Creating a multi exposure length HDR composite

[christian]

Very much thanks for the Info Ivo, it was just for interest, i think the next object in range would be the orion nebula or the andromeda galaxy to test, but there i must wait, because it´s very deep now and i don´t have a view to the east. I was confused because I tought that the result would be bader, if i take another stack to the new stack with another exposure, it would disort the result, but eventually it will be sorted out, i will test i will see a difference.
Oh dithering is no problem with the mgen, just turn on and when i see my data i think i must not take darks

[micheleorsini]

Hi!
I have a problem with this technique. Here is my story: just for testing how to to HDR, I made this test
1. took two sets of frames of M31, one with 400 ISO (4 minutes, newton 200mm f4 with IDAS) and one with 100 ISO (same params)
2. stack them separately
3. post processed them separately with Startools and my usual workflow, saved as tiff
than I tried to work with the layer module as described here but the result is horrible!

hdr.jpg (422.25 KiB) Viewed 13322 times

playing with kernel radius does not solve the problem. What's going on? any other information needed?

thank you

Michele

[admin]

Hi Michele,

Thank you for attaching that image.
After testing with 1.4, I noticed I had a bug that made the Max Contrast algorithm no longer work (the Max Contrast filter is definitely the preferred filter for doing HDR composites).
This should now be fixed in the latest alpha (see download section).
I have also increased the sensitivity for the kernel radius as well.

Do let me know if this helps!

[micheleorsini]

oh man it worked!

-----------------------------------------------------------
StarTools 1.4.324
Sun Sep 18 16:43:42 2016
-----------------------------------------------------------
File loaded [/Users/morsini/var/20160902_M31/20160902_M31.tiff].
---
File loaded in layer module [/Users/morsini/var/20160902_M31/20160902_M31_HDR.tiff].
--- Layer
Parameter [Layer Mode] set to [Blend]
Parameter [Cap Mode] set to [Clip]
Parameter [Brightness Mask Mode] set to [Where composite is light, use bg]
Parameter [Filter Type] set to [Max Contrast]
Parameter [Blend Amount] set to [100 %]
Parameter [Mask Fuzz] set to [1.0 pixels]
Parameter [Filter Kernel Radius] set to [51.0 pixels]
Parameter [Offset X] set to [3.0 pixels]
Parameter [Offset Y] set to [-2.0 pixels]
Parameter [Brightness Mask Power] set to [1.00]
File saved [/Users/morsini/var/20160902_M31/20160902_M31_composition_3.tiff].

thank-you!!!
that's what I call awesome customer support

20160902_M31_composition_3.jpg (457.65 KiB) Viewed 13312 times

[micheleorsini]

I'm writing again on this topic because I got a result that is different from what I expected. I did as this post suggest with two images of Orion Nebula running Startools 1.4.327 on a Mac OS.

Here is the first thing you see

layer 1.jpg (361.86 KiB) Viewed 13150 times

so I tried to increase kernel radius and I got this

layer2.jpg (415.12 KiB) Viewed 13150 times

finally, I reduced the blend amount and things got better but I was expecting something different

layer3.jpg (403.5 KiB) Viewed 13150 times

any thought/comment?

Michele

[admin]

Hi Michele,

That's indeed what I would expect for minimum distance to 1/2 unity (1/2 unity is grey); it selects the correct pixels. The two images are fairly extreme when it comes to differences in exposure (not a bad thing of course). How does the other option (Max contrast) perform?
Know that you can also Copy the result (which you may think is a bit stark), then Paste that as foreground or background (e.g. you just used the output as a new input!). This effectively performs multiple iterations of the combine and may yield results that are more to your liking.

Does this help?

[micheleorsini]

Thank you Ivo,
I'll try as you suggested

[micheleorsini]

you-are-awesome! go Startools