Author Topic: Frame Adaptation Function  (Read 5254 times)

Offline Michael Hernandez

  • Newcomer
  • Posts: 8
    • Darkfire
Frame Adaptation Function
« on: 2010 November 04 09:56:11 »
Ok, I've searched all over the program for the Frame Adaptation Function as shown in the video tutorial "Mosaic Construction using Star Alignment" and can't find it. In the video, the function was already open on the desktop. Can someone point me in the right direction?
Michael Hernandez

Offline Nigel Ball

  • PixInsight Addict
  • ***
  • Posts: 277
    • Astrophotography by Nigel
Re: Frame Adaptation Function
« Reply #1 on: 2010 November 04 16:15:33 »
Michael

Are you referring to the video tutorial on the PI website? If so then this is using the PixelMath function to match the frames

This is now done more easily by checking the Frame Adaption tickbox on the Star Alignment function

HTH

Nigel
Nigel Ball
Nantwich, Cheshire, United Kingdom

Takahashi FSQ-106 at f/8, f/5 and f/3.6 on AP900, Nikon 28 mm and 180mm f/2.8
SBIG STL-11000M, Astrodon LRGB, 5nm Ha
ST-10XME, Astrodon HaLRGB
www.nigelaball.com

Offline Michael Hernandez

  • Newcomer
  • Posts: 8
    • Darkfire
Re: Frame Adaptation Function
« Reply #2 on: 2010 November 04 19:54:58 »
Michael

Are you referring to the video tutorial on the PI website? If so then this is using the PixelMath function to match the frames

This is now done more easily by checking the Frame Adaption tickbox on the Star Alignment function

HTH

Nigel

I had checked the Frame Adaptation tickbox, which gave a valiant effort but fell short of being seamless in the mosaic. Then I watched the video on the PI website, which seems to refer to older functions, which apparently are no longer available. PI did no better at balancing the backgrounds and highlights than ImagesPlus. The best mosaic tool I've found so far, especially at balancing the dynamic brightness range, is the photomerge tool in PSCS4. Unfortunately, I can't use it for this data set because one of the frames did not have enough overlap to satisfy Photoshop. I am hoping someday I'll be able to merge these four quadrants.....
Michael Hernandez

Offline Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 7111
    • http://pixinsight.com/
Re: Frame Adaptation Function
« Reply #3 on: 2010 November 05 01:16:22 »
Hi Michael,

A few questions:

- Are your images linear, or are they stretched (nonlinear) images? In PixInsight, you must always build your mosaics with linear (raw) images.

- Have your images dissimilar additive gradients? If so, no linear frame adaptation routine (as the one implemented in PixInsight's StarAlignment tool) can solve the problem. You must remove all additive gradients prior to mosaic construction.

- Similar to the previous question: Are your images correctly calibrated? Accurate flat-fielding (that is, uniform illumination on each mosaic frame) is an essential precondition for successful frame adaptation.

If your images are linear, well calibrated and have no gradients, then our frame adaptation routine will yield a seamless mosaic that will be strictly a linear image. As you probably know, linearity is a very important procedure that will allow you to process a large mosaic just as you can process a single raw image in PixInsight.

See this tutorial, where I describe the step-by-step procedure to build a large four-frame mosaic:

http://pixinsight.com/forum/index.php?topic=2011.0

In the above example, the images were linear and calibration was very accurate. The final 3x3 mosaic was generated by Antonio PĂ©rez with the same procedure described above:

http://www.datsi.fi.upm.es/Hyperion/objetos_web/IC1318_ing.html

Hope this helps.
Juan Conejero
PixInsight Development Team
http://pixinsight.com/

Offline RBA

  • PixInsight Guru
  • ****
  • Posts: 511
    • DeepSkyColors
Re: Frame Adaptation Function
« Reply #4 on: 2010 November 05 02:08:24 »
Mike, you need to review the slides from my AIC presentation :P

It was kind of like: 1) good flat calibration + 2) gradient removal + 3) linear adaptation.

I've built quite a lot of mosaics, not all of them have been successfully built with PI alone, but that's also due to the fact they're often rather massive mosaics 12, 24, 52 frames...

1 isn't very hard to achieve, at least getting close enough. 3 is done by the software. All you have left is 2.

Now, Juan says that IF your images are correctly calibrated and IF you've removed the gradients correctly, PI wil succeed. It will. What Juan doesn't tell you is that PERFECT gradient correction sometimes isn't easy. Sometimes it can be very hard. Sometimes, for the purpose of putting together a seamless mosaic while still inear, it's simply IMPOSSIBLE

That's why when building a mosaic I spend A LOT of time working on that (gradient correction). When most people are usually content with whatever their first DBE run gives them, I rarely use the first, second or third try, and keep modeling the background until I'm happy (I do not mean I run the DBE 3 or 4 times, but that I try, don't like it, go back and try it differently, and so on). On non-mosaics one doesn't need to be brutally accurate. When building mosaics, you have to. If you can't then you have to default to proceed with what you have and MANUALLY build your own background model afterward, something you cannot do with PI alone.

I'm familiar with Photoshop's Photomerge. I've used it a few times to put together already-processed images to build a larger mosaic, but for smaller mosaics you still haven't processed, I definitely suggest starting with the PI workflow. The DBE is a very powerful tool. See in what areas the seams are noticeable and model your background accordingly, by adjusting the parameters, replacing samples in different locations, etc. Always, ALWAYS look at your screen-stretched background model and see if it shapes according to the background and the areas where the seams are more noticeable. Rinse and repeat. Often, the time you spend on that won't be wasted (going the PS way and building your background model manually isn't any less time consuming), plus you'll become very good at using the DBE tool!





Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4729
Re: Frame Adaptation Function
« Reply #5 on: 2010 November 08 11:23:44 »
Mike, you need to review the slides from my AIC presentation :P

not to hijack the thread, but are these actually available yet? i have not received any email from the conference organizers regarding the slide availability...

Offline RBA

  • PixInsight Guru
  • ****
  • Posts: 511
    • DeepSkyColors
Re: Frame Adaptation Function
« Reply #6 on: 2010 November 08 15:30:50 »
Mike, you need to review the slides from my AIC presentation :P
not to hijack the thread, but are these actually available yet? i have not received any email from the conference organizers regarding the slide availability...

Probably within the next few days or couple of weeks at most I would guess.