PDA

View Full Version here: : Help my Eta Carinae is too bright


Spookyer
07-08-2017, 10:04 AM
Hi all

I am trying to process and Eta Carina mosaic which I got plenty of data for. Unfortunately when I combine my 3 channels (Ha,OIII,SII) in Pixel maths I am clipping some of the image as I appear to have too much signal in the brighter areas of the neb.

I went back and tried to stretch my three base images less but when I do that I lose some of the fainter areas. Is there a way to stretch the faint areas without stretching the bright areas as much? If I just use the middle slider like I usually do in Histogram Transformation I am getting too much signal in the bright areas as well.

Brett

Lognic04
07-08-2017, 12:47 PM
I would try maskedstretch(only stretches dimmer parts of the image), or make a range mask and stretch separately.

RickS
07-08-2017, 12:54 PM
Hi Brett,

Try turning on the "rescale result" option in PixelMath. That will keep the result from being truncated if it goes out of range.

I've never had a problem like this but I always do the colour combine with linear data and then stretch after that.

Cheers,
Rick.

Spookyer
07-08-2017, 01:26 PM
Here is what I have so far. Seems to be clipped out in places. What do you think?

https://flic.kr/p/Xu2BZa

Brett

RickS
07-08-2017, 02:01 PM
Yes, I agree.

Rescaling the PixelMath result may give you a quick fix. In the longer term I'd recommend doing the colour combination while your data is still linear. You really have to do that for RGB if you want accurate colour and I can't think of any reason to do NB colour any differently.

Cheers,
Rick.

Spookyer
07-08-2017, 02:53 PM
I have been following the Light vortex tutorials, with his theory being you stretch the channels first and get the mean and median values similar so that when you combine them one channel doesn't overwhelm the others. If you combine first you must stretch the whole image at once, how do you stop the strongest channel dominating the image?

Brett

RickS
07-08-2017, 06:29 PM
I use LinearFit to match the three linear channels before combining them. An alternative is to combine the channels and do an unlinked STF stretch - you can easily check with STF what this is going to look like with a screen stretch.

Spookyer
07-08-2017, 06:40 PM
thanks Rick, will have a look at that. I do a Linear Fit in my workflow but not really sure why exactly.

Brett