2017 eclipse - HDR

[kkretzsch]

Hi Rob,

what i did was to use pixelmath to replace the moon with the color of the corona that was just outside of the moon's edge. i believe that removing the brightness discontinuity represented by the moon helps the HDRMultiscale transform work better without ringing.

I have a similar problem with a ringing effect after applying the HDRMultiscale to the HDRcomposed image. How did you replace the moon with the color of the corona using pixelmath? Did you create a special mask to define the moon? Which expression did you use?

Thanks
Klaus

[pfile]

i just used the readout cursor to measure the corona near the edge of the shadow of the moon, and yes i used a circular pixelmath equation to fill it in. IIRC i had to do it twice because the shadow of the moon is the union of all the shadows of the moon in each sub exposure and of course the moon moves across the face of the sun.

i took the circular PM equation from one of david ault's presentations but either made it an inequality or made it a really thick line (i am not in front of the computer i did the processing on... can't remember). i think a circular mask made with his PM expression would have worked just as well.

the result still has kind of a circular white area around the moon which is not present in other images of the eclipse. i think this might be an HDRT artifact of its own, but i never really got to the bottom of it. i guess i thought it looked kind of cool so i just left it.

rob

[kkretzsch]

i just used the readout cursor to measure the corona near the edge of the shadow of the moon, and yes i used a circular pixelmath equation to fill it in.

Thanks Rob, this indeed solved my problem. I used the following PixelMath expression

iif( sqrt((x()-x0)*(x()-x0) + (y()-y0)*(y()-y0)) < r, h_R/G/B , $T )

, where (x0,y0) is the center of the moon shadow, r is the radius (in pixel units) and h_R/G/B the intensity values of the R/G/B channels. After applying  this PixelMath expression the HDRMultiScale trafo produced much better results. I applied LHE and Larsson-Sekanina to the stretched image (MaskedStretch) to protect the bright structures close to the sun surface.     

[pfile]

that looks great - the contrast in the corona is really strong.

one thing that eluded me was to get all the way to where druckmuller is with his corona images. however i believe that he has written his own software for processing corona images... one characteristic of his images is that they are very flat - the large scale halo around the sun is pretty much gone. i played around a little bit trying to extract a large scale image to try to divide or subtract, but i was not successful with that.

rob