. links after the uploads are done.

https://youtu.be/ooqmr1s9_O8
rottielover
 
Posts: 22
Joined: Wed May 24, 2017 2:26 am

Re: Video of me processing, please help

Postby rottielover » Mon Jun 26, 2017 5:28 pm

Here they are split into 3 separate videos

Andromeda video - https://youtu.be/52G9Nu-5aas

Anteres Video - https://youtu.be/IHgKd2ItnPM

North American Nebula Video - https://youtu.be/SAfMP5dZlRk
rottielover
 
Posts: 22
Joined: Wed May 24, 2017 2:26 am

Re: Video of me processing, please help

Postby admin » Tue Jun 27, 2017 1:04 am

Having a look now...

Thanks for this!
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
 
Posts: 1553
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne

Re: Video of me processing, please help

Postby rottielover » Tue Jun 27, 2017 2:47 am

Thank you Ivo,

Here is the FTS files from Deep Sky Stacker, maybe that would help also:

https://www.dropbox.com/sh/bvufqtvqfwfr6kw/AAAjQdXiMXlhqrdMwryLbvA4a?dl=0
rottielover
 
Posts: 22
Joined: Wed May 24, 2017 2:26 am

Re: Video of me processing, please help

Postby admin » Tue Jun 27, 2017 2:55 am

Ok, I went through your Antares video.

To sum things up, you're doing very well. Your data, by the looks of it is great and you're doing very well without a mount. I understand you're getting to grips with the tools and what they do. However even without, perhaps, understanding in-depth what they do, you're guiding them well.

First off, one thing that should define the context within which you use the tools, is that you are processing a wide field. You already mention this.

When processing wide fields there are 3 main things to take into account;
  • Fast undulating gradients
  • Undersampling (vs oversampling for narrow fields)
  • Busy star fields

You're experiencing all 3 in your image.

Going through your video, I can make the following comments;

AutoDev. The first AutoDev is to quickly let you inspect the image for potential problems. Take your time.
Here you look for stacking artifacts (you already do this), try to gauge how good/bad light pollution/gradients are (important for widefields), try to gauge if your data is oversampled/undersampled (this has consequences for binning and decon) and more.

You decide to bin (this indeed helps with noise). You rightfully crop the image (to get rid of stacking artefacts).
The reason why you might do that second AutoDev, is simply because getting rid of stacking artifacts will mean that AutoDev no longer needs to show them - it can now allocate dynamic range to show other "problems". For now, we haven't been using AutoDev at all for the purpose of our final/"real" stretch. We've just been (ab)using its dynamic-range-allocation-skills to try to bring out any glaring problems (e.g. any problems that can be shown to you by stretching the image in a particular way). Problems you can more easily pick up this way are stacking artifacts, gradients/vignetting, dust specks/donuts, etc.

You go onto Wipe, to get rid of the light pollution, skyglow, gradients and any other biases.
You have a tricky problem here; because your data set is a wide field, gradients can be very noticeable and undulate rapidly. Specifically, if there is any light pollution on the horizon this can show up quickly as a fast undulating gradient near one of the corners. You can see an example of that in the bottom/right, where there is a gradient that rapidly falls off. This is a tricky, non-trivial situation to model and subtract.

There are varying solutions, each with their own pros and cons.
The easiest is to simply recognise that it is a real light source that has invaded your faithful recording and to simply crop it out.
The second easiest solution is to bump up the aggressiveness parameter in Wipe, or bump up the aggressiveness near the corners (the Vignetting preset does that for example).
More elaborate solutions involve masking (which some AP'ers consider cheating or "painting") to combine different Wiped images together.
Much depends on what your preferred way of conducting AP is and who you are trying to impress.

You go on to use the RoI feature of AutoDev effectively, making a valid consideration that you wish to highlight that particular nebulosity as the center piece of the image. Rather than as a problem visualisation tool, you are now using AutoDev to perform the final stretch. You can now do that because you are now satisfied that problems have now been dealt with; you are now certain that AutoDev will only stretch signal.

On to decon now.
It is important to note that the binning will have made the data undersampled; multiple units of detail have now been crammed into a single pixel. This is the opposite of being oversampled, where a single unit of detail is "smeared out" over multiple pixels.

The purpose of deconvolution is to reverse this "smearing out" of single units of detail. The cause of the smearing can be anything from atmospheric conditions to optical train light scattering particularities. But the fact remains; if there is no longer smearing out of detail in your data (either because it's a wide field, or because you binned the data) then deconvolution can't do much for you.

In fact when dealing with widefields, it can make another problem worse; busy star fields (see further down below). The latter may also be exacerbated by using the Sharp module to enhance small detail. If using it on a wide field, try reducing the lower scales - there is plenty of small detail already (it's a wide field!), but larger scale detail is less visible as the busy star field is distracting. You can use the Sharp module (and Life module - see below) to only enhance large scale structures by only enhancing the larger scales.

You use the Contrast module, where you increase the Dark Anomaly Filter. You, rightly so, don't see much difference doing that. The Dark Anomaly Filter here is very similar to the one in Wipe - it keeps Contrast (and Wipe) from getting confused by small darker-than-real-background specks, dead pixels or noise. Halos may form otherwise (where Contrast/Wipe back off as they try to protect that aberrant pixels from clipping). Your data is of good quality, so no tweaking was needed.

As a side note, here I have to remark that, to any untrained observer, you certainly seem to know what you're doing. You are already effectively expressing your vision for your data using the various tools. You do so in a way that is non-destructive to your data and yields a valid interpretation of your data. Understanding the tools better will of course help home into your personal taste much more.

The Flux module is better left alone, especially in wide fields. The reason is that Flux uses self-similarity to detect and interpret detail; gas flows/knots/shock fronts at a larger scale tend to behave like gas flows/knots/shock fronts on a smaller scale. The Flux module uses that similarity to detect and enhance detail across scales. The trouble is that, in widefields, at smaller scales stars take over and are too numerous. Stars are simply point lights and don't behave/look consistent across different scales.

For wide fields like these, I can heartily recommend the Isolate preset (with full mask set!) in the Life module. It will push back heavy and distracting star fields and re-focus the viewer's attention on larger scale structures.

Finally, you go into the Color module. You mention the colors don't look like other renditions found on Reddit, which can be a good thing, deppending on your preferences (you can emulate the "Reddit" way by changing the "Style" away from Color Constancy.
The reason why StarTools' color renditions look different (by default) is that StarTools records, recovers and maintains color ratios as they were recorded. Other software - regrettably - stretches color along with the luminance detail, resulting in wildly distorted hues and saturation values. Things in outer space don't magically change color just because you decided on a particular exposure length or stretch curve. M42's core is not white, it is green. M31's core is not white, it's yellowish.

You will notice that in your rendition, the sky is peppered with all manner of star temperatures. As is the case in reality (in fact StarTools relies on this fact to achieve reasonable default color balance).
Though this is slowly changing, unfortunately, many AP'ers still depict star colors as uniformly yellow or white.

After noise reduction, you can indeed modify the star field, though some times this is not even needed if the Isolate preset has been used.
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
 
Posts: 1553
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne

Re: Video of me processing, please help

Postby admin » Tue Jun 27, 2017 2:56 am

Downloading Andromeda now...
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
 
Posts: 1553
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne

Re: Video of me processing, please help

Postby rottielover » Tue Jun 27, 2017 3:51 am

Taking into account your advice I quickly reprocessed two of the images. Not sure if this post will work to show them but here goes....

Anteres Reprocessed


Image
rottielover
 
Posts: 22
Joined: Wed May 24, 2017 2:26 am

Re: Video of me processing, please help

Postby rottielover » Tue Jun 27, 2017 3:51 am

and Andromeda reprocessed

Image
rottielover
 
Posts: 22
Joined: Wed May 24, 2017 2:26 am

Re: Video of me processing, please help

Postby admin » Tue Jun 27, 2017 3:58 am

Thanks again for uploading the data and video.

First off, I also like diffraction spikes ;) Besides giving a visual cue with regards to brightness of a star, they can also be great for color balancing sanity checking (the "rainbow"/spectrum artifacts can really help).

--- Auto Develop
To see what we got.
I can see a slightly soft focus (or otherwise trouble coalescing point lights into points rather than circles). I can further see "inverse" coma , stacking artefacts and vignetting (as you noted)..
--- Bin
To increase signal and make use of "useless" oversampling/focus blur.
Parameter [Scale] set to [(scale/noise reduction 35.38%)/(798.89%)/(+3.00 bits)]
--- Lens
Rather than going straight for the cropping tool, I'm going to use the Lens module to "pinch" the image, undoing some of the star deformation near the corners. The welcome side effect of this is that, in order to make the image square again (and not a pin cushion), the autocrop feature will crop away some parts of the image, which happen to contain the stacking artefacts.
Parameter [Curvature Linked] set to [87.44 %]
--- Wipe
Now we can do a Wipe - I used the vignetting preset, but bumped up the Aggressiveness a little as I could see a slight remnant in the center. The Dark Anomaly filter also needed bumping up.
Use the Temporary AutoDev to see a bit better what the effect is.
Parameter [Dark Anomaly Filter] set to [3 pixels]
Parameter [Aggressiveness] set to [80 %]
--- Auto Develop
Back to AutoDev to do the final stretch.
In your video, you are worried about the noisy background. Noise is the bane of our existence and will always be present.
Tony Hallas talks indeed about color mottle in one of the talks I have seen. First off, if it is the video where he uses Adobe Camera RAW, please run far, far away from the advice in that video.

The color mottle (which is not the same as color noise!) is wholly self-inflicted by his use of ACR and applying noise reduction before stacking. Mottle is a distinct pattern spread out over more than 2x2 pixels, whereas noise is a roughly random deviation from a latent "real" value which affects at most a patch of 2x2 pixels (if the data is from an OSC or DSLR and not binned).

Dithering is obviously a good idea to avoid CCD pattern noise, but Hallas says he needs it to get rid of the color mottle which is of his own making in the first place.

If your data is mostly background with a small DSO on top, you will definitely want to use the RoI feature to optimise the stretch just for the DSO, and not the background. You can also increase the "Ignore Fine Detail <" parameter which will make AutoDev ignore finer noise.

--- Deconvolution
You could attempt a little deconvolution (StarTools tends to know how far it can go by default). I found it made the stars a little tighter, but not much more.
Default parameters.
--- Life
Try using the Life module's Isolate preset (with full mask set) to push back the stars and leave M31.
Parameter [Strength] set to [69 %]
--- HDR
It use the Reveal All algorithm (1.4.x versions) in an attempt to enhance the dust lanes just a tiny bit.
Parameter [Small Detail Precision] set to [Max]
Parameter [Channels] set to [Brightness Only]
Parameter [Algorithm] set to [Reveal All]
Parameter [Dark/Bright Response] set to [Full]
Parameter [Detail Size Range] set to [911 pixels]
Parameter [Strength] set to [1.2]
--- Color
Final color calibration. I used the MaxRGB mode to check for areas of too much green and adjusted the Green Bias Reduce slider accordingly until green was not dominant any more. I killed off any remaining green with the Cap Green feature.
Parameter [Cap Green] set to [To Yellow]
Parameter [Dark Saturation] set to [8.80]
Parameter [Bright Saturation] set to [5.50]
Parameter [Saturation Amount] set to [170 %]
Parameter [Green Bias Reduce] set to [1.16]
Parameter [Red Bias Reduce] set to [1.37]
--- Wavelet De-Noise
Mostly default parameters.
Parameter [Grain Size] set to [5.3 pixels]
Parameter [Read Noise Compensation] set to [21.59 %]

Andromeda.jpg
Andromeda.jpg (560.2 KiB) Viewed 104 times
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
 
Posts: 1553
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne

Re: Video of me processing, please help

Postby admin » Tue Jun 27, 2017 4:03 am

If you'd like to upload the Antares data set, I'd be happy to have a look at it as well!
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
 
Posts: 1553
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne

Next

Return to Image Processing Troubleshooting

Who is online

Users browsing this forum: No registered users and 1 guest

cron