The 'Halo' effect around bright areas

bud martin budzynski

Following on from what James had said on SGL regarding a faint halo around bright edges, I've been studying my Lunar images taken with the 120MM and noticed these inside various craters. These are early days yet for this hardware -much has still to be learnt as to best settings etc.

If you zoom in on the small craters in this attached image you can see the effect.image.jpg

jamesf

I can definitely see that at 200% size. It's particularly obvious in the craters in the centre section of the image, but I also think I can see it inside the right hand lip of the craters at the bottom left.

James

stuw

Has the image had wavelets applied?

Sent from my HTC One X using Tapatalk 2

bud martin budzynski

Yes, this image was processed via AS!2 and R6. I'll run the original avi in Virtual Dub tomorrow to see if any of the individual frames show the same effect.

Will give you a positive answer sometime tomorrow.

Sam@ZWO

Is it over exposured when capture?

jamesf

I have a similar problem with images of Jupiter. The finished (sharpened) image has a pale ring outside the limb. In the unsharpened image this appears as a more fuzzy halo around the limb. I think I have an example frame from the captured avi that shows a blur around the limb which is causing the problem, but I haven't worked out how to split one single frame from the avi yet. Once I've done that I'll post all three images.

James

bud martin budzynski

Hi Sam, I've just been looking at some 120MM Lunar avi's and can report that the 'halo' effect is present in the source video stream. I extracted the sharpest frame from the avi and had a good look at it in PhotoShop CS5.

James, if you have Virtual Dub (free) then all you need to do is:-

File>Open video file

... move the slider at the bottom of the screen (or use the across arrows ) to find a suitable frame then

Video>Copy source frame to Clipboard (Cntrl + 1)

.... then open it up ( ie Paste) it into any image software.

The sample frame I looked at shows the full halo inside small craters as well as on the outside edges of brightly lit (overexposed) crater rims.

I don't have any footage from the MC camera to make comparisons but I did take some Moon avi's with the IDS e2v Ruby sensor which shows no halo effect but, of course, the pixel size of this camera is 5.6um as opposed to the 120's 3 .75um so perhaps is more tolerant of overexposure leading to bleeding over to the adjacent pixels. More research required.....

Bud

jamesf

I shall try downloading VirtualDub this evening and see what I can sort out.

James

jamesf

Ok, here we go. This first image is a single frame from a 50fps capture run with the ASI120MC, Skywatcher 127 Mak and 2x barlow. The gain was fairly high for this, but according to the histogram in FireCapture I was not reaching saturation. I was however following Darryl's advice of trying to get the histogram quite full, so it was up around 85% to 90%. You'll see there's a "halo" effect around the bright limb. This occurs at all frame rates and gain settings I tried, but it's most visible in this run which is why I've used it despite not being a particularly great capture. It's not blurring due to motion of the image, which at this point in the capture was fairly stationary.

This next image is how a stack of half of the 5000+ frames captured comes out. Clearly the halo is in enough good frames to easily make it through the stack. This image happens to be from Registax, but AS!2 does the same thing

And finally, how the above image appears once wavelets are applied

My feeling is that the problem is somehow related to light intensity, but I haven't worked out how at the moment. It's as if the bright limb of the planet is bleeding into adjoining pixels, but not in the same way that it does if you completely blow out the image with too much exposure, for example.

James

stuw

I'm not getting anything like this on my captures. I've attached a frame to show what I mean.

jamesf

I don't get them all the time, but it's not at all clear what combination of circumstances leads to the effect.

James

stuw

Are you using the IR filter?

bud martin budzynski

Hi James & Stu, ok, we have an interesting mix here .... can you look at your respective text files that FIrecapture produces and see what your ....

a) frame rate

b) histo

c) shutter

d) gamma

e) focal length

..... is.

My results are similar to James'.

There are some other factors like the 'raw' checkbox and what ROI was being used at the time.

Shame we don't live close enough to each other -I'm sure we could solve this in no time at the local pub.

Cheers

Bud

jamesf

I believe at some point or other I have solved all the world's problems in the pub. Unfortunately I could never remember the solution the next morning :)

I'll dig out the file and check what my settings were.

James

jamesf

I am using an IR filter, yes. Here are the capture settings for the image I posted earlier. I have others with longer exposures and lower gain settings that show the same problem. I'll dig one of those out a well.

James

FireCapture v2.2 beta (Build: #37) Settings

-----------------------------

Camera=ASI120MC Camera (ZWO Design)

Filter=L

Profile=Jupiter

Diameter=40.29"

Magnitude=-2.37

CMI=295.6� CMII=20.4� CMIII=100.9�

Filename=Jup_20130219_211057.avi

Date=20130219

Start=211057

Mid=211157

End=211257

Duration=120s

Frames captured=5914

Binning=no

ROI=640x480

FPS=49

Shutter=20.02ms

Brightness=8

Gain=95

Gamma=50

Histogramm(min)=0

Histogramm(max)=227

Histogramm=89%

Noise(avg.deviation)=0.00

AutoAlign=false

PreFilter=none

Limit=120 Seconds

stuw

I know it says Moon, but I forgot to change it over from the day before!

FireCapture v2.2 beta (Build: #37) Settings

-----------------------------

Camera=ASI120MM Camera (ZWO Design)

Filter=L

Profile=Moon

Filename=Moon_204140.avi

Date=190213

Start=204140

Mid=204310

End=204440

Duration=179s

Frames captured=38887

Binning=no

ROI=320x240

FPS=217

Shutter=2.785ms

Brightness=10

Gamma=51

Gain=51

Histogramm(min)=4

Histogramm(max)=181

Histogramm=70%

Noise(avg.deviation)=5.71

AutoAlign=false

PreFilter=none

Limit=180 Seconds

jamesf

Here's one with a much lower gain setting (and slightly lower histogram fill) that also shows the problem:

FireCapture v2.2 beta (Build: #37) Settings

-----------------------------

Camera=ASI120MC Camera (ZWO Design)

Filter=L

Profile=Jupiter

Diameter=40.29"

Magnitude=-2.37

CMI=275.2� CMII=0.1� CMIII=80.7�

Filename=Jup_20130219_203725.avi

Date=20130219

Start=203725

Mid=203825

End=203925

Duration=120s

Frames captured=3582

Binning=no

ROI=640x480

FPS=29

Shutter=33.04ms

Brightness=8

Gain=68

Gamma=50

Histogramm(min)=0

Histogramm(max)=214

Histogramm=83%

Noise(avg.deviation)=0.00

AutoAlign=false

PreFilter=none

Limit=120 Seconds

James

jamesf

Any ideas why I have 0 for the "Noise" line and you don't, Stu?

James

stuw

No Idea, maybe it would be an idea to ask Torsten what it means.

jamesf

Google suggests:

"That noise value is the simple standard deviation of neighboring pixels. Don't

read this as an absolute value and it's also not related to the histogram max

value. It's based on Emil's research (see CN forum) and it's only purpose is to

give an indication if switching to 12bit recordings would give any advantage

compared to 8bit. This would be the case if the value is < 2 when it turns red

in the preview window."

No idea why mine should be zero though.

James

~~Next Page »~~