›

Deconvolution Question

Wayne Hixson
in PixInsight
Love your tutorials, I almost feel like I'm back in college! I try to follow along with my own images as you go. One thing I've always had trouble with is deconvolution, to the point I never use it, so really have enjoyed starting that lesson. I have 2 masters, an Ha and an O3 - do you do deconvolution on each sub or after you combine?

Wayne

Comments

  • BlockHeadBlockHead
    Hmmm... NB is a different creature as you mention in your other message. 
    Typically what is done is to create a synthetic luminance (perhaps just out of the channel with the brightness signal- or a weighted blend of all of them) and then deconvolve that one image. Subsequently you color it with your mapped color image. That is the approach I would use. I would caution, however, to really consider how beneficial deconvolution will be to the image. Many diffuse nebula that lack high signal small features will not benefit. Another thing to consider is that deconvolution, like with my data, works well with critically sampled or overly sampled data. (big fat stars and features). 

    The combination of star de-emphasis plus contrast adjustments to the nebula can be just as powerful as deconvolution- and without the dangers its comes with. 

    But for a globular cluster, planetary nebula, or a galaxy... deconvolution is wonderful.

    -the Blockhead
  • Wayne Hixson
    I guess I should have mentioned what I was shooting - it was the Crescent Nebula in Cygnus. I thought decon might work well with the details in that nebula.
  • BlockHeadBlockHead
    Ngc 6888 does have small details, but usually requires an oversampled image with very good signal. So platescales of less than 0.6”/pixel and a moderate light bucket.

    Even with the image I worked on the nebula really wasn’t bright enough to benefit much from deconvolution.

    http://www.caelumobservatory.com/gallery/n6888.shtml

    In fact deconvolution will enhance stars of the image. So for an object like this I think the combination of de-empasizing stars plus contrast enhancements of the nebula with other tools is better. This is just my style to go light if at all on on object like this.

    If your data is narrowband, then you are ahead because stars are not as much of an issue.

    -the Blockhead
Sign In or Register to comment.