Seeing holes in stars after HDR Composite

[sctall]

Hi All

I am having issues where after I do an HDR Composite on 3 exposure times on M42, I end up with stars that have holes in the middle.
Or at least that what it looks like.


My process was, I had 3 different exposure lengths.
30 sec, 60 sec, and 240 sec. several subs for each.
I used BPP separately for each exposure time, using their calibration subs.
Once I had the calibrated images from BPP, I performed a DBE, BN, and CC on each.
I then used the HDR Composite tool to combine the 3 images.   No HDR Transform done yet.
I checked the 3 individual exposure images, and they did not have this issue.

I have put some .fit files up on dropbox if anyone wants to look at this.
They have had the DBE, BN, and CC done.  Also a couple needed the Linear Fit.

The 3 files are titled appropriately, and the HDR showing the issue, is the HDR_comp.fit file.
There is also a jpg of my attempt to complete the processing, hoping it might clean itself up.  But no.

There is some noise on these images.

The files can be found at:  https://www.dropbox.com/sh/vhlut01p9p81qjn/TKRpil3XP8

I processed this earlier, by not doing the HDR Composite until I had performed some Noise reduction and stretching.
Then, Separating a Lum image after the HDR Composite.  When I did it this way I did not have this issue, but I found that after I combined the LRGB, that the core had a big artifact in the middle.  By doing the HDR Composite just after the DBE, BN,CC, I found that I did not get that issue, but I see the holes in the stars now.

Scott T.

[jkmorse]

Scott,

Just out of curiosity, what capture software did you use.  I have been having similar issues with images captured by MaximDL.

Jim

[sctall]

Jim

I use APT (Astro Photography Tool).

I have not seen this on any other images. But this is the first time I have used the HDR Composite, to combine multiple images.
Are you seeing it on the calibrated images or only after you combine with HDR Comp?

Scott.T

[jkmorse]

Mine happens on raw captures, without any processing so I am having a different issue.

Jim

[georg.viehoever]

Probably you should do the HDRCombination before DBE, BN, and CC . I am not sure if the images loose some properties that HDRCombination requires (ADU strictly proportional to photons collected) if you do these processes first.
Georg

[sctall]

Thanks Georg

I tried it on the base images just after the BPP processing.  I keep getting a

Error: Inconsistent HDR composition detected (bad linear fit, channel 0)
<* failed *>   message.
I ran Batch Linear fit on the 3 files, they succeeded, and then I loaded the 3 new linear fit files into HDR_Comp, and I fail.

What exactly does the Linear fit tool do?

Scott.T

[georg.viehoever]

Hi Scott,

A linear fit between image X and Z does this: it determines a and b in  formula y=a*x+b, such that the pixels in image X are transformed to to image Y using the formula, and Y most closely matches image Z.

I am not sure why HDRCombine fails for you. Usually that means that those images are not really linear, or that they come from different sources with incompatible characteristics (for instance a CCD camera and a DSLR), or that they are not properly aligned, or that they are not properly calibrated. Difficult to say without having the images.

Maybe you upload the calibrated images somewhere, and maybe someone has a look.
Georg

[pfile]

i looked at the 'processed' images above - they are out of focus and the stars are really "flat". also despite the different exposure times, the images don't really seem that different. the trapezium is not blown out in any of them. i wonder if HDRComposition is not able to properly compute what data to take from what image since they are all so similar.

rob

[sctall]

Hmm!!!

Don't know what to say. They were all taken on the same night, all I did was shorten the exposure time.
All were processed separately with BPP.
My 4 min image , shows the trapezium completely blown out.
How are you measuring this?
The 3 images although it has the blown out stars does make a huge difference with HDR_Comp. If I use it on just the 240sec image, the core is still blown out..
Darn, I thought I had my focus pretty good,
I don't have a flattener, so I can't control the focus across the whole image.

After further testing, the 60 sec image is the one that does not fit.  I can use the other 2. Don't know why.

I guess I will start all over with the BPP and see what happens.
I did put the orig images that have had no processing beyond the BPP output up on drop box.
They are the _base files

Scott T.

[pfile]

i had downloaded the _base files. you can see from my screenshots what your 240s image looks like with no STF. the trap seems to be okay. in contrast also attached is a 1200s Ha image from my f/4.4 system. you can see that the trap is totally blown out; no STF is applied here either.

rob

[sctall]

OK I see

I could never get to that point here.
LP is too bad.

Can you do the HDR_Comp with the 3 _base images?

Scott T.

[pfile]

yeah - i tried with the _base images and had the same problem as you are having…

rob

[sctall]

OK Rob

Thank you
Not sure where to go from here...

Scott T.

[Andres.Pozo]

I have downloaded your images and I think that your problem is that the images are not aligned. HDRComposition requires that images are aligned.

The second problem is that the brightness of the images doesn't seem to be proportional to the exposure. The 60s image is brighter that the 240s one.

Your images have other problems also very important:

• The images are very badly focused. This is the first step in astrophotography. All other techniques (guiding, dithering, calibration, ...) are useless if you don't focus the telescope.
• In all the images the brightest stars have an strong reddish cast.
• In the 240s image there is a strange color noise in the Trapezium area.

In any case, I have aligned the three images and the HDRComposition runs successfully. I have attached the result of applying these processes:

• Align the three images
• HDRComposition with "Binarizing threshold" of 0.3
• DBE, BackgroundNeutralization & ColorCalibration. (I should have applied DBE before HDRComposition)
• HistogramTransformation for stretching the image
• HDRMultiscaleTransformation for flattening the brightness in the Trapezium
• CurvesTransformation for increasing the contrast
• SCNR for removing a residual green cast

[sctall]

Thanks Andres

I did a SA on the 3 images prior to the HDR_Comp and it worked.

As far as for the focusing, I could use some advice. I see the focus other than the flattening across the image as pretty good. So I am missing something here.

I use a mask and APT software which has a tool to focus with. The way it works is you align 2 circles on top of each other and it's focused.

The seeing was not great, but better than some nights, I could not get it exact, but it was pretty close.
So I am confused about why it looks out of focus.

Anyhow, is there a tool I can use to look at a sub, and evaluate the focus before I start to run a planned integration?
For example: I looked at a 240s sub with the FWHM script and saw a huge difference between a jpg and a .CR2 sub. And I saw a huge diff between , loading as greyscale a .CR2, and loading same image in RGB, and then converting to greyscale. The .CR2 after converting to greyscale had a pretty good value. The loading as greyscale ( which comes from the settings in DSLR format). have a pretty high value.  see attached for both .CR2 files

Since I do not fully understood all the image analysis stuff, I don't understand why the same image show different results.
And of course what would the right way to do this?

As far as FWHM goes, what is a good value to shoot for in a NOT ideal location.  Somewhere around 2-3 arc/sec?

Is this a tool to use, or is the FWHM value so subjective it  is  just is not relevant. And are there other tools? What I want to do here is just start to "focus" on just focusing, to see if I can dial this in better.

appreciate any help.

Scott T.