Before and after Ripley - I might need help...

[pmesquita]

Hello all,

I finally decided to upgrade to the Ripley version and this weekend I strated to process some more of my deep space phots with the Hyperstar, 600d and Hyperstar.

I followed all the tutorial steps I was using with the previous version - Darks & master Dark, Flats and Master Flat, Debayer Script (RGGB for canon) - and when I finally got to the Image Integration result (before I go though all of Harry's and Warren's tutorials), I did a comparison stretch of my 1st Cone (done with the previous PI version) and 2nd Cone (done with the current Ripley PI version).

And I noticed something that I also saw on the M78 Image I processed this weekend: there seem to be some RGB vertical stripes on the 2nd Image Integrated Cone nebula. I'm publishing here a comparison of the Image Integrated stretches of my 1st Cone and 2nd Cone. Bear in mind the 1st cone has 20 x 45" Iso1600 lights and the  the 2nd cone has 81 x 45" Iso1600 lights. So I was really expection a much better image...and not the veritical RGB stripes I also saw on my 2nd M78 processed this weekend too.

If you check on the two images you can see that the original Cone with only 20 lights has better quality than the more exposed one. This was not what I was expecting...:-(...

I'm suspecting of the deabyering...r some button that I wrongly pressed on the 1st steps (but I've processed two images thsi weekend and the odds of making the same mistake twice are very thin)...

Please help anyone...


cheers
paulo

[Alejandro Tombolini]

Hi Paulo,
Have you used MasterBias? It is necessary.
Saludos, Alejandro.

[pmesquita]

Dear Alejandro,

I've done all my sessions (about 18) without Bias, just darks and Flats, and I've been quite sucessful. I'm sure this problem has to do with something else, either a mistake I've done (since I'm still in experienced and I might have forgotten some butoon), or some change from one PI version to another...

The image on the left was alos done withou Bias and you can see it doesn't have the vertical stripes...

cheers
paulo

[Josh Lake]

Just a guess, but because the image is horizontally flipped, the master dark is not matching up with the light, so the vertical lines are a result of the dark not being flipped as well?

[pmesquita]

Not really, but thatnks for the hint.

the image was flipped when already Image Integrated. And the M78 I did this weekend, after installing Ripley, also came out with the the vertical stripes.

And one important thing: both the Cone and M78 were done in different days, each with their own dark session and Flat session right outside in the cold, after the lights session.

cheers
paulo

[pmesquita]

adding the M78 Image integrated result of 80 Lights (with Darks, Flats and debayered), stretched with STF..also this weekend with the newly installed Ripley.

I also added the almost finished M78 (only need to do SatBoost). You can still see the remains of the original "vertical lines" on the dark background.

You can see the problem of the vertical lines better here.

cheers
paulo

[georg.viehoever]

...I've done all my sessions (about 18) without Bias, just darks and Flats, and I've been quite sucessful. ...

Unless your camera really has no bias, you have to use bias frames. It is in the math that dark scaling uses. Only if your bias is ==0, you can omit bias frames.
Georg

[pmesquita]

but Georg, how can we explain that befoire, using the exact same technique, same exposure, same ISO (everything the same) I did about a dozen of much better image Integrations and went on to finishing the photos with better resuilts, and now with the 1st two processings on Ripley...I'm getting the vertical lines? It just doens' make sense. If it were the BIas (and I'm not saying I don't need them for improving my photos...they are my next step)...I would have had the same problem before...and I didn't...:-)

cheers
paulo

[ruediger]

I'm suspecting of the deabyering...r some button that I wrongly pressed on the 1st steps (but I've processed two images thsi weekend and the odds of making the same mistake twice are very thin)...

I also faced similar problems when upgrading from 1.7 to 1.8 because of a preference change of the FITS file format, namely "Coordinate origin". In 1.7 it was "Lower left corner", but in 1.8 "Upper left corner".

This has the interesting side effect, that CFA master files created with 1.7 have to be vertically flipped before using them in 1.8. If you don't do this, ImageCalibration might produce all kind of weird looking results. The weirdest error however was, that dark calibration might output "no correlation between master dark and target frames", which finally helped me to find the reason.

I use the new 1.8 FITS preference setting as is and use ImageContainer and FastRotation to flip old files generated with 1.7. Now everything is fine.

Rüdiger

[pmesquita]

hI Rudiger,

thanks for the tip, but I don't it's the solution. Because the master files (dark and Flat) used with these two problem sessions were all done on the same night and only processed this weekend after installing PI1.8.

so I guess I'm still flat in the dark...(copywriter pun...)

[georg.viehoever]

but Georg, how can we explain that befoire, using the exact same technique, same exposure, same ISO (everything the same)...

If Juan improved/changed the Dark Scaling/something else to provide better results than before (assuming the user provides bias), the improvement may cause worse results if you dont provide bias. You may have been lucky in your previous attempts.

Just create bias files - they cost you almost nothing. Or check that your bias indeed is ==0 (costs you 5 minutes). The mathematics of image calibration require bias files unless your bias is ==0!

Georg